The Limits of Legitimacy: Dissenting Opinions, Media Coverage, and Public Responses to Supreme Court Decisions

Citation preview

Page viii → Page ix →

Acknowledgments I am grateful the opportunities provided by the Departments of Political Science at the University of Michigan, DePaul University, and the University of Kentucky. Staff, faculty, and colleagues alike contributed to the work presented in this book in ways both concrete and immeasurable. The book is also much better off for the support I received from the Gerald R. Ford Fellowship and other grants made available during my time at Michigan. I have received invaluable feedback on the project from a variety of scholars at conferences and seminars. Their questions, suggestions, and advice have enabled me to develop my research from a simple idea into a fully formed project in a manner that I could not have anticipated. I am particularly grateful for the feedback of colleagues in Robert Axelrod’s course on framing at the University of Michigan, where the idea for the project first took shape. Adam Weissler and Vanshika Vij, my research assistants, contributed in a variety of ways to this project. I thank them for all of their comments, effort, and time. Phoebe Ellsworth helped to improve my facility with research design at the early stages of this project. She brought a constructive interdisciplinary perspective to the project, encouraging me to think about both judicial behavior and experimental methods in new ways. Ted Brader offered provocative feedback at many stages of the project, forcing me to think carefully about the relationships among public opinion, media, and the Supreme Court. Chuck Shipan has been a great mentor, and I am grateful to have taken courses with him, worked with him, received his insight about the project, and most of all to have forged a close relationship with him. His guidance in many ways is responsible for much of the research that informs this book. Nancy Burns has been a wonderful source of support. Her thoughts, insight, and willingness to help whenever I needed it were essentialPage x → to this project. She constantly challenged me to think more broadly and deeply about my work. Her enthusiasm is infectious. She is a great adviser and mentor. Because writing a book is such a monumental endeavor, professional support and academic feedback are not nearly enough. This book would not have been possible without the love and support of some very special people in my life. My mom and dad spent many years sacrificing for me and instilling in me the traits that are now borne out in these pages. They have always encouraged me to aim high, follow my dreams, and make them a reality. They’ve been there for me whenever I’ve needed them, from my first days of kindergarten to the final stages of this process. So too has my sister, Molly, a wonderful (now) young woman and friend that I’m proud to have for late night phone calls. I’d like to thank all of you for everything you mean to me. Finally, one other person has been instrumental in everything about this project and she, too, has been there for me since my first day of kindergarten. My wife, Katrina, has been an irreplaceable part of my life; she has spent the past few years encouraging me and supporting my work, even as we maintained a long distance relationship, while we prepared for our wedding, and eventually as I finished this manuscript over many long nights. Without her, none of this would have been possible.

Page 1 →

Chapter 1 Introduction Interest in the U.S. Supreme Court has perhaps never been higher than it was on June 28, 2012. On the final day of its term, the Court was prepared to announce its decisions in a series of cases that would determine the fate of the Patient Protection and Affordable Care Act (ACA), more commonly known as Obamacare. Health care reform had generated substantial controversy since at least 2009, when Congress began debate on a series of related bills. Disagreements lingered long after President Barack Obama signed the ACA into law on March 23, 2010. Challenges to the law worked through the judicial system quickly, before most of its provisions could take effect. In November 2011, the Supreme Court granted certiorari to three such cases. The Court allocated an unprecedented three days in March 2012 for oral arguments in these cases, at which point it became clear that five justices had serious reservations about the constitutionality of the ACA. Two factors conspired to heighten interest in the decision. The Court’s deliberation generated a substantial amount of drama, since the notoriously secretive institution had given few hints as to how it would rule. Indeed, most observers expected the fate of the law to hinge on the vote of a single justice (most likely Anthony Kennedy, who often sided with the majority in high-profile cases). The decision would be announced on the final day of the Court’s term. At the same time, the consequences of the Court’s decision were monumental: in all likelihood, the Court would determine the ultimate fate of the law. For opponents of Obamacare, the ruling represented what appeared to be the final chance to kill a law they had denounced yet failed to defeat for years. For supporters, the ruling would either consolidate their victories or thwart what had been decades of agitation in favor of a national health care law. If the Court struck down the law, there was little chance the Congress could pass a constitutional piece of health care legislation in the near future, if ever. Page 2 →On the morning of June 28, the courtroom filled with observers. Outside, thousands of activists talked, chanted, and stewed. Major media outlets had focused on the litigation for months, with coverage peaking in recent weeks. Reporters waited inside the courtroom and on the courthouse steps. Cable news offered live, uninterrupted coverage (and speculation) throughout the morning. One of the leading websites for judicial news—scotusblog.com—reported 866,000 readers on its live blog and 5.3 million hits during the day.1 On June 29, national newspapers followed up with front-page reporting and analysis. Coverage of the ruling continued to saturate the news long after it was announced. In a landmark decision, the Supreme Court upheld most major provisions of the ACA, including its most controversial piece—the “individual mandate” that all Americans purchase health care insurance. Surprising observers, the Court found the mandate beyond the scope of Congress’s power to regulate interstate commerce but a permissible exercise of its power to levy taxes. Chief Justice John Roberts wrote the majority opinion, joined by a coalition of four justices with whom he rarely aligned. The Court’s ruling paved the way for Obamacare to take effect, another step in the remaking of health care law in the United States.

Public Reactions to Supreme Court Decisions This book begins with a question that is deceiving in its simplicity: Do high-profile Supreme Court decisions shape public opinion? In recent decades, a great deal of attention has been paid to the persuasive power of the judiciary—specifically, its ability to increase support for controversial policies.2 This interest is driven in part by a curiosity about interplay between institutional legitimacy—the reservoir of durable goodwill that the U.S. Supreme Court, more than most political institutions on earth, enjoys (Woodson, Gibson, and Lodge 2011)—and policy support. An intriguing question—one with immense normative implications—underlies this research: Might the Court use its legitimacy to build consensus surrounding even the most controversial public policies? The potential for the judiciary to create popular consensus even as it exercises its constitutional responsibilities offers dazzling promise for an American polity plagued by polarization and gridlock.3

Imagine how this process might have played out in the case of national health care legislation. Prior to the Court’s ruling, polls showed sharp polarization over the ACA, with many Americans expressing strong approval for Page 3 →the law and even more evincing vehement opposition (Campbell and Persily 2012). These differences came through most clearly along partisan and ideological lines, as many conservatives doubted the legitimacy and constitutionality of the legislation. At the same time, Americans maintained a high degree of confidence in the Supreme Court. Four out of every five citizens affirm that the institution can usually be trusted to make decisions that are “right for the country as a whole” (Gibson, Caldeira, and Spence 2003b, 540). The justices, as the final arbiter of the health care dispute and the constitutionality of the ACA, were better equipped than perhaps any other set of actors in the federal government to lessen polarization. They had the ability to confer a symbolic stamp of institutional legitimacy—drawing on the Court’s reservoir of goodwill—on a contentious public policy while also offering a reasoned constitutional analysis to persuade the most attentive observers. A decision to uphold the signature legislative achievement of a Democratic president by a Court with a majority of Republican appointees would send a further signal that the institution was not swayed by partisan motives (as some critics had accused in 2000, after Bush v. Gore). If the judiciary possessed the ability to increase support and acceptance of Obamacare, the process might have transpired as follows: Policy ambivalence—Large numbers of Americans view the ACA as unconstitutional and unwise Institutional trust—Many of these Americans view the Court as the institution best equipped to make judgments about the constitutionality of legislation; many of these Americans trust the Court to make decisions that are in the best interests of the country as a whole Decision—The Court upholds the ACA as constitutional Awareness—The vast majority of Americans are aware of the decision that upholds the ACA Legitimation—The decision confers a symbolic stamp of legitimacy on the health care law, increasing the belief that the law is constitutional among its detractors and creating a consensus to this effect Policy agreement—The decision increases agreement with/support for the health care law (which, in turn, adds to its popular legitimacy [Mondak 1994]) This process—and the promise associated with it—have driven a great deal of research on the Court and public opinion over the past three decades. What could offer a better way forward for an American political system Page 4 →plagued by gridlock than an institutional actor to whom Americans turn for policy guidance and the resolution of constitutional disputes? Or so the story goes. But the theory that posits the Supreme Court as a consensus builder has regularly failed to live up to empirical reality. The majority of Supreme Court decisions pass with little popular notice, to be sure, but even when the institution weighs in on a high-profile controversy, its influence is circumscribed at many points. The Obamacare ruling provides a telling example: support for the ACA remained relatively unchanged even after the Court, led by Republican appointee John Roberts, deemed the legislation constitutional. This public response casts considerable doubt that the Court can increase support for the most controversial public policies. Even more troubling for our understanding of Court influence is the fact that a surprisingly large number of rulings inflame public sentiment to a degree that one would not expect given the trust that Americans place in the judiciary. For instance, a landmark 2005 decision, Kelo v. City of New London, portrayed by the Court as a limited extension of its property rights precedents, led to widespread indignation. Public support for government takings of private property, which the justices upheld in Kelo, appeared to decrease after the decision (Nadler, Diamond, and Patton 2008). Why might this be the case? Why does the Court, given its relatively robust institutional standing, not only fail to create consensus but in many cases spark outrage? The question lies at the heart of this book and motivates much of what follows. How can we best understand many instances in which a high-profile judicial decision fails to foster policy agreement (defined as an increase in popular support for a policy upheld by

the judiciary)? The question is not entirely novel. In a comprehensive study of public opinion and the Court, Thomas Marshall (1989) argued that rulings regularly failed to increase support for the policy positions endorsed by the justices. Moreover, decisions had little consistent effect on popular attitudes. More recently, Persily, Citrin, and Egan (2008) collected evidence that showed the uneven effects of Court influence across a wide range of constitutional disputes. Experimental results have told a different story, however, with a more consistent portrait of the judiciary as a consensus builder. The typical explanation for these divergent conclusions points to popular awareness of rulings as the Court’s main obstacle to building consensus. Most Americans pay little attention to rulings; it is therefore unsurprising that the judicial branch’s ability to persuade is much more circumscribed in the real world than it is in the laboratory. But even this explanation fails to offer an account for backlash (declines in support for the policies the Court upholds). So while questions about Court influence are not novel, the approachPage 5 → that this book takes to understand it is. It is an approach motivated by the puzzle of backlash and by a need for more careful attention to the factors that shape public responses to high-profile decisions.

A Link between the Court and the Public We start with a simple observation, albeit one that has important consequences for our understanding of the judiciary’s ability to build popular consensus: the link between the Court and the public is an indirect one. Few Americans take the time to read the written opinions released by the justices, and many fewer still have the opportunity to sit in the courthouse and hear the announcement of decisions. With cameras not permitted in the courtroom, and with the reluctance of justices to describe their perspectives on specific cases outside of the courtroom, most of the information available about landmark rulings originates from sources other than the Court. These sources include leading politicians and other political elites, legal experts and parties to cases, and issue activists with vested interests in outcomes. But the national news media bears primary responsibility for describing and interpreting the work of the justices and other sources for the public. Indeed, the vast majority of Americans learn about the decisions of the U.S. Supreme Court not by reading case law but by picking up newspapers, turning on their televisions, and monitoring the Internet. Does careful attention to the press inform our ability to understand the persuasive power of the institution? Or in other words, why should we care about the content of Supreme Court media coverage? The reasons are threefold. First, the vast majority of information Americans learn about Supreme Court decisions comes from the news media. Unlike a president, who occasionally speaks directly to the American people with primetime television addresses, campaign events and advertisements, and a State of the Union message, Supreme Court justices maintain low public profiles. And though justices are increasingly “going public” to fulfill personal objectives (Davis 2011), the Court itself offers no formal justifications for its decisions beyond those found in its written opinions. Furthermore, these written opinions, which represent the public’s only direct access to rulings, rely heavily on legal analysis. Most Americans lack the time, motivation, and legal expertise required to make sense of the Court’s decisions. Instead, people look to major media outlets, which aim to interpret decisions in an accurate, simple, and entertaining manner. Second, some evidence indicates that the content of media coverage alters attitudes toward Court decisions. In a carefullyPage 6 → designed experiment, Clawson and Waltenburg (2003) show that media effects influenced attitudes toward the Court’s ruling in Adarand v. Pena: framing the affirmative action decision as reinforcing the principle of equal treatment increases support for it among white respondents, while framing it as a dramatic setback for African Americans does not. The results suggest that the persuasiveness of a Court endorsement depends on its media contextualization. And third, there are reasons to expect experimental treatments and actual media coverage of Court rulings to differ systematically. Unlike laboratory stimuli, real-world press coverage of judicial decisions focuses on a narrow range of newsworthy controversies; offers detail, complexity, and multiple perspectives; and includes abundant criticism of the Court. Thus, for ordinary Americans, the law of the land receives meaning as much through the publications of journalists as it does through the pens of the justices. Take the Obamacare ruling as an example. While the decision, written by Chief Justice Roberts, might well have assuaged the doubts of the many Americans who questioned the ACA’s constitutionality, other pressures

robbed the decision of its ability to create consensus. Among the most important of these factors were the strong preexisting attitudes held by many Democrats and Republicans about Obamacare. But fully understanding responses to the decision also requires considering how the public interpreted its meaning and impact. What judgments did the Court make, and how did they affect the fate of Obamacare? What decision-making procedures did the institution employ to arrive at its ruling? How did each justice reach his or her conclusions, and how did the justices come together in coalitions to render the decision? What standards did the Court use to evaluate the ACA’s constitutionality? What standards did the dissenting justices use? Was the ruling situated in a larger political context, replete with references to the reactions of Democrats and Republicans? Or was it a carefully circumscribed legal analysis offered by the judiciary? A few common themes animated coverage of the Obamacare case. The press highlighted the Court’s finding that the ACA represented little more than Congress’s exercise of its power to tax.4 A Los Angeles Times headline read, “Supreme Court Upholds Healthcare Law as Tax Measure.” CNN simplified the ruling with the headline, “Supreme Court: Mandate Penalty Is a Tax.”5 Many accounts explained in considerable detail Roberts’s reasoning, which rejected the argument that the law was unconstitutional in compelling individuals to purchase a commodity—health care insurance—against their will. The media framed the decision as envisioning a more limited role for Congress: to levy a small tax on those who did not purchase health care, without making their actions criminal. The press also noted the four dissentingPage 7 → justices’ vehement objections to the ruling.6 Accounts made clear that, given the opportunity, these justices would have struck down the entirety of the ACA. The press highlighted the dissenters’ finding that Congress intended to regulate commerce in health care, which was beyond the scope of its constitutional power (and on this point, Roberts agreed). And finally, the press highlighted the surprise outcome, which saw the reliably conservative Roberts break with the other conservatives on the Court to uphold the law. According to the New York Times, the case featured the first vote to unite Roberts, Ruth Bader Ginsburg, Stephen Breyer, Sonia Sotomayor, and Elena Kagan against their four conservative colleagues. Some reports even suggested that Roberts changed his vote, deciding only in the days before the decision to uphold Obamacare.7 Each distinct frame that the media employs to characterize judicial decisions—using the language of the Court majority, explaining the objections of dissenters, depicting the institution as nonpartisan—may place unique pressures on popular opinion, amplifying or diminishing the judiciary’s power to persuade. These characterizations of Supreme Court rulings are among the most important ingredients that shape popular understanding of the law. But scholars have little sense about the content of coverage afforded landmark Supreme Court decisions, the factors that shape this content, and its effects on public opinion. In fact, in many studies of political behavior and the judiciary, the press is all but invisible. And when experimental research acknowledges that the press may matter, it commonly does so by presenting brief news reports to subjects with the intention of communicating simply that the Court has ruled on an issue. Media reports are rarely this straightforward. But of the shape and impact of such coverage, we have little idea.

Overview and Argument This book is motivated by the puzzle of how to understand the panoply of reactions that have followed landmark Supreme Court rulings in recent decades. To do so, it offers a portrait of the media coverage through which the most Americans learn about and make sense of case law. The research follows in the tradition of sociolegal studies, which commonly trace the stature of specific laws in American society, as well as a new line of public opinion research that explores the effects of judicial decisions on attitudes. The book’s approach differs from other lines of research, however, by exploring in detail the nature of media coverage afforded landmark decisions. It is one of the Page 8 →first studies to link Supreme Court decisions, press coverage, and public opinion. By taking seriously the role played by the media in this process, it offers new insights, with implications for a diverse range of fields, including studies of political communication, accounts of public opinion and the judiciary, and models of judicial voting behavior. The research presented in the following pages seeks to answer two primary questions. First, what is the nature of press coverage afforded the Court, and through what mechanisms do cross-ruling variations in coverage come

about? Second, what role does, if any, does press coverage play in shaping approval for the policies the Court upholds? The main argument of the book is that the depiction of Supreme Court rulings varies in response to voting signals sent by the justices themselves; because such coverage offers complex and often critical portraits of rulings, the Court faces limits on its ability to increase popular support for the policies it upholds. These limits do not necessitate a loss of institutional reputation, as the Court’s legitimacy remains relatively stable over time. Rather, such legitimacy functions as a weak persuasive currency in debates over the wisdom of Court-endorsed policies. Chapter 2 highlights gaps between theory and empirical reality in the study of judicial persuasion. It chronicles the numerous limits on the judiciary’s ability to persuade but makes the case that one of the most powerful limits—the depictions of Supreme Court decisions in the national news media—has received very little attention. Part 1 focuses on the central role the press plays in interpreting and defining Supreme Court decisions for most Americans. Chapter 3 describes the model of dissensus dynamics, which suggests that the content of media reports is determined by the characteristics of judicial decisions, most notably the voting outcomes on the Supreme Court. Chapter 4 provides the first test of dissensus dynamics, using a case study of two takings law decisions: Lingle v. Chevron, Inc. (2005) and Kelo v. City of New London (2005). The chapter uses a most-similar design, demonstrating that despite the substantive similarities in the rulings, differences in their voting outcomes lead to distinct media coverage of them. The cases provide evidence that press coverage follows not only from votes but also from the arguments raised by the majority and dissenting coalitions in their written opinions. Chapter 5 offers a more general test of dissensus dynamics, looking at the way in which national newspapers report on important Supreme Court decisions. It provides evidence that the press portrays dissensual rulings in more unfavorable terms than consensual ones. It also shows that press coverage responds to the makeup of the Court’s voting coalitions, with dissensual and ideologically divided rulings receiving the most negative coverage. Chapter 6 offers a further test of dissensus dynamics,Page 9 → focusing on cable news reports. Much like newspapers, cable news outlets respond to votes on the Supreme Court. In particular, dissensual decisions engender incivility and negativity in television news reports. Part 2 applies the findings from part 1 to explore public reactions to Supreme Court decisions. It details the central role that media coverage plays in shaping public responses, highlighting new insights we may glean when we view the press as the lens through which Americans come to understand judicial decisions. These chapters pay particular attention to the microlevel mechanisms at work, tracing how the currency of institutional loyalty loses its consensus-building value when the press criticizes rulings (as it so often does). Chapter 7 introduces a framework for evaluating the influence of media coverage on popular reactions to Supreme Court decisions. The framework—built on research in public opinion and political psychology—explores the effects of frames and cues and the countervailing pressures that arise when a trusted institution like the Court endorses a policy only to have the media frame the decision as activist, unwise, or reckless. Though Americans commonly view the judiciary as trustworthy and unbiased, new evidence shows that the Court loses persuasive currency on these occasions. Two original experiments demonstrate that even small changes in news coverage have important effects on the popular responses to decisions. The experiments also provide evidence that Court decisions do indeed serve a cuing function, altering public opinion in a positive direction, as well as that media coverage of these decisions affects the relative level of support for them. Chapter 9 employs a powerful research design to study reactions to the Obamacare decision. The panel study allows me to trace individual-level attitude change during the summer of 2012. The findings offer a unique insight: the Court was unable to draw on its reservoir of goodwill to persuade; rather, Americans used their attitudes toward the health care law to guide their support for the institution in the aftermath of the decision. Throughout the book, the need for a careful examination of how the press covers the Court becomes clear. By examining popular reactions to the type of critical coverage that regularly follows dissensual decisions, the book reshapes our understanding of the judiciary’s power to persuade. It presents a portrait of a Supreme Court whose ability to legitimate policies is, contrary to conventional wisdom, only indirect, mediated through the often critical lens of the national press. Though the Court indeed maintains a relatively durable reservoir of trust, this

legitimacy supplies it only a limited, indirect tool for fostering popular consensus.

Page 10 →

Chapter 2 Institutional Legitimacy, Public Reactions, and the Backlash Puzzle To understand popular responses to high-profile Supreme Court rulings, consider first the unique standing of the institution. Americans view the Court as inherently legitimate—trustworthy, unbiased, and worthy of protection from politicized interference. When compared with other institutions in American politics and other political institutions worldwide, the Court enjoys significant popular support. This fact leads to a question: Might the institution, as a credible source with a well of diffuse support, use its decisions to increase public approval of controversial policies (Mondak 1990, 1994)? This question hints at intriguing possibilities for consensus building in a polarized society. One landmark study involving Roe v. Wade, the Court’s controversial decision that legalized abortion, uses the image of the institution as a “republican schoolmaster,” exploring whether it can instruct the public in a manner that is “gentle but effective” (Franklin and Kosaki 1989, 781). Others consider the possibility that Court decisions increase policy compliance (Johnson 1967), policy acceptance (Gibson, Caldeira, and Spence 2005; Woodson, Gibson, and Lodge 2011), and popular approval of controversial positions. Given the Supreme Court’s standing in American politics, many view the bench as having a unique ability to resolve political disagreements. For these reasons, it is not unexpected that no theory of public response can account for one type of reaction to Supreme Court decisions: backlash—a short-term decrease in support for a policy after the Court endorses it. As puzzling as it may seem, backlash occurs with surprising frequency. Consider the many instances in which public support for a policy has vanished following the Court’s endorsement. Lerman (2008) demonstrates that a series of Court rulingsPage 11 → (beginning with Miranda v. Arizona) had a negative influence on support for the rights of the accused. Similarly, public support for capital punishment increased after the Court struck down the death penalty as unconstitutional in Furman v. Georgia (Hanley 2008). A series of rulings that upheld affirmative action programs did little to stem popular disapproval for race-based preferences (Le and Citrin 2008). Court decisions that gave legal status to same-sex partners and overturned sodomy laws halted increases in popular support for gay rights (Egan, Persily, and Wallsten 2008). Texas v. Johnson, which upheld First Amendment protections for flag burning, inflamed popular sentiment (Hansen 2008). And the Supreme Court’s ruling in Kelo v. City of New London, a case dealing with property rights, led to public backlash because Americans viewed its interpretation of eminent domain as a violation of property rights (Baron 2007; Nadler, Diamond, and Patton 2008).1 Existing accounts of legitimation do not anticipate negative opinion responses to high-profile Supreme Court rulings. Indeed, there are few reasons to expect low levels of support for the policies endorsed by a legitimate institution like the Court. Why might the endorsement of a policy from a high-credibility institution dampen support for that policy? Scholars have suggested but never tested two possible explanations. Egan and Citrin (2009) speculate that backlash arises from one-sided elite debate following a judicial ruling, where supportive elites remain silent while those in opposition control the public discourse. Franklin and Kosaki (1989) suggest that when the public uniformly opposes a policy before the Court endorses it, discussions of the policy can become even more critical after a ruling is released. Nonetheless, the prevalence of negative responses to high-profile judicial rulings presents a puzzle worthy of further attention.

Institutional Legitimacy and Policy Agreement Americans hold the Supreme Court in relatively high esteem compared with the other branches of government, a fact that offers the (unrealized) promise that the institution might serve as a consensus builder. Favorable attitudes toward the Supreme Court arise out of perceptions of procedural fairness, which people value in legal disputes (Tyler 1990, 2006; Baird and Gangl 2006). People express steady support for courts when judges employ unbiased decision-making procedures. By and large, Americans believe the Supreme Court uses principled decision making to resolve disputes, making it one of the most legitimate judicial institutions in the world (Woodson, Gibson, and

Lodge 2011). Page 12 →Popular views about courts take shape early in life, as people learn about the roles and responsibilities of judges in the American court system. According to Federalist No. 78, “There is no position which depends on clearer principles.В .В .В . To avoid an arbitrary discretion in the courts, it is indispensable that they should be bound down by strict rules and precedents, which serve to define and point out their duty in every case that comes before them.” Judges implement these principles with the goal of ensuring public deference and compliance (Davis 1994). The Supreme Court carefully manages its public image by highlighting the majesty and dignity of judicial proceedings. Depictions of the Court rely on symbolic language and imagery. Among the most important visual symbols are pictures of judges in robes, gavels, the Lady Justice, and the Supreme Court building itself, which help to emphasize the distinctiveness of the institution (Brigham 1987; Woodson, Gibson, and Lodge 2011). The justices use language to emphasize the unbiased application of law; indeed, even some of the most notorious Supreme Court decisions frame their outcomes as matters of legal principle. In Plessy v. Ferguson (1896), which upheld segregation by race, the Court begins with an innocuous characterization of its work: “This case turns upon the constitutionality of an act of the General Assembly of the State of Louisiana.” The Court then concludes with this judgment: “If one race be inferior to the other socially, the Constitution of the United States cannot put them upon the same plane.” Similarly, the Court emphasized that its decision in the Obamacare case was a legal one, not a political one. At times, the justices seek unanimity to further legitimize rulings, though this practice has declined in recent decades (Epstein, Segal, and Spaeth 2001). John Roberts, in his Senate confirmation hearings, reinforced the image of justices as principled decision makers, arguing, “Judges are like umpires. Umpires don’t make the rules; they apply them.”2 The ideal of mechanical jurisprudence—that judges apply simple legal principles in an unbiased fashion to arrive at the correct decision—is, of course, a myth. The attitudinal model of judicial decision making sees rulings as influenced by the personal preferences of judges (Segal 1997; Segal and Spaeth 2002). Others scholars contend that judges act strategically to achieve their most preferred policy outcomes (Harvey and Friedman 2006; Clark 2009). And recent research suggests that both ideological and legal principles shape court rulings (Bailey and Maltzman 2008). But the American belief in procedural fairness is not simplistic. People believe judges can exercise discretion yet adhere to legal principles in their decision making (Gibson and Caldeira 2011). Most Americans agree that judges use discretion to apply the Constitution to modern problems and that their personal Page 13 →beliefs play a role as they do so. Yet Americans tend to view judicial actions as principled, if not mechanical. What might undercut the symbolism and perceptions of procedural fairness integral to the popular legitimacy of the Supreme Court? Three of the most likely explanations fall short. There is little evidence that sophisticated observers of politics view the Court as ignoring legal principles in its decision making. In fact, political awareness increases support for the Supreme Court as a consequence of sophisticates’ exposure to legitimating symbols like the black robes worn by the justices, the Lady Justice, the scales of justice, and the Supreme Court building itself (Caldeira and Gibson 1992; Woodson, Gibson, and Lodge 2011). Similarly, some have suggested that the spectacle of confirmation hearings can undermine the image cultivated by the Court (Davis 1994); however, more recent evidence suggests that exposure to them increases the institution’s legitimacy under certain conditions (Gibson and Caldeira 2009a, 2009b). Finally, there is conflicting evidence as to whether the Court damages its legitimacy by becoming enmeshed in political controversies and releasing unpopular decisions. Grosskopf and Mondak (1998) suggest that Texas v. Johnson, which preserved the constitutional right to burn the flag, substantially damaged confidence in the Court, the result of a negativity bias whereby people assign asymmetric penalties to the institution depending on their views of its decisions. But other evidence suggests that positivity bias exists: the effect of exposure to any news about the Court means that popular decisions increase diffuse support for the institution, but similar decreases in legitimacy do not follow unpopular rulings (Gibson, Caldeira, and Baird 1998; Gibson and Caldeira 2009a). Gibson, Caldeira, and Spence (2003b) show that in one of the most controversial and politicized decisions of all time—Bush v. Gore—agreement with the outcome had little impact on how people viewed the Supreme Court. They argue that the Court may damage its legitimacy only by

issuing a series of high-profile and unpopular decisions in succession. Diffuse support for the courts in America, then, arises out of the popular understanding of judicial responsibilities and the belief that judges exhibit a commitment to procedural fairness. These beliefs are reinforced by the language, symbolism, and imagery surrounding the courts. Americans view the Supreme Court favorably compared to other institutions, and these attitudes remain quite durable. Easton (1965) defines legitimacy (sometimes known as diffuse support) as “a reservoir of favorable attitudes or good will that helps members to accept or tolerate outputs to which they are opposed or the effects of which they see as damaging to their wants” (273). As such, the concept involves a Page 14 →more stable set of attitudes than those measured with items about “support” for or “approval” of an institution. Other research characterizes legitimacy as loyalty toward an institution, since short-term outcomes have little effect on long-standing support for it (Gibson, Caldeira, and Baird 1998; Gibson, Caldeira, and Spence 2005). When the Court releases decisions, its legitimacy may also be considered a form of source credibility that enables it to influence popular attitudes (Mondak 1990, 1992, 1994; Bartels and Mutz 2009; Egan and Citrin 2009). Scholars employ a number of measures to evaluate institutional legitimacy, but the standard is a six- (or sometimes four-) item scale developed by Gibson and colleagues. The institutional legitimacy scale is specific to the Supreme Court, with items about its proper jurisdiction, the fairness of its decision making, and the decisionmaking procedures it employs. They include: If the U.S. Supreme Court started making a lot of decisions that most people disagree with, it might be better to do away with the Supreme Court altogether. The right of the Supreme Court to decide certain types of controversial issues should be reduced. The Supreme Court can usually be trusted to make decisions that are right for the country as a whole. The decisions of the U.S. Supreme Court favor some groups more than others. The U.S. Supreme Court gets too mixed up in politics. The U.S. Supreme Court should have the right to say what the Constitution means, even when the majority of people disagree with the Court’s decision. Because the Supreme Court lacks the power to enforce decisions, it relies on its legitimacy to ensure their implementation. And in a number of political controversies, judicial legitimacy has proven essential. For example, despite the unhappiness that many Americans evinced with decisions released by the Court in the 1930s, Franklin Roosevelt’s plan to add justices was rejected as a politicized attack on the judiciary (Caldeira 1987; Shesol 2010). Bush v. Gore settled a political and constitutional controversy with little long-term cost to the Court’s reputation (Gibson, Caldeira, and Spence 2003b). The legitimacy of the institution remains relatively high. Judicial legitimacy matters not just because it serves to safeguard Court jurisdiction and ensure the implementation of decisions but also because it Page 15 →holds promise as a tool through which the Court may sway public opinion and foster national consensus. According to the policy agreement hypothesis, Supreme Court decisions persuade the public to support the specific policies the decisions endorse. The Court, by issuing a decision that upholds a controversial policy as constitutional, confers the legitimacy of the institution on the policy itself. Attitudes toward the policy become more favorable along a number of dimensions: people express higher levels of acceptance for the policy, greater willingness to comply with it, and, most important, more support for it. I define policy agreement as a short-term increase in support, along a variety of dimensions, for a policy endorsed by the Supreme Court. This increase in support occurs as a result of symbolic legitimation—the association between the institution and the policy—and persuasive legitimation (Mondak 1994; Bartels and Mutz 2009). American support for Court-endorsed policies, then, is expected to arise out of the links among perceptions of procedural fairness, the institution itself, and its decisions.

But while the theory of policy agreement is well developed, testing it proves difficult. Scholars have yet to agree on how to define the Court’s ability to build consensus. Does it consist of acquiescence to decisions? Does it alter compliance with laws? Does it alter support for the policy as well? Each definition has distinct consequences for our understanding of Court persuasiveness (Woodson, Gibson, and Lodge 2011). Another difficulty is practical: tests of opinion change demand pre- and postdecision data on public opinion. But the erratic timing of rulings makes such data scarce (Marshall 1989). For these reasons, experimental settings saturate the study of policy agreement, giving scholars control over measurement intervals and information environment. A number of experimental studies provide some evidence of agreement and legitimation effects. Scholars demonstrate the Court’s ability to increase policy support in cases involving censorship (Mondak 1992, 1994), search and seizure (Mondak 1992, 1994), educational policy (Mondak 1994), abortion (Zink, Spriggs, and Scott 2009), school prayer (Zink, Spriggs, and Scott 2009), affirmative action (Bartels and Mutz 2009), flag burning (Bartels and Mutz 2009), bankruptcy law (Zink, Spriggs, and Scott 2009), taxation (Mondak 1992), campaign law (Mondak 1992), electoral disputes (Gibson, Caldeira, and Spence 2005), and funding for the arts (Hoekstra 1995). Some survey data provide additional evidence of agreement and legitimation effects. For example, Franklin and Kosaki’s landmark study (1989) shows that Roe v. Wade helped to legitimate abortions for health-related purposes, even as many Americans expressed deep reservations about the decision. In other cases, policy support develops over a long period of time, as the publicPage 16 → gradually becomes more comfortable with the consequences of rulings (Egan, Persily, and Wallsten 2008; Murakami 2008). By and large, though, support for the policy legitimation hypothesis comes from laboratory studies that facilitate control of source, message, and measurement.

Limits on Policy Agreement In spite of evidence that demonstrates Supreme Court persuasiveness, “no single hypothesis provides a universally applicable explanation for how and when courts affect public opinion” (Persily 2008, 8). Popular reactions to Court rulings vary considerably. In many cases, the public does not express immediate approval of rulings but rather reacts with indifference. Over time, the public may express more support for controversial decisions, as it did in the case of Brown v. Board of Education (Murakami 2008). Other decisions lead to divergent responses from various segments of the public. And most puzzlingly, certain decisions inflame public sentiment and precipitate backlash. The Supreme Court’s power to persuade is moderated by a range of factors. Salience The most significant factor that shapes the public response to judicial rulings is their salience. The vast majority of Supreme Court decisions receive minimal attention from the press. From 1946 to 1995, only about 15 percent of all rulings received a front-page mention in the New York Times, despite the fact that the paper offers some of the most comprehensive coverage of the institution (Epstein and Segal 2000). These decisions are among the only ones to reach the less attentive portion of the public (Berkson 1978) and, as such, the only ones capable of generating widespread policy agreement (Murphy and Tanenhaus 1968). A number of Court rulings fail to persuade the public because most people remain unaware of them (Marshall 1989; Franklin and Kosaki 1995). And even when Americans become aware of rulings, they must accurately interpret the outcomes for legitimation to occur (Murphy and Tanenhaus 1968). Because salience plays an integral role in shaping public responses, experimental studies struggle to estimate accurately the prevalence and extent of policy agreement. Most of the evidence in favor of the policy agreement hypothesis inflates the salience of Court actions by tasking subjects with reading about decisions (Mondak 1990; Hoekstra 1995; Bartels and Mutz Page 17 →2009; Zink, Spriggs, and Scott 2009). In actuality, the majority of judicial decisions pass with little notice from average Americans, stripping the Court of its power to persuade. Preexisting Opinion and Interpretive Context When Americans learn about Court rulings, preexisting opinions further moderate their reactions. The Court may

do little to persuade strong opponents of a policy, though the institution may increase policy acceptance and acquiescence (Gibson, Caldeira, and Spence 2005). Consider the case of Bush v. Gore, which did little to increase support for George W. Bush’s Electoral College victory among Democrats but helped to legitimize his election, effectively ending the legal controversy (Gibson, Caldeira, and Spence 2003b). The Court similarly lacks the power to legitimate policies when facing a polarized public. Franklin and Kosaki (1989) demonstrate how the interpretative context through which people come to understand rulings—which takes shape in response to discussions with others—alters opinion responses. For example, because Catholics largely disapprove of discretionary abortions, Catholic opinion about Roe v. Wade developed in reaction to this social environment, causing individual reactions to move toward the group position (and away from the Court). The authors offer the structural response hypothesis, which suggests the existence of multiple opinion reactions to any Court ruling. According to the authors, “Members of microenvironments are positively affected by the group norms. Thus the effect of group interaction is to increase agreement with the modal response within the immediate social environment. When between-group variance is high, we would expect group members to move in opposite directions, leading to polarization. When between-group variance is low, then more uniform shifts in the population are the likely result” (Franklin and Kosaki 1989, 763; see also Stoutenborough, Haider-Markel, and Allen 2006). The structural response hypothesis helps to account for divergent reactions to judicial decisions across not only microenvironments but also partisan contexts and media environments. Sequence The order in which the Court releases rulings on a given issue impacts public responses as well. Johnson and Martin’s (1998) conditional response hypothesisPage 18 → posits that only initial rulings on salient political controversies have the power to engender attitude change. Because these rulings help individuals to fully elaborate relevant attitudes, their opinions remain stable even after subsequent Court rulings (Chaiken 1980; Petty and Cacioppo 1981). More recent research draws on the receive-accept-sample model (Zaller 1992; Zaller and Feldman 1992) to refine the conditional response hypothesis (Brickman and Peterson 2006). Decision Characteristics Finally, research suggests that a variety of decision-specific factors affect public opinion about judicial rulings. In accordance with their demand for procedural fairness, Americans view rulings more favorably when the justices eschew political bargaining and uphold existing law (Caldeira 1986; Baird and Gangl 2006). The public expresses more approval for rulings when the Court makes strong persuasive arguments (Mondak 1990, 1994). People are more likely to approve of unanimous decisions that uphold precedent (Zink, Spriggs, and Scott 2009). And the connection of rulings to legal principles similarly affects the likelihood of popular acceptance and acquiescence (Gibson, Caldeira, and Spence 2005). The idiosyncratic characteristics of decisions also affect their public reception. No simple model of opinion response fully characterizes the reactions to Roe v. Wade, where people evaluated not only the policy outcome but also the facts of the case, the identities of the parties, the identities of the justices themselves, and the arguments they raised. The majority’s opinion, for example, which protected abortion under the right to privacy found in the Constitution’s “penumbras and emanations,” has come under attack, undermining the persuasiveness of the decision (Sirico 2010). Might Americans have expressed more support for a decision framed in different terms? And because many cases that come before the Court are the result of concerted efforts to win public and judicial support through the strategic use of claimants, popular approval of rulings may increase when the Court sides with sympathetic parties (Nadler, Diamond, and Patton 2008; Carpenter 2012; Carrubba et al. 2012). All told, these findings demonstrate the limits of policy agreement. The Supreme Court’s ability to persuade is constrained, first and foremost, by the extent to which Americans are aware of its decisions. But even in the most high-profile cases, preexisting attitudes, interpretive lenses, and decision characteristics moderate the influence of rulings on public opinion. The Page 19 →Court’s legitimacy, it seems, confers it with only a

circumscribed power to persuade.

Improving Models of Supreme Court Influence The Supreme Court, which Americans value for its commitment to procedural fairness, maintains a measure of diffuse support yet only sometimes increases support for policies. What accounts for the literature’s divergent portraits of Court persuasiveness? First, issues of measurement have proven problematic. Scholars define policy agreement differently across studies. In some cases, they look at the Supreme Court’s ability to increase acceptance and acquiescence (Gibson, Caldeira, and Spence 2005). At other times they focus on compliance with decisions (Johnson 1967). Many more studies examine support for policies that Supreme Court endorses (Franklin and Kosaki 1989; Hoekstra 1995; Johnson and Martin 1998; Bartels and Mutz 2009). When these measures are unavailable, scholars rely on items that focus on the extent to which Americans agree with specific decisions (which eliminates the possibility of measuring predecision attitudes). On its own, none of these measures is objectionable, and each taps into an important dimension of public reactions that is worthy of attention. But understanding the persuasive power of the Court depends, in large part, on defining the standards of persuasion—something that has not always been done. As such, one must take care when comparing Court persuasiveness across studies that employ different measures of agreement and legitimation.3 For the purposes of clarity, this book focuses on the institution’s ability to alter support for and approval of specific public policies. Furthermore, a full test of the policy agreement hypothesis demands pre- and postdecision measurements of opinion, which are frequently unavailable. Scholars deal with data shortcomings in a number of ways. When only postdecision surveys are available, some studies explore the general tenor of attitudes on an issue to infer whether people view rulings in favorable terms (e.g., Nadler, Diamond, and Patton 2008). Franklin and Kosaki (1989) use another approach, controlling for the effects of rulings by dividing survey respondents into groups that are aware of the decision in Roe v. Wade (the treatment group) and those that are not (the control group). This technique allows them to make inferences about the impact of the ruling, but only across groups that have different levels of political awareness. They cannot estimate the effect of Roe v. Wade on American opinion as a whole. When Page 20 →pre- and postdecision opinion data are available, scholars explore how the Court may influence attitude change (e.g., Marshall 1989). The shortcomings of this approach lie both in lack of control (since any event that transpires between the two surveys can cause attitude shifts) and lack of uniformity (since there exist neither standard nor acceptable intervals at which to measure opinion surrounding a ruling). Most studies, however, deal with measurement issues by using an experimental approach, which allows control over the pre- and postdecision measures of opinion. But the use of laboratory settings contributes to another major problem plaguing the study of Court persuasiveness: a troubling divide in conclusions that is driven by methodology. According to one study, “With few exceptions, experimental work finds much stronger effects of exposure to Supreme Court decisions than do observational studies” (Egan and Citrin, 2009, 7; see also Clawson, Kegler, and Waltenburg 2001). Most evidence for the policy agreement hypothesis comes from laboratory studies that maximize control (but see Baas and Thomas 1984), while most evidence that finds constraints on the Court’s ability to persuade comes from survey data that emphasize external validity. Little work aims to reconcile the divide between these methods (see table 2.1). Nonetheless, we may attribute experimental evidence in favor of policy agreement to two factors. First, experimental studies artificially raise awareness of Supreme Court decisions. Most Americans express little interest in and knowledge about the vast majority of Court rulings (Marshall 1989), but experimental settings task subjects with reading about decisions nonetheless. Scholars have seldom sought more externally valid approaches such as embedding ruling descriptions in a larger news environment (but see Zink, Spriggs, and Scott 2009); as such, experiments depict policy agreement as prevalent. By tasking subjects with reading directly about rulings, experimental studies also obscure important differences in salience across issue domains. The news media afford the most coverage to rulings involving the First Amendment, civil rights, and the right to privacy (Davis 1987; Blake and Hacker 2010); experimental studies, conversely, explore the Court’s persuasiveness across a range

of issues—including search and seizure, educational policy, taxation, election law, bankruptcy law, and funding for the arts (Mondak 1992, 1994; Hoekstra 1995; Zink, Spriggs, and Scott 2009). Evidence of Court persuasion in these cases provides little information about legitimation in cases where public attitudes have crystallized. More important and with few exceptions, literature on policy agreement neglects entirely the role the media plays in translating decisions to the public. In experimental settings, subjects typically read a brief description of a ruling Page 21 →or a stylized news report on it. These materials inform subjects about the content of a decision but do little to convey the complexity of information and diversity of perspective that commonly characterize press coverage. A handful of studies on media coverage of the Court focus on the prevalence and accuracy of coverage but make few links between the press and popular reactions to specific rulings (Newland 1964; Slotnick 1993; Davis 1994, 2011; Slotnick and Segal 1998; Brickman and Bragg 2007). The few studies that do aim to connect justices, journalists, and public opinion find that not only the substance of a judicial decision influences popular attitudes but also the way in which it is framed (Mondak 1994; Allen and Haider-Markel 2006; Baird and Gangl 2006). But because most existing theories of policy agreement ignore press coverage,Page 22 → their conclusions diverge. Studies that verify Supreme Court influence fail to consider how unfavorable media coverage might circumscribe legitimation. Studies that find evidence of backlash struggle to offer a systematic account of its causes, instead relying on case-specific explanations. At various times, this work suggests that backlash may arise from the violation of sacred American values (Hanson 2008), the Court’s failure to engage with popular sentiment in its opinions (Baron 2007), or the identities of particularly sympathetic victims (Nadler, Diamond, and Patton 2008). Table 2.1. Literature on Supreme Court Persuasion Study Method Court Influence Baas and Thomas (1984) Experimental Null effects Mondak (1990) Experimental Legitimation Mondak (1992) Experimental Legitimation Mondak (1994) Experimental Legitimation Hoekstra (1995) Experimental Legitimation Baird and Gangl (2006) Experimental Constrained legitimation Zink, Spriggs, and Scott (2009) Experimental Constrained legitimation Bartels and Mutz (2009) Survey experiment Legitimation Franklin and Kosaki (1989) Survey Constrained legitimation Marshall (1989) Survey Null effects Rosenberg (1995) Survey Null effects Johnson and Martin (1998) Survey Brickman and Peterson (2006) Survey Stoutenborough, Haider-Markel, Allen (2006) Survey Murakami (2008) Survey Lerman (2008) Survey Gash and Gonzales (2008) Survey Luks and Salamone (2008) Survey Hanley (2008) Survey Mayeri et al. (2008) Survey Le and Citrin (2008) Survey Hanson (2008) Survey

Constrained legitimation Constrained legitimation Constrained legitimation Null effects Backlash Backlash Null effects Backlash Null effects Backlash Backlash

Mullin (2008)

Survey

Null effects

Egan, Persily, and Wallsten (2008) Green and Jarvis (2008) Nadler, Diamond, and Patton (2008) Goux, Egan, and Citrin (2008)

Survey Survey Survey Survey

Backlash Null effects Backlash Null effects

Mate and Wright (2008)

Survey

Constrained legitimation

The Approach of This Study This book employs a multiple-method approach to explore popular understanding of judicial decisions. Part 1 focuses on the media in detail. What determines how the press chooses to cover decisions? How do reporters frame rulings? What consequences might these choices have for the Court and for our understanding of policy agreement? By bringing the press back into the study of the Supreme Court decisions, we can better understand the divergent responses its decisions seem to engender. We also gain insight into the ways in which the existing literature fails to characterize adequately the process that leads to popular responses. These responses depend not only on the symbolic association of the institution with a policy—moderated by preexisting attitudes, interpretative contexts, and decision features—but also on the information communicated to the public about these decisions. As the media environment becomes more favorable toward the Court—highlighting the basis of its legitimacy, emphasizing its adherence to law and commitment to procedural fairness, explaining its decision-making rationale, and giving voice to supporters—its decisions become more likely to legitimate policies. The coming pages present an account of how the press covers the Supreme Court. They offer an explanation of how decision specific coverage takes shape across a wide range of cases in response to voting outcomes on the Court.

Page 23 →

Part One The Court and the Press

Page 24 → Page 25 →

Chapter 3 The Dissensus Dynamics Model How the National News Media Cover Supreme Court Decisions For news organizations preparing to cover the Supreme Court’s health care ruling, the morning of June 28, 2012, brought a number of challenges. With popular interest in the case at a crescendo, journalists faced a demand for immediate and accurate reporting about the outcome. Did the law survive? In what form? And on what basis? From a reporter’s perspective, the Court would not be particularly helpful in answering these questions quickly and accurately. The justices would release their written opinions—majority, concurrence, and dissent, 193 pages in total—all at once. They would make little effort to simplify or summarize their findings. Indeed, the shifting rationale contained in the majority opinion would prove so complex that two cable news networks would report erroneously that the Court had struck down the ACA as unconstitutional. Media coverage of the U.S. Supreme Court takes shape in a constraint-laden environment. News organizations focus on a handful of high-profile rulings each year, aiming to draw public interest while also providing simple yet accurate information about legal affairs. These demands challenge journalists who cover the Court, all the more so because the institution does little to makes its actions transparent. As a result, the press relies on the voting signals sent by the justices to guide its coverage of rulings. The following pages describe a model of Supreme Court journalism based on the observation that constraints on the Court beat and in the newsroom lead to distinctions in decision coverage depending on the size and makeup of majority voting coalitions. This model is designed to offer new insight Page 26 →into the content of coverage afforded high-profile Supreme Court decisions, about which scholars know very little. The model extends our understanding of elite-driven accounts of political communication while demonstrating important differences in how the press reports on the Court versus other political actors.

On the Supreme Court Beat The Supreme Court choreographs a complex strategy for dealing with the press. The Court carefully guards its legitimacy—the relatively high levels of diffuse support Americans express for it—by emphasizing the legal and symbolic basis of judicial authority (Johnson 1967; Caldeira and Gibson 1992; Woodson, Gibson, and Lodge 2011). The Court highlights its majesty and dignity, which it conveys “through the nature of the physical setting, the usage of ritual, and the style of communication” (Davis 1994, 12; see also Brigham 1987). But the Court eschews more aggressive public relations strategies.1 It makes minimal effort to simplify or frame rulings for public discussion. It offers little transparency in its decision making. It has long resisted cameras in the courtroom. Perhaps this is for the best, as Americans may recoil at the behavior of the justices (their visceral disagreements with one another, for example, or the sight of Justice Clarence Thomas napping during oral arguments). So the Court pays at best minimal attention to media demands. According to Linda Greenhouse, who spent two decades covering the high court for the New York Times, the institution “is quite blithely oblivious to the needs of those who convey its work to the outside world” (Greenhouse 1996, 1559). In fact, a review of the procedures used by the Supreme Court makes clear how little concern the institution has about packaging its work for media consumption. Each year, the Court receives more than 8,000 case petitions, yet it only decides to hear about 80 of them per term (Carpenter 2012). To hear a case, at least four justices must vote to issue a writ of certiorari, though these votes remain private. Thus, the Court simply neglects to grant certiorari without explanation in the vast majority of cases it considers. The justices typically set aside a portion of one day to hear oral arguments in each case on the docket. At any point thereafter, the Court can announce its ruling in the case, and only in recent decades did the Court begin to

schedule announcement days on its calendar (Davis 1994). On the morning of the announcement, the Court distributes written opinions Page 27 →through its Public Information Office and in recent years on its official website, supremecourt.gov. Along with these formal opinions comes a case syllabus, prepared by the reporter of decisions at the Court. This syllabus is not formal law but rather is intended to summarize the Court’s main holdings,2 drawing directly from the written opinions. Upon release of a decision, one justice verbally summarizes the holding in the courtroom; on rare occasions, other justices may speak, offering dissenting opinions (Witt 1990). Because the Court building has seats for only 250 observers, many activists and reporters are forced to wait outside to hear the first news about significant decisions. The justices of the Supreme Court take no questions and offer few clarifications about their opinions (though they may illuminate their thinking when they choose to speak in oral arguments, when they announce a ruling, or when they make outside appearances). They do little to simplify the complex legal analysis contained in their writings (Ginsburg 1995). They have never allowed cameras inside the courtroom (an archive of audio recordings exists at oyez.org). And the Public Information Office, created in 1935 to institutionalize relations with the press, provides minimal content to the media beyond that produced by the jurists. Unlike other press offices, it makes no effort to frame stories about the Court or its rulings, instead accommodating the media only to the extent that doing so benefits the institution’s image. According to a former public information officer for the Court, “We’re not spin doctors, as it were. It is a very different office from the other two branches” (Davis 1994, 47). Because the Court “speaks once and is silent” (Newland 1964, 35), reporters face a unique and challenging environment. The resources necessary to offer year-round coverage of the institution are considerable, both for journalists and news organizations as a whole. Full-time Court reporters must be well versed in legal affairs (Hess 1981). Some but not all journalists who cover the Court receive a coveted media credential, granting them access to a seat in the courtroom for oral arguments and decision announcements. Their training often includes journalism and law school, coupled with years of experience in the field, though reporters disagree about whether a law degree aids them in covering the Court (Davis 1994, 67). A news outlet must be able to devote a full-time member of its staff to report on an institution that releases a small number of rulings each year, only a handful of which will generate public interest. For these reasons, leading national news organizations (New York Times, Washington Post, ABC, NBC, CBS, CNN, Fox News, MSNBC, and National Public Radio, and so forth) offer most of the original reporting on the Court. Page 28 →Reporters focus most of their coverage on the cases that reach the Court’s docket (Slotnick and Segal 1998). To shape this coverage, they typically familiarize themselves with the background of the cases to which the Court grants cert (Greenhouse 1996). Such background includes the case history, factual circumstances, legal controversies, and political implications. The case history encompasses the process by which the case makes its way through the judicial system until the high court grants cert. Factual circumstances of the case involve the scenario that gives rise to the legal challenge before the Court. The legal controversy is among the most important and complex matters with which reporters have to deal. It includes the legal matters highlighted by the cert petition, which commonly involve a conflict between the language or application of a law and the Constitution. Finally, journalists also familiarize themselves with the political implications of a dispute, as many cases draw public attention for their political consequences more than their legal ones. To understand the extent of background knowledge necessary for journalists on the Supreme Court beat, take the Affordable Care Act (ACA) case as an example. The case history began with a challenge joined by 26 states and the National Federation of Independent Business to the ACA’s constitutionality. The federal District Court for the Northern District of Florida found the law’s mandate that individuals purchase health insurance to be an unconstitutional exercise of federal power. The Court of Appeals for the Eleventh Circuit upheld this ruling on appeal but found that the mandate could be severed from the rest of the ACA, allowing the remainder of the law to stand. The Supreme Court then agreed to hear an appeal of the Eleventh Circuit’s decision, National Federation of Independent Business v. Sebelius. The factual circumstances of the case were straightforward—the states challenged the major portion of the ACA (the individual mandate), which had not yet taken effect. In this case, the Court allowed the challenge to proceed, but in some cases the Court finds that petitioners do not have

legal standing to challenge a law (Roberts 1993). The central legal dispute in NFIB v. Sebelius involved whether the health care mandate exceeded the scope of Congress’s powers to regulate interstate commerce or to levy taxes. The case’s political significance was well known and monumental; indeed, many more Americans were likely interested in the political battle over national health care than in the Court’s interpretation of the Commerce Clause. News organizations covered all of these aspects of the ACA case to provide a full and accurate portrait to their audiences. All told, journalists covering the Supreme Court face a complex environment (Greenhouse 1996). They must bring expertise to bear on controversiesPage 29 → with a variety of dimensions. They must familiarize themselves with case specifics as well as history, law, and politics. And they must do so subject to the restrictive media standards developed by the Court itself (Witt 1990).

Constraints in the Newsroom The journalistic pressures placed on Supreme Court reporters are just as great as the expertise required of them. Coverage of the Supreme Court must meet a variety of professional objectives. It must be simple, accurate, and timely yet comply with the criteria for newsworthiness by generating drama and emphasizing conflict (Gans 1979; Davis 1994). Reporters covering the Court face a number of challenges as they aim to fulfill these objectives. Simplicity According to a veteran Supreme Court journalist, “A judge may take three pages to discuss a minute point of law, while a reporter may have three sentences to explain the meaning and impact of the entire decision” (Knoche 1987, 268). The need for simplicity presents a particular challenge for journalists who cover the Supreme Court, whose rulings delve into legal minutiae. The news audience has little preexisting knowledge of the legal and constitutional issues at stake in a given case, meaning that coverage must explain not only an outcome but also the background necessary to understand it (Greenhouse 1996). Consider one of the landmark holdings of the 20th century, Roe v. Wade, where the Supreme Court struck down bans on abortion as incompatible with the Constitution. To provide Americans with an accurate account of the decision, the press had to explain the majority’s reasoning, which was based on a complex legal analysis. Though the Constitution does not explicitly mention abortion, the Court found that abortion rights were protected as part of the fundamental right to privacy enjoyed by all Americans. But the right to privacy itself also does not receive explicit mention in the Constitution. Rather, the Court reasoned that the right to privacy is ensconced in the document in multiple places, including in “penumbras” and “emanations” found in the First, Fourth, and Fifth Amendments protecting private associations, security of person and property, and a zone of privacy from self-incrimination, respectively (Griswold v. Connecticut [1965]). At the same time, the Court ruled that the state has two legitimate interests in regulating Page 30 →abortion: protecting prenatal life and a woman’s health. Because the Court found that the balance of these interests with the right to privacy changes throughout a pregnancy, it used a trimester system to determine the restrictions that can constitutionally be placed on abortions (Roe v. Wade [1973]). The case is extraordinarily complex from a legal perspective, yet reporters faced the daunting task of simplifying its outcome for the millions of average Americans who took an interest in it. Reporters also had to simplify a range of other information in their coverage, including the facts of the case and its implications for Americans and their political system. But even less complex cases require a great deal of simplification in the news. Every case that reaches the Court has a dense history behind it (e.g., Carpenter 2012). Because understanding every ruling requires some historical context, and because every ruling may be framed along both political and legal dimensions, the need for simplicity places acute demands on Supreme Court reporters (Goldstein 2012). Accuracy

Court reporters could more easily simplify judicial rulings if they did not need to achieve accuracy. For example, a simpler portrait of the holding in Roe v. Wade would note that the Constitution protects a “right to abortion.” While easy to understand, this account makes two mistakes: it confuses a right to abortion, which is not an explicit part of the Constitution, with the implicit right to privacy, and it implies that the right to abortion is absolute (which it is not, since the Court found that the right to privacy must be balanced against state interests in regulating abortion to preserve the health of the mother and life of the fetus). In such complex scenarios, simplicity and accuracy can be at odds; indeed, there are numerous examples of the media’s errors in reporting on judicial decisions (Newland 1964; Greenhouse 1996; Goldstein 2012). The media commonly mischaracterize the legal meaning of Court actions by, for example, reporting cert denials (the refusal to hear a case) as decisions on the merits (rendering a judgment about a case’s outcome; Slotnick and Segal 1998). But news organizations strive for and often achieve accuracy in their coverage. They do so by relying first and foremost on the expertise of reporters with experience and legal training (Hess 1981; Davis 1994). These reporters familiarize themselves with the facts of a case, arguments made, and potential decisions before the Court releases its ruling (Sherman 1988). In so doing, they are well prepared to make sense of it. The coverage of the Obamacare decision from scotusblog.com, a leading Page 31 →website for judicial reporting, provides an example of the care that many news organizations take to report on decisions in simple and accurate terms. The website’s publisher describes the strategy for reporting in a timely and correct manner on the ruling: Today, our entire team will work together in person, including Lyle [Denniston], who is the most experienced member of the press corps, having covered the Court for more than five decades, and who has written about the case with incredible depth; and four lawyers who have collectively argued thirty-four Supreme Court cases and taught Supreme Court litigation at Harvard and Stanford for more than twenty years. And nothing will to go onto the Live Blog about the Court’s ruling until I say I agree with it. If we make a mistake, I will be personally responsible.В .В .В . Within a few seconds of getting the opinion, I realize from reading the syllabus—and announce into the conference call—that the government has lost the Commerce Clause argument, but that there is much more going on that is going to require careful study.В .В .В . It takes me almost one minute exactly to analyze the decision. After about twenty seconds, I am almost certain that the government has won under the tax power. But I worry the opinion itself would have important nuance or qualification. So I read all the first sentences of each paragraph in the critical part of the opinion announcing the tax holding. It is clear that although the Court has rejected the government’s Commerce Clause argument, it has upheld the mandate without qualification. I turn to Lyle—who has been focusing on the Commerce Clause section of the syllabus on his copy, but also skimmed the tax power discussion in the syllabus—and say, “They win under the taxing power.” Lyle responds, “Yes.” Kevin Russell has gotten a third copy of the opinion from an NBC runner; he agrees. (Lyle then turns immediately to writing the overwhelming majority of our team’s analysis of the case, as he has throughout.) I dictate to [SCOTUSblog editor Amy Howe]: She repeats it to me to confirm, and publishes the update announcing the decision to our readers. (Goldstein 2012) The account hints at the various demands placed on news organizations when the Court releases a landmark decision. The early stages of reporting—providing the first summary of the ruling minutes after it is released—focus Page 32 →on speed, simplicity, and accuracy. But despite the demand for the former two, many media outlets strive for the latter as well. In the case of Obamacare, scotusblog (and other organizations) used a team of experienced reporters to parse the decision and confirm the analysis.3 In cases where the demand for

information is less immediate, news outlets might rely on the analysis of a single experienced reporter to interpret a ruling. Timeliness The need for fast and accurate reporting on high court rulings is an ever-present worry for reporters. On decision days, the justices announce their rulings beginning at 10 A.M. Eastern Time. This gives print reporters approximately half a day to file their reports for the next morning’s paper. In this time, the reporters must organize their preexisting information on the case, interpret the ruling and the opinions of dissenting justices, speak with legal experts to clarify and expand on interpretations, and seek out other interested parties for their reactions (Greenhouse 1996). They must then craft a story worthy of publication. In recent years, the time demands have become much more acute. Twenty-four-hour cable news channels often broadcast their first reports within minutes of a decision announcement. The Internet can yield its first headlines in under a minute. In the ACA case, CNN and Fox News put up banner headlines announcing the outcome about 50 seconds after the decision’s release. Other organizations followed suit within a minute. Unfortunately, CNN and Fox News filed erroneous reports suggesting that the Court had invalidated the ACA. Their mistakes stemmed from an incomplete reading of the decision: while its first few pages described the law as an unconstitutional exercise of Commerce Clause powers, its later pages declared the law constitutional since it imposed a tax. Again, the complexity of judicial decisions, coupled with the media’s need for timely reporting, created this situation. But even in less monumental cases, reporters face the challenge of portraying the work of the justices in simple and accurate terms within a few hours or minutes of the release of the decisions (Davis 1994). Drama The goal of newsworthiness also informs the work of Supreme Court reporters (Katsch 1983; Davis 1994). So coverage must create drama that draws attention from potential readers, viewers, or listeners. Indeed, most news Page 33 →organizations see attracting an audience as a central goal (Gans 1979). Depending on the case, journalists may develop drama by highlighting human-interest stories, reactions to a ruling, or tension on the Court itself. Reporters often draw human-interest stories from the facts of the case (Slotnick 1991). For example, much of the coverage of Atwater v. Lago Vista (2001) focused on the legal violation that gave rise to the case—the failure of a woman to wear a seat belt in her car, leading to her arrest as her two children watched from the backseat. These circumstances allowed the media to describe prohibitions against unreasonable searches and seizures. Similarly, the press focused on ordinary homeowners in its coverage of the eminent domain case, Kelo v. City of New London (2005). USA Today described the petitioner’s view that “protecting her Victorian dream house from the urban renewal bulldozer was always going to be an uphill fight.” CNN aired an interview in which the guest described the stakes of the case, noting, “The house, you know, is not a commodity to my father. My father came over from Italy in 1962 and, you know, land to him means that he’s rich and that he’s got all the gold in the world. You take that away from him, and you know, he has nothing.”4 Such interviews provide the necessary drama to keep viewers interested while making plain the stakes surrounding the Supreme Court’s work. Similarly, reporters may focus on reactions to rulings to create drama. In many cases, interested parties serve as compelling sources. Much of the coverage involving the health care decision emphasized the reactions of President Barack Obama and other politicians. Other reports looked to activists who had an interest in the case. The Washington Post, for example, posted a story, “Health-Care Ruling Reactions from the Supreme Court Steps.”5 At other times, reactions provide dramatic visual cues. For this reason, CNN reported live from the University of Michigan in 2003, when the Court upheld a portion of the school’s race-conscious admission policy. In the background, students marched and chanted in support of the ruling.6 Finally, most cases provide some measure of intrigue from the Court itself. This intrigue begins with oral

arguments and continues with the tension in the days leading up to a decision announcement. In many cases, disagreements among the justices also lend drama to a decision. On rare occasions, the reading of oral dissents from the bench produces unscripted moments of anger even after the release of a ruling (Toobin 2012, 80–81). Conflict Finally, reporters face the challenge of emphasizing conflict in their coverage of the Court. Doing so helps to create a dramatic narrative while following Page 34 →the institutionalized practice of highlighting multiple perspectives. Because each case involves a dispute between two or more parties, the opportunities to portray conflict are ample. Reports may focus on the disagreements between the petitioners and respondents in a case, between their legal teams, or between various amicus curiae who have taken sides before the Court. Many cases also engender disputes outside the courthouse building—between interest groups, activists, and politicians. And journalists can further highlight conflicts within the legal profession, such as between lower courts that have disagreed about the case or between the Supreme Court justices themselves. Once again, the Obamacare case allowed the media to portray all manner of conflict, including disputes between the federal government and the states, President Obama and Republicans, hospitals and insurance companies, and appeals courts and the Supreme Court. Some headlines on the case made plain such conflict: “Conservatives Turn on Roberts,” “Decision on Obamacare Fuels Obama-Romney Presidential Campaign,” “Corporate Winners, Losers under Healthcare Ruling.”7 More dramatically, a number of news organizations explored the conflicts between the justices, manifest by the splintering of the Court’s conservative wing with John Roberts’s decision to write for the majority. As one story noted, Roberts “originally sided with the four justices who thought the individual mandate was unconstitutional, then changed his mind and wrote the majority opinion for the liberals who wanted the law to stand. And even as he signaled he was siding with the left side of the bench, the justice who was thought to be the swing vote, Anthony Kennedy, lobbied Roberts intensely but to no avail.”8 The vote switch had everything that Court reporters desire: a simple and accurate way to frame the case outcome in a manner that emphasized drama and conflict. One longtime reporter, taking stock of how the press covers the Court, views conflict as central. “The overwhelming impression that journalism about the Court—including my own—probably conveyed to the casual reader was of an institution locked in mortal combat, where sheer numbers rather than force of argument or legal reasoning determined the result” (Greenhouse 1996, 1551–52).

The Model of Dissensus Dynamics Unique challenges face news organizations and reporters who cover the U.S. Supreme Court. Chief among these are the need to present complex judicial decisions to the public speedily, in a newsworthy manner. The Court itself Page 35 →does little to aid journalists in meeting these objectives. The institution does not frame its decisions for public consumption, does not clarify them upon their release, and does not even provide notice as to when it will rule. The model of dissensus dynamics details the factors that influence how media organizations shape coverage of judicial rulings in this unique environment. The model is based on a simple premise: given considerable constraints, news organizations rely on straightforward and readily available pieces of information that allow them to balance a range of objectives in their portrayals of Supreme Court rulings. These objectives include achieving accuracy and simplicity while emphasizing drama and conflict. The dissensus dynamics account does not address salience. It does not explain what makes the press more attentive to some decisions than others (though other research examines precisely that question; see Franklin, Kosaki, and Kritzer 1993; Davis 1994; Franklin and Kosaki 1995; Spill and Oxley 2003). Rather, it focuses on the content of coverage alone. Given that the press pays attention to a Supreme Court decision, what determines the nature of that coverage?

The Scope of the Model Dissensus dynamics focuses on the content of coverage of high-profile Supreme Court rulings from leading national news organizations. There are a few reasons for the model’s scope. First, I explore decision-centered coverage since the justices’ main responsibility is to resolve legal disputes with their rulings. The study of rulings grounds the work of the Supreme Court in specific legal and political controversies. But not all rulings receive equal attention. Most press coverage focuses on a handful of high-stakes cases each year that reach both attentive and less attentive audiences (Berkson 1978; Greenhouse 1996). These cases, along with Supreme Court confirmation hearings, constitute the main source of popular knowledge about the Court (Gibson and Caldeira 2009a). Dissensus dynamics further concentrates on national press organizations because they are the most visible and influential sources of judicial reporting. The training of the reporters at these institutions, coupled with the resources at their disposal, allows them to offer some of the most comprehensive accounts of judicial branch action (Davis 1994). Their reports not only reach a large number of interested Americans but also powerfully affect subsequent coverage. Indeed, many smaller news organizations that lack Court reporters republish the work of these major media outlets. Furthermore, I expect (and test) that similar factors influence the Page 36 →content of reporting across leading national news outlets since they operate with analogous goals and under comparable constraints as they cover the Court. Exploring the dissensus dynamics model fully (across a wide range of high-profile rulings) requires generalizable hypotheses, portable measures, and a broad sample. These things are rare in the sociolegal literature, where most studies focus on a single case or a narrow area of the law. And indeed, it is easy to envision how an inquiry into press coverage might analyze the ACA case alone (e.g., tracing the various frames used to describe the ruling, outlining the legal arguments discussed, cataloging statements of praise and criticism from interested parties, accounting for the balance of partisan sources used by the press). But reporters on the Supreme Court beat confirm the systematic nature of their coverage. According to one journalist, the development of the initial report on a decision follows a pattern: It’s almost a formula story: Supreme Court upheld, struck down, did x, y, z. Then I try to give a sense of the vote and perhaps implications. This would include a quote from the majority, a little background of the case, a quote from the dissent, and then more background. It’s not real hard to do. It’s almost like a science. (Davis 1994, 83) The question then becomes how to catalog a diverse range of reporting, touching on a variety of legal controversies, according to a portable standard of coverage content. To speak more generally about a range of judicial decisions, I focus on the tenor of published reports. I expect that coverage of a high-profile Court ruling can vary from favorable to unfavorable depending on the frames it employs, arguments it raises, sources it quotes, and tone it evinces. To be sure, a broad conceptualization such as this has its shortcomings, particularly in its loss of issue-specific detail.9 But the comparative value of a cross-case study far outweighs these shortcomings since the determinants of coverage are likely operate across cases and because very few existing studies have attempted to explore these determinants. Shaping Coverage of Supreme Court Decisions What factors determine the extent of favorable coverage afforded to a Supreme Court ruling? Recall again the premise of dissensus dynamics: given considerable constraints, news organizations rely on straightforward and readily available pieces of information that allow them to balance a range of objectivesPage 37 → in their portrayals of Supreme Court rulings. The information from which news outlets can draw includes information released by the Court (case syllabus, written opinions, oral summaries, voting outcome) and materials provided by outside sources (statements from lawyers and scholars, issue activists, parties to the case, politicians, and other interested observers). But outside source material becomes valuable for media outlets only after they have an understanding of a decision, which requires them to rely first and foremost on information provided by the Court. As experienced legal reporters begin to make sense of a ruling, they then call on sources for further context, interpretation, and reaction. Indeed, Court reporters attest that they first consult the written work of the justices after its release (Davis 1994; Slotnick and Segal 1998).

The information released by the Supreme Court is thus the most important for media coverage of rulings. But each piece of this information has a distinct value. First, consider the written opinions. In these pages, the justices lay out the entirety of their legal analyses. The majority opinion represents controlling law. In it, the justices review the facts of the case, the legal controversies it involves, their views of the relevant laws, and their holdings in the case. The opinion relies heavily on footnotes, referencing other cases that come to bear on the decision. The majority opinion can vary considerably in length, from a few words (e.g., Claiborne v. U.S. [2007]) to dozens of pages (Roe v. Wade [1973]). Other justices often file concurring or dissenting opinions that perform similar functions to the majority opinion—analyzing the case, making a determination about the law, responding to other arguments—but do not represent controlling law. Reporters may draw the most comprehensive and accurate portrayal of a ruling from its written opinions. These opinions also help capture specific points of conflict between the justices where they exist. But they are particularly poor at providing a simple, dramatic, and rapidly interpretable account of a decision. Even reporters with extensive legal training may find it impossible to quickly read and accurately summarize entire written opinions (Greenhouse 1996). This task is made all the more complex by the fact that a single opinion may constitute the holding of different coalitions of justices at different points. A Supreme Court judge may choose to join some parts of one opinion (like that issued by the majority) and select parts of another (i.e., the dissenting opinion). She may also file her own concurrence or dissent in addition to joining parts of other opinions. In the Obamacare ruling, the opinion of the Court included four parts, all written by Chief Justice John Roberts but signed by various coalitions (four justices joined parts I, II, and III-C; two joined part IV; and none Page 38 →joined parts III-A, III-B, and III-D). In addition, Justice Ruth Bader Ginsburg filed an opinion that concurred in part with Roberts and dissented in part and that was joined by Justice Elena Kagan in its entirety and by two other justices in part. Four justices signed a dissenting opinion, while Justice Clarence Thomas added his own dissent. The reporting of CNN and Fox News on the ACA makes plain another risk of relying on written opinions: the loss of accuracy. Erroneous reports based on written opinions are most likely to be filed when journalists lack the expertise to interpret opinions and when time pressures are great. As such, the language of written opinions exerts a relatively small effect on the favorable coverage a decision garners. Similarly, oral summaries provided in the courtroom on decision day provide little value to journalists, who must be present and take accurate notes. In most cases, these summaries are not particularly newsworthy since cameras are not allowed to capture the scene, disagreements are rare (except when a justice reads an oral dissent), and the justices offer little information beyond what can be found in their written opinions. Therefore, the most simple, accurate, and easy to understand pieces of information available to reports on the Supreme Court beat are the case syllabus and voting outcome (Davis 1994). But the case syllabus only presents text from the Court’s opinion (and not the dissents), so what it gains in simplicity it loses in its ability to capture conflict. Reporters cannot accurately portray disputes on the Court from the case syllabus alone. However, the final voting outcome in a case provides a straightforward indicator of conflict—the size of the majority and dissenting coalitions. In any given case, the voting outcome provides a wealth of information about the diversity of opinion surrounding a decision. A unanimously decided case sends a strong signal to reporters that the legal and political principles at hand are without controversy, so much so that nine ideologically diverse Court justices reach agreement about them (see Epstein, Segal, and Spaeth 2001). In these cases, the balance of coverage characteristics is weighted heavily toward simplicity and accuracy. Reporters are unlikely to search widely for critics of such decisions because of the strong judicial indication that they are uncontroversial. Journalists themselves may further be leery of criticizing such rulings given the justices’ expertise in constitutional law. Five–four rulings, conversely, provide a clear indication that the Court’s decision is controversial, contested, and potentially incorrect. In these cases, the balance of coverage characteristics is weighted more heavily toward drama and conflict. Indeed, if the justices themselves—whose analysis is purportedly based on a careful interpretation of law alone—cannot reach Page 39 →agreement, it is likely that many other political partisans will have strong reasons for disagreeing with the Court. In these cases, judicial dissent should have the effect of limiting the favorable press coverage

afforded a ruling. As such, dissensus dynamics suggests that the most influential factor that shapes coverage of rulings is that which provides a framework for capturing drama and conflict in a simple, accurate, and timely matter: judicial voting outcomes. Reporters will use frames that reflect positively on a ruling when the justices signal their agreement about the legal principles at stake with large majority voting coalitions, while decisions that engender dissent on the Court will lead reporters to frame them in an unfavorable light. Dissent Hypothesis: As the number of dissenters increases, Court rulings will garner less favorable media coverage. The dissent hypothesis echoes Greenhouse’s argument that “sheer numbers” guide the media portrait of conflict on the Court (1996, 1552). It is not, however, only the size of dissenting coalitions to which the press likely pays attention. Most journalists familiar with the Court can readily identify justices by their ideological predilections. For many years, the press tagged Justice John Paul Stevens as the Court’s most liberal jurist, and Justice Antonin Scalia has earned a reputation as its conservative leader. Similarly, the press often divides the Court between its liberal and conservative wings, with one or two justices considered swing votes (a role currently occupied by Justice Anthony Kennedy). Because of the judges’ well-known ideological preferences, they send another signal with their votes: the ideological controversy that may surround a case. Consider two hypothetical decisions, both decided by five–four votes. In the first, a broad coalition of the Court’s two most liberal members and three moderate conservatives vote in the majority. In the second, the Court’s five most conservative members align against its four more liberal justices. The “ideologically diverse” decision indicates to the press that while controversy exists about the ruling, it is rooted in neither partisan nor ideological disputes. But the “ideologically divided” decision suggests the presence of both legal and political controversy. The press may highlight criticism of it from not only unsympathetic law professors but also from liberal Democrats and interest groups that disagree with it. Recall again the ACA decision, whose majority opinion was written by the normally conservative Chief Justice Roberts and signed by four liberal justices. The five–four vote demonstrated clearly the presence of legal and perhaps political controversy (though the case would have received a wealth Page 40 →of critical attention regardless), but the ideologically diverse majority coalition likely insulated the Court from some of the most vehement criticism (see chapter 1). Elites found it more difficult to attack the opinion as an example of liberal bias when a conservative justice (appointed by a Republican president) penned it. Indeed, a large amount of press coverage after the ruling centered on a resurgent Republican push to repeal the law rather than to denounce the decision. Ideological Diversity Hypothesis: As the Court’s majority coalition becomes more ideologically diverse, a ruling will garner more favorable media coverage. The dissent and ideological diversity hypotheses provide the basis for dissensus dynamics, linking the demands of Supreme Court journalism with the most powerful tool that the press can use to make sense of rulings: the votes of the justices themselves. The model suggests that these votes will have the strongest influence on the framing of judicial decisions because they provide a path for the media to emphasize drama and conflict in a simple, accurate, and timely manner. Simply put, Court rulings that bring together large and ideologically diverse coalitions will receive more positive depictions than those with a small and narrow majority. Though no scholarship offers a comprehensive account of how the press covers a cross-section of Supreme Court rulings, other studies provide some support for the basic tenets of dissensus dynamics. Veteran journalists suggest that to the extent that they adhere to consistent procedures in reporting on judicial branch decisions, they reference voting outcomes and draw from the majority and dissenting opinions where available (Davis 1994; Greenhouse 1996). More important, the political communication literature demonstrates the central role that political elites (including

the justices) play in framing issues in the media. This work verifies that reporters of political news rely on highstatus political sources in crafting coverage (Sigal 1986; Baum and Groeling 2005). Such behavior can be traced to the constraints under which journalists operate—the need to report in a timely and accurate manner under conditions of uncertainty—which foster a reliance on authoritative, legitimate sources of information. In the semisymbiotic relationship between elites and the press, the former has the upper hand (Gans 1979; Schlesinger and Tumber 1994). According to Schudson (2002, 255), “The basic orientation of social scientists is that political news making is a reality-constructing activity that follows the lead of government officials.” Page 41 →Within this framework of elite influence, two lines of research add nuance. The first involves indexing theory, which explores on the range of content presented in the news. The theory suggests that elites not only shape the central content of coverage but also define the scope of perspectives presented by the media (Bennett 1990; Bennett and Livingston 2003; Bennett, Lawrence, and Livingston 2006). In other words, a diverse array of opinions and perspectives receive coverage only to the extent that they are espoused by authoritative elite voices. A second model of elite influence, cascading activation, makes distinctions about the degree of elite influence based on the relative status of sources. Entman (2004) demonstrates that elites’ power to shape the content of press coverage is stratified: framing perspectives spread from high-level elites to other elites, the news media, and eventually the public. The model explains “how thoroughly the thoughts and feelings that support a frame extend down from the White House through the rest of the system—and who thus wins the framing contest and gains the upper hand politically” (Entman 2004, 9). How do these insights inform our understanding of journalism at the Supreme Court? Consider first that in the realm of legal affairs, the justices occupy the highest place in the American political system. As members of an “official decision circle” (Bennett 2006, 468) with a high degree of credibility (Druckman 2001; Bartels and Mutz 2009), Supreme Court justices exert a powerful influence on the media’s coverage of law and politics. At the same time, the range of perspectives espoused by the justices also affects press coverage. The Court can most effectively frame legal decisions when it endorses a unified perspective with large and ideologically diverse majority coalitions. Dissensus dynamics, then, follows from a tradition of political communication research that sees high-level elites as the most influential source of frames for the press. But it also recognizes the constraints facing journalists who operate in the singular environment the Court fosters. These journalists rely first and foremost on judicial voting outcomes to guide their coverage of high court rulings.

Alternative Hypotheses The model of dissensus dynamics views elite disagreement as expressed through the votes of the justices as the most important factor that influences coverage across a range of rulings. But alternative explanations for decision coverage exist. I explore four of the most powerful alternative explanations to demonstrate why they lack the explanatory value of dissensus dynamics. Page 42 →Alternative Hypothesis #1: Nonjudicial Elites Influence Coverage Does the press shape its coverage of Supreme Court rulings in response to political elites other than the justices? There are a number of reasons to expect that such might be the case. First, existing research shows the influence of executive branch sources on coverage (Bennett 1990; Entman 2004). Given the visibility of a sitting president, he is certain to have an impact on coverage of the Supreme Court when he expresses strong opinions. Furthermore, by highlighting the views of other politicians, the press may focus on interbranch conflict to enliven its reporting. And political opinion about high court actions gives the news sources an opportunity to use partisan frames that make coverage easier to understand and more dramatic for their audiences. Conversely, evidence shows that the press pays attention to the relative status of its sources. According to Entman, “The spread of frames is stratified; some actors have more power than others to push ideas along to the news and then to the public” (2004, 9). While presidents may enjoy the highest status in foreign policy debates, the

Supreme Court is seen as the ultimate arbiter of legal conflict (Gibson, Caldeira, and Spence 2003b). We would expect the justices to exert the most powerful influence relative to other elites (including the president) in shaping how the press covers Court rulings. But political elites outside the judicial branch are unlikely to affect coverage for an even more important reason: they have limited incentive to disagree with the justices. The judiciary maintains a wealth of support and strong institutional legitimacy (Gibson, Caldeira, and Spence 2003b; Gibson and Caldeira 2009a, 2009b). Americans strongly dislike outright attacks on the Court (Caldeira 1987; Shesol 2010). So political elites, who are often unwilling to the pay costs associated with attacking the judiciary, instead choose to express support for decisions with which they agree and remain mostly silent when they disagree. For this reason, unanimous rulings receive favorable coverage not only because the justices set the terms of debate for the media but also because such rulings make partisan elites reluctant to attack the Court.10 Even in closely divided decisions, elites have limited incentive to launch outright attacks on the Court. Their criticisms do not have any direct effect on the law in the United States, and even the president—who has the authority to disobey judicial decisions—risks his reputation in doing so. Most Americans strongly dislike strident attacks on the Court, even when they disagree with its decisions. Page 43 →The Obamacare ruling provides an example of elite reluctance to criticize the Court. The presence of four dissenters allowed Republican elites some measure of cover to denounce the ruling. The presence of dissenters also enabled Republicans to denigrate Roberts’s vote “switch” (which may have received less attention if the ruling had not been decided by a five–four vote). But few Republican elites showed a willingness to excoriate the decision despite their antipathy to it. At a press conference after the ruling was announced, presidential candidate Mitt Romney did little to attack the Court or its decision but focused more broadly on the negative aspects of the ACA and the need for its repeal. The New York Times covered Romney’s response in detail: “Mitt Romney declared Thursday that he would вЂact to repeal Obamacare’ if he was elected president, saying that he agreed with the dissenting justices in the Supreme Court ruling on Thursday.В .В .В . Mr. Romney might have been hoping for a different Court decision, but his campaign staff was not complaining.”11 The measured response of many elites even in this most divisive of cases demonstrates the Supreme Court’s power to set the terms of public debate about its rulings. Other elites are simply unwilling to attack many decisions outright and aggressively, meaning that the elites most responsible for shaping their coverage are the justices themselves. Alternative Hypothesis #2: Political Context Matters But perhaps the tenor of decision coverage changes depending on the political circumstances at the time. The press may, for example, frame decisions in more conflictual terms when the Court seems out of step with other political actors (regardless of whether these elites express their disagreement with the Court). The media might scrutinize more carefully a liberal outcome, agreed upon by a liberal judicial majority, when Americans have elected a conservative president and Congress. This scenario makes the frame of ideological conflict readily available to reporters. The political context explanation in the Obamacare case would suggest that unfavorable coverage of the ruling stems from the media’s ability to frame it as anathema to the Republican-controlled House of Representatives. But the political context explanation falls short on three counts. First, it flies in the face of existing research that shows that journalists rely on elite sources to shape their coverage (Bennett 1990). Without partisan elites willing to criticize the Court, it is unlikely that the media could successfully employ partisan conflict frames in their coverage. Indeed, the most powerful Page 44 →elite sources available to the press are the justices themselves. The explanation also demands much of journalists, who would face the task of making inferences about the ideological characteristics of a ruling, a Court majority, and other politicians without the benefit of elite sources. Given the complex nature of legal decisions, it is difficult to characterize them along an ideological spectrum, particularly

for reporters, who face considerable constraints (Greenhouse 1996). According to dissensus dynamics, the best avenue for characterizing decisions involves reference to the voting coalitions formed by the justices. And finally, the existence of ideological conflict between the Court and prevailing political mood is unlikely absent division among the justices themselves. How would journalists frame conflict between a unanimous Court and conservative politicians when any unanimous decision includes the votes of the most conservative jurists? Once again, journalists are more likely to rely on the voting signals sent by the justices themselves than on journalists’ own interpretations of the political environment. Alternative Hypothesis #3: Issue Area and Interest Group Participation Matters Another explanation for the nature of decision coverage revolves around the issues that come before the Court. Two potential challenges to the dissensus dynamics model exist with respect to issue controversy: first, certain issues themselves are inherently controversial, causing the media to seek out diverse perspectives whenever the Court rules on such matters (and irrespective of the final vote in a case); and second, attentive interest groups may shape popular discussions of rulings, standing ready to weigh in on decisions even when the Court speaks with one voice. I consider each of these challenges in turn. First, are not certain rulings likely to invite challenge if they violate long-standing cultural values about a policy area (Entman 2004, 14)? Are not some controversial political matters—like the case of Obamacare, or other rulings involving abortion rights or the death penalty—likely to generate unfavorable coverage when taken up by the Court, no matter the specifics of its rulings? The media may have incentives to play up conflict over these matters to the extent that they capture public attention and ensure readership or viewership. These incentives would hold regardless of the specifics of a Court decision. Some evidence suggests that this perspective has merit, at least with respect to coverage volume. Research shows that the media’s attentiveness to Court rulings varies dramatically depending on the issue at Page 45 →hand. Cases involving social policy, civil rights and liberties, criminal justice, and First Amendment law receive outsized attention in the press (Katsch 1983; Davis 1987; O’Callaghan and Dukes 1992; Slotnick and Segal 1998; Hall 2009; Blake and Hacker 2010), which suggests that news outlets might cover rulings differently depending on the issues involved. However, volume and content are separate matters. While evidence is robust that the press covers in greater detail certain issues on the Court’s docket, very little research shows that the content of this coverage differs, and none demonstrates that distinctive factors inform coverage for controversial and banal issues. In fact, there are strong reasons to expect that the content of press coverage should remain dependent on judicial votes, even once we control for issue specifics. In even controversial cases, the press relies on elite sources to bring conflict to life. If high-profile elites lack the motivation to criticize the Court about decisions involving sensitive issues, coverage will also strike a deferential posture toward the Court. And as discussed earlier, these high-profile elites have incentives to refrain from criticizing a united Court even when the issue at hand is controversial. The justices use consensual decisions to temper criticism in other ways as well (Epstein, Segal, and Spaeth 2001). In achieving unanimity or near unanimity, the justices raise the costs for critics by depriving them of the ability to side with some (dissenting) justices and use their analysis against the majority. Consensual decisions further reduce controversy and criticism because, in attracting the votes of many justices, they are more likely to be narrowly tailored and carefully delineated. This form of judicial restraint throws into question the simple analysis that a decision may be controversial because of the issue it involves rather than its substance. Supreme Court justices have a great deal of discretion at their disposal: which cases to hear, which legal challenges to consider, which constitutional remedies to prescribe and proscribe. For any given issue, the justices may shape a ruling in a manner that attracts varying degrees of support. Consider the issue of the death penalty, among the most controversial in American politics today, defined by polarized and durable attitudes and, more important, committed activists on both sides of the debate (Hanson 2008). Though the Court decided 78 cases involving the death penalty between the 1981 and 2007 terms (Spaeth 2010), only a handful received substantial attention and criticism in the mainstream media. For reporters who cover the Court, then, the concept of inherent controversy in a case is less instructive than whether the justices make controversy apparent with their votes.

But if political elites are reluctant to denounce judges for fear of being penalized for overstepping their responsibilities in a constitutional government,Page 46 → other actors may lack similar qualms. Perhaps controversial issues bring about unflattering portrayals of judicial decisions because attentive interest groups willingly step in to drive the media narrative. An analogous perspective has been advanced by Eskridge, who argues that in recent decades, “organized interest groups have proliferated, producing more monitoring of judicial decisions” (1991, 338). While Eskridge focuses on the role of interest groups in shaping congressional responses to rulings, it is not hard to translate his perspective into one germane to the press. Consider the death penalty issue again. Any ruling involving the issue is likely to draw the attention of committed activists on both sides, some of whom will disagree with the Court. The activists have a set of responses they may offer, among them direct lobbying for a congressional response, but also seek to mold public opinion. One might hypothesize that for these reasons, coverage of any decision involving a controversial issue like the death penalty will present it in a more unfavorable light than one involving an uncontroversial issue, all else equal. And yet there are strong reasons to remain skeptical that even powerful interest groups will, absent a Court decision in their favor or internal division on the bench, drive the content of coverage that a ruling receives. First, the relative status of elite sources matters to news reporters as they craft coverage, giving the justices a larger megaphone than interest group leaders (Entman 2004). Second, reporters need to portray complex legal decisions with accuracy, which leads to an even greater reliance on the perspective of legal experts over policy-focused interest groups. Third, reporters enjoy a measure of autonomy, which, far from making them reliant on interest group sources, allows them to shape coverage according to their own paradigms (Terkildsen, Schnell, and Ling 1998). As we have seen, these paradigms are often formulaic in nature, as journalists lean heavily on the justices themselves even, it appears, as interest group leaders stand ready to respond when controversial issues come before the Court. The question of interest group influence is ultimately empirical and will be examined in the coming chapters. But much as we would expect with other political elites, strong reasons exist to anticipate that the justices will wield the most significant influence in media framing of their decisions. Alternative Hypothesis #4: Decision Characteristics Influence Coverage If news organizations cannot cover elite conflict over Supreme Court rulings absent elite sources, perhaps they can simplify the equation by providing distinctPage 47 → coverage depending on the characteristics of the decisions themselves. The media might exhibit more deference to the Court when the chief justice pens a decision, offering more favorable coverage. The press might report more positively on rulings that uphold existing laws and respect precedent, thereby avoiding the dreaded label of judicial activism (see Kmiec 2004; Green and Jarvis 2008). Or it might frame decisions that have a particular ideological bent in a positive light. For example, many press organizations may be particularly sympathetic to rulings decided by liberal majority coalitions. Each of these explanations is, on its own, plausible. Each demands only a limited amount of knowledge and offers a preexisting framing structure for reporters, and each may present a compelling storyline for the audience. At the same time, the explanations all require elite sources to speak about decision characteristics. Reporters may indeed structure their coverage around both the voting outcome in a case and specific features of a decision, affording more positive stories to rulings that have been decided by large majority coalitions and respect existing precedent, for example. But, in accordance with the dissensus dynamics model, the press still relies primarily on the criticism of dissenting justices to frame certain types of rulings in a negative light.

Summary Coverage of high-profile Supreme Court decisions plays an important part in shaping popular understanding about the law, the judicial branch, and political controversies. But despite the unique constraints on Supreme Court journalism, scholars know little about how coverage of rulings takes shape. To remedy these shortcomings, dissensus dynamics explains the content of coverage across a variety of cases as a consequence of voting outcomes on the Court. Because news outlets face time pressures as well as the need to

report in simple, accurate, and dramatic terms on rulings, they rely on the informative signal of judicial votes to guide their coverage. When the divisions between the justices become sharp, they express those divisions with dissensual decisions, which in turn cause the press to highlight elite disagreement. As a result, dissensual decisions receive more unfavorable portrayals than those decided by large judicial majorities. Because of press reliance on elite voices, the influence of judicial votes has the most powerful impact on the content of coverage, even once we account for the opinions of other political actors, the characteristics of decisions, and the issues under consideration. The following chapters offer a series of tests to explore the dissensus dynamics model and its implications. Chapters 5 and 6 review evidence for Page 48 →the major prediction of the model—that media framing of rulings follows from voting outcomes on the Court. Using a random sample of high-profile rulings released between 1980 and 2008, these chapters explore the content of national newspaper and cable news coverage—specifically, the extent to which it is framed in favorable terms. But first, chapter 4 tests, refines, and expands the dissensus dynamics model using a most-similar case study of two eminent domain rulings. To explore how media coverage of this area of law changed from the spring of 2005 (before either ruling) to the summer (after both had been announced), the chapter tracks the frames the national press used to discuss the issue and the influence of the Court’s rulings on these frames. It provides evidence that confirms some of the predictions of dissensus dynamics, as the media discussion of the issue and the Court became more critical following the dissensual and ideologically divided outcome in Kelo v. City of New London.

Page 49 →

Chapter 4 “All Private Property Is Now Vulnerable” A Case Study of Consensus, Dissent, and Media Coverage At the end of its 2004 term, the Supreme Court redefined takings law in the United States. First, in May, it broadened the power of the government to seize private property even if the seizure did not substantially advance legitimate government interests (Lingle v. Chevron U.S.A. Inc.). One month later, the Court expanded the takings power by permitting the seizure of private property for the purposes of private economic development (Kelo v. City of New London). The Washington Post admired the former decision for its candidness and restraint but argued that the latter was unjust though ultimately correct.1 Other media outlets would be even less forgiving in their coverage of Kelo. This chapter explores the dissensus dynamics model through a case study of these two eminent domain rulings. It shows how the newsworthy features of Kelo raised the salience of government takings. It demonstrates that the press’s portrait of property rights shifted dramatically between April and July. But most important, it shows how different voting outcomes in the cases strongly influenced their coverage. The unanimously decided Lingle did little to disturb vague media narratives about the invulnerability of American property rights, even as it rendered these narratives hollow. But the five–four outcome in Kelo had both direct and indirect effects on coverage. Internal disagreements on the Court caused the press to call into question the wisdom of the decision and raise alarms about property rights in the United States. Dissenting justices most effectively influenced media coverage when they employed evocative language to describe the ramifications of the decision.Page 50 → At the same time, though, the presence of judicial disagreement caused reporters to highlight other critics of the ruling, from high-ranking members of Congress to ordinary homeowners. These findings provide new insight into the process that generated singular discomfort with Fifth Amendment law after Kelo. They provide evidence for the basic tenet of dissensus dynamics: the press fashioned its coverage of eminent domain rulings in response to the level of internal disagreement among the justices themselves. But they also add to our understanding of the model, demonstrating how coverage of a legal controversy changes over time and how the effects of judicial dissensus are borne out, as the press highlights only the most newsworthy frames espoused by Supreme Court justices.

The Case Study Approach Because no existing studies have explored the media coverage of Lingle and Kelo in tandem, a case study of their coverage provides a wealth of analytic leverage. Case studies allow researchers to identify causal mechanisms and clarify concepts in a manner not afforded by large-N research (Mahoney 2000; Gerring 2007); they can also provide confirmatory evidence for a working hypothesis (Seawright and Gerring 2008). A case study further provides a level of detail not afforded by large-N research (Gerring 2007). It enables us to trace the specific frames that the media employed to characterize property rights law during the spring and summer of 2005. In so doing, we can make stronger links between the information presented to the public and the considerations that influenced opinion about the controversy (Nadler, Diamond, and Patton 2008). This chapter appraises how the national press framed property rights law after Lingle and Kelo. I expect that Court voting coalitions influence the content of press reports, with more dissensual and ideologically narrow rulings garnering unsupportive coverage. While exploring this hypothesis, I provide a more robust portrait of press coverage. What frames does the media use to characterize Supreme Court decisions? What influence do written opinions have on coverage? How do decisions change the coverage of an issue over time? What other mechanisms translate rulings into press reports?

A most-similar design helps answer these questions. This approach enables the researcher to isolate an explanatory variable while controlling for other confounding factors in the cases under investigation (Gerring 2007; Seawright and Gerring 2008). Ideally, a most-similar design would control Page 51 →for a variety of case features, listed here in order of importance: issue under consideration, substance of decision, ideological direction of decision, time period, legal significance of decision, Court membership, political climate, and involvement of other political actors. At the same time, the cases selected should differ in voting coalitions—one consensual, the other dissensual and ideologically divided. Lingle v. Chevron and Kelo v. City of New London meet these criteria quite well. In these cases, the Court dealt with the government’s ability to “take” property from private parties, in accordance with the Fifth Amendment of the Constitution. The cases were decided exactly one month apart with the Court offering broad interpretations of government power in both instances. Many scholars consider Lingle and Kelo to be among the most significant property rights rulings in decades (Baron 2007; Nadler, Diamond, and Patton 2008). But despite these similarities, the justices themselves diverged in their interpretations of the cases, splitting five–four in Kelo while reaching a unanimous agreement in Lingle.

The History of Lingle and Kelo On February 22, 2005, the Supreme Court heard arguments in two cases that would soon form a major part of its takings jurisprudence. Both cases dealt with the Fifth Amendment’s Takings Clause, which provides, “Nor shall private property be taken for public use, without just compensation.” The Court has traditionally interpreted this clause to give the government wide latitude in seizing property for public use and to require the government to pay fair market value for such property (Sax 2005; Dana 2006; Sterk 2006; Berman v. Parker). In Lingle v. Chevron U.S.A Inc., the Court was adjudicating a dispute on appeal from the U.S. Court of Appeals for the Ninth Circuit. The case involved a challenge to a 1997 law enacted in the state of Hawaii to curb high prices on gasoline. The law placed a limit on the amount of rent that oil companies could charge to lease their service stations. Gasoline refiner Chevron U.S.A. challenged the act as an unconstitutional regulatory taking of property,2 and two lower courts sided with Chevron, finding that the Hawaii law did not “substantially advance a legitimate state interest.” The Supreme Court heard arguments as to whether the “substantially advance” test, which had been applied in earlier property rights decisions, was an appropriate one for evaluating regulatory takings. On May 23, 2005, the Court released a unanimous verdict, written by Page 52 →Justice Sandra Day O’Connor, that struck down the “substantially advance” formula. In finding that earlier applications of the formula were unclear, the justices ruled that courts should not evaluate the purpose of regulations to determine whether they constitute a compensable taking. Though the Supreme Court remanded the case for further consideration, the decision debilitated Chevron’s challenge to the rent-control act. Lingle v. Chevron substantially broadened government power to regulate property and conduct takings without having to demonstrate a “legitimate” interest in doing so (Barros 2006; Baron 2007; Merrill 2010). Similarly, the Court offered a broad interpretation of government power under the Takings Clause in Kelo v. City of New London. At issue in the case was an economic development plan approved by the city of New London, Connecticut, to deal with its financial distress. The plan capitalized on pharmaceutical giant Pfizer’s arrival in the area by creating space for a hotel, restaurants, retail outlets, and offices, among other things. The city intended to use its eminent domain power to acquire land from homeowners who were unwilling to sell, in turn providing them with compensation for their properties. Two of these homeowners, Susette Kelo and Wilhelmina Dery, challenged the city’s use of eminent domain as a violation of the Fifth Amendment’s “public use” requirement. On June 23, 2005, the Court ruled against the petitioners, finding that New London’s economic development plan did indeed serve a “public purpose.” Writing for the majority, Justice John Paul Stevens noted that public benefits of the economic development plan fulfilled the public use requirement. Furthermore, he emphasized deference to the New London legislature, affirming an expansive government power to employ

eminent domain where lawmakers deem appropriate (as long as they provide compensation for seized property). Four justices raised strong objections to the decision. One dissenting opinion, written by O’Connor and signed by Justices William Rehnquist, Antonin Scalia, and Clarence Thomas, charged the majority with abandoning a “long-held, basic limitation on government power.” They argued that economic development projects served a private purpose, not public use. Ominously, they warned, “Under the banner of economic development, all private property is now vulnerable to being taken and transferred to another private owner” (Kelo v. City of New London, O’Connor dissenting, 494). In a separate dissent, Thomas offered a textual analysis of the Fifth Amendment to show that the majority had misinterpreted it. He further suggested that the decision would give new power to wealthy interests at the expense of minorities and the poor (Kelo v. City of New London, Thomas dissenting). Page 53 →To make meaningful inferences about the effect of judicial votes in these two cases, one must be confident that Lingle and Kelo are similar in many of their other characteristics. Table 4.1 catalogs many of the most important features of the cases, which I discuss in further detail. Most-Similar Features Lingle and Kelo are among the small number of cases decided by the U.S. Supreme Court that concern the Takings Clause of the Fifth Amendment. Though the circumstances and details surrounding these cases differ, they typically involve the government’s ability to curtail or eliminate private property rights for public goals. Indeed, a wide range of legal scholarship acknowledges the similarities between Lingle and Kelo, considering their joint implications for takings law in the United States (Ely 2004; Wroth 2005; Barros 2006; Breau 2006; Fletcher 2006; Baron 2007; Nadler, Diamond, and Patton 2008). Nonetheless, the cases deal with different aspects of takings: Kelo with the physical seizure of private property known as eminent domain and Lingle with regulatory takings—whether a legal regulation is so onerous as to constitute a compensable taking of property (Baron 2007). This distinction has little legal or practical significance; according to O’Connor, regulatory Page 54 →takings and eminent domain appropriations are “functionally equivalent” (Lingle v. Chevron, 539; see also Kent 2011). Both regulatory takings and eminent domain decisions adjudicate whether the Takings Clause of the Fifth Amendment permits specific invasions of private property rights. Table 4.1. Comparison of Two Takings Clause Rulings Lingle v. Chevron Kelo v. City of New London Fifth Amendment; Fifth Amendment; Issue regulatory takings eminent domain Substance Prescribes a new test to evaluate regulatory larifies the meaning of “public use” of decision takings Direction of Substantially broadens government takings Possibly broadens government eminent domain power decision power Legal New test New application significance Hearing: February 2005 Hearing: February 2005 Time period Decision: May 2005 Decision: June 2005 ourt Rehnquist court Rehnquist court membership Political Unified Republican government Unified Republican government climate

Perceived victim / orporations/oil companies losing party Voting 9–0 outcome

Middle class homeowners 5–4

The Supreme Court’s rulings in Lingle and Kelo reinforce the parallels between the regulatory takings and eminent domain questions. In the most reductive terms, Lingle remanded the dispute over Hawaii’s rent control act to a lower court while making it more difficult for challenges to the act to succeed. Kelo allowed the use of eminent domain for economic development to proceed in New London and other cities. Both cases offered broad interpretations of the government’s power under the Fifth Amendment (Fletcher 2006; Ely 2004). They rendered it significantly more difficult for property owners—whether companies or individuals—to block government takings. Indeed, scholars have viewed Lingle as noteworthy for eliminating the due process “substantially advances” test, affording government more latitude and requiring less justification for takings (Barros 2006; Baron 2007). Others view Kelo as redefining “public use” to similar ends (Kelo v. City of New London, O’Connor dissenting, 504–5; Thomas dissenting, 520–21; Kanner 2006). The temporal proximity of the Lingle and Kelo decisions further underscores their similarities. Argued on the same day, the cases were decided exactly one month apart, near the end of the Court’s 2004 term. The same nine justices heard oral arguments and voted in each case, ensuring that the different press reactions to each cannot be attributed to the Court’s membership itself. Similarly, the makeup of the other branches of government remained constant between the two cases, with Republicans controlling both houses of Congress and the presidency. Most important, the minimal time span between the Lingle and Kelo decisions ensures that the popular understanding of property rights remained largely identical. Between May 23 and June 23, no major shifts occurred in the way in which the media or public discussed property rights in the United States, and no other major controversies erupted involving the Takings Clause.3 These broad similarities make Lingle and Kelo strong candidates for a most-similar case study design, allowing us to rule out the alternative explanations that differences in the content of coverage stem from distinctions in their issue areas, decision characteristics, political contexts, or legal significance. Distinctions and Other Accounts of Differences in Coverage Though there are many reasons that recommend the Lingle-Kelo comparison, the case study method requires controlling for all possible alternative Page 55 →explanations. I consider other prominent differences between the cases besides their voting outcomes: the legal significance of the controversies, their newsworthiness, and popular perceptions of the losing parties. First, the legal similarities between the cases are substantial albeit imperfect.4 But even though the decisions deal with distinct aspects of takings law, they had similar implications for the popular understanding of property rights. For journalists, Kelo was alarming because the ruling upended the media’s (inaccurate) portrayal of these rights as inviolable. But so too did Lingle. Why did the media not sound the alarm about this decision? The legal distinctions between the rulings may also have rendered public displeasure with Kelo less likely; indeed, the media could have portrayed the precedent-altering Lingle as the more aggressive and troubling decision from a legal perspective (Baron 2007). For these reasons, the legal differences between the cases likely had a minimal impact on distinctions in their coverage. Second, one may question whether case-specific facts influenced coverage. This is undoubtedly true in one respect: Kelo presented a more dramatic and newsworthy controversy for three reasons: the division on the Court, animated by sharp rhetorical jabs included in the justices’ opinions; the various state and federal responses to the ruling that kept it in the public eye throughout the summer; and, perhaps most significantly, the identities of the parties involved in the dispute. Because these factors made the media’s extensive attention to the case likely, I focus on the content of such coverage and not its volume (a feature about which much has been written

previously). Though Lingle and Kelo differed in the drama they generated, they had a number of substantive similarities that could have allowed the media to portray these Fifth Amendment holdings as consistent with one another. One other important difference between Lingle and Kelo complicates the analysis of their press coverage: the identity of the losing parties (Nadler, Diamond, and Patton 2008). The Chevron corporation, a major oil refiner, was unlikely to elicit sympathy from either the press or the public after the Supreme Court ruled against it in Lingle. Conversely, petitioners Susette Kelo and Wilhelmina Dery made ideal protagonists as humble homeowners seeking to defend their properties from the invasion of the New London municipal government. It was somewhat unusual for a Takings Clause case before the Supreme Court to result in the justices siding with a government institution at the expense of a sympathetic set of claimants, such as ordinary citizens or middle-class homeowners, as table 4.2 shows. But though it is difficult to come by case study comparisons that control for the sympathetic nature of the losing party while holding constant a range of other factors, we can gain some analytic leverage by exploring the most-similar decisions: Page 56 →Hawaii Housing Authority v. Midkiff (1984); Berman v. Parker (1954); and a case before the Supreme Court of Michigan, Poletown Neighborhood Council v. City of Detroit (1981). In Hawaii Housing Authority v. Midkiff, a unanimous Court upheld a state government plan to break up an oligopoly in landownership and redistribute property to a wider population of residents. The justices characterized Page 57 →the losing party in the case as a group of “large landowners.” These landowners were quite clearly less sympathetic than the middle-class homeowners who came before the Court in Kelo. However, the losers in the Midkiff case could nonetheless be seen as victims in the sense that their personal properties came into the crosshairs of the state government, which seized property without their consent. In fact, many notable parallels exist between the decisions in Kelo and Hawaii Housing Authority v. Midkiff that, while animating public antipathy toward the former, failed to provoke an outcry over the latter. The Court even referenced its Midkiff precedent in Kelo, noting that “a legislative determination that the taking of land, even developed land, as part of an economic development project is a вЂpublic use’ and in the вЂpublic interest’” (Kelo v. City of New London, 5). According to scholars, much of the public outcry over Kelo came in response to three values threatened by the decision: the sanctity of the home, government overreach, and private rights (Nadler, Diamond, and Patton 2008, 304–5). At least two and possibly all three of these values were similarly threatened by the Midkiff decision two decades earlier. Yet despite the triumph of the state government in the case, media criticism of the ruling was muted, and no popular backlash followed. Table 4.2. Prominent Supreme Court Takings Decisions, 1981–2005 Holding Losing Party Sympathetic Losing Party? Hawaii Housing Authority v. Midkiff (1984) Keystone Bituminous Coal (1987) First English v. Los Angeles County (1987)

Plan to transfer land from large landowners Large landowners upheld

Unclear

Restriction on coal removal is not a taking oal company

No

Government must pay retroactive compensation after removing restriction on Government flood area

No

Nollan v. California Coastal Commission (1987) Lucas v. South Carolina Coastal Council (1992) Dolan v. City of Tigard (1994) Palazzolo v. Rhode Island (2001) TahoeSierra (2002) Lingle v. Chevron (2005) Kelo (2005)

onditions on building permits must be related to proposed construction

Beachfront landowners

Unclear

Regulation that denies all economically productive use of land is a taking

Government

No

onditions on building permits must be Government related to impact of proposed construction

No

Landowner can challenge regulations that predated purchase

No

Government

Moratorium on construction is not a per se Developers taking

No

Regulation need not substantially advance hevron corporation government interests

No

Forced sale of property to promote economic development upheld

Middle-class homeowners Yes

Source: Adapted from Nadler, Diamond, and Patton (2008, 293–94). Of course, large landowners, even when they are the victims of eminent domain, are not similarly sympathetic to Kelo’s losing parties. To find a more direct comparison, we have to search more broadly. Perhaps the most sympathetic claimants in major eminent domain cases before American courts can be found in a 1954 Supreme Court case and a 1981 dispute that reached the Michigan Supreme Court. It is hard to imagine more tragic victims of eminent domain than either the thousands of poor African Americans displaced by a Washington, D.C., plan that gave land to a private developer or the 4,000 working class residents of Detroit’s Poletown neighborhood who were forced to sell their homes to make way for a General Motors plant. In each case, a court nonetheless upheld government’s exercise of its takings power. Evidence of extensive negative media coverage of these cases would suggest the important role that the sympathy engendered by losing parties can play in shaping press coverage. But such coverage did not materialize. The Berman decision went largely unnoticed by the press and public (Nadler, Diamond, and Patton 2008). The Poletown case did attract a measure of national attention—the New York Times ran three stories about the case in March 1981, with two of them portraying the Michigan court’s ruling in negative terms—but nothing of the magnitude or degree that followed Kelo. Another way to explore the divergent press depictions of Lingle and Kelo—and the degree of difference that can be attributed to judicial voting coalitions versus perceived victims in each case—involves entertaining two Page 58 →counterfactuals.5 First, how would press coverage of Kelo have changed if the petitioners elicited less sympathy? On the one hand, it is likely that a large portion of the human-interest coverage of the case would have evaporated, since the controversy would have had held less interest for journalists looking to attract an audience. And public antipathy would have diminished to the extent that Americans did not view the ruling as violating the property rights of sympathetic claimants. But a number of stories involving Kelo framed the case as having little to do with its victims and more to do with its legal shortcomings. These frames, raised by O’Connor and

others in their dissenting opinions, would have remained. A second counterfactual lies at the heart of this analysis: How would a unanimous decision in Kelo have changed its press coverage? We can draw on other cases to increase our analytical leverage in answering this question. Consider that the Supreme Court’s eminent domain jurisprudence, which covered twelve cases as of 2005, has featured a variety of petitioners. Almost all of these cases passed without substantial negative attention in the press, and almost all were decided by consensual votes on the Court (Sax 2005). None of these cases, in fact, offered the compelling narrative of Kelo, so it is not surprising that they attracted less coverage. But what is surprising is the substance of such coverage and the emergent portrait of property rights law. The media’s portrayal of private property under siege can be traced in important part to Kelo’s status at the time as the Court’s only five–four ruling in Takings Clause jurisprudence. Of course, in spite of all of this evidence, it is unwise to claim that every difference in the coverage afforded to Lingle and Kelo can be attributed to the votes of the justices themselves. Did the identities of the losing parties in these cases play a role in the amount of attention they attracted? Very likely yes. Did these perceived victims affect the content of coverage? To some extent, yes. Did they animate public backlash against the decisions? Most definitely. However, the votes of the justices themselves also played a comparatively large role—indeed, perhaps the most prominent role—in this equation. Though it was unlikely the public would have a great deal of praise for any decision that sided with a local government as it seized the land of middle-class homeowners, the Court could have fostered a depiction of its Kelo decision as limited, deferential to both precedent and local authority, and principled had it achieved unanimity or near unanimity in the case. At the very least, the decision might have passed with little public notice. But instead, the presence of dissent on the Court and the strategic choices made by the dissenters led the media to portray the ruling in sharply critical terms. The decision was depicted as far-reaching and overaggressive, ignorant of Page 59 →precedent, unprincipled, and immoral. By framing the decision in a distinct and much more unfavorable manner than did the Court majority, the press opened the door for the public backlash that followed.

Testing the Effects of Judicial Votes in Lingle and Kelo The mainstream news media afford limited coverage to property rights as a general rule. But the Supreme Court’s decision in Kelo v. City of New London triggered a spike in attention, as news outlets regularly discussed eminent domain. Why did Kelo lead to such different reactions than Lingle? How did media discussions of property rights evolve from the spring of 2005, before these decisions, to the summer? What frames did the press highlight? How did it portray the Court’s rulings? To answer these questions, I gather coverage of the property rights issue from 10 leading media organizations. To ensure the validity of my inferences, I combine a narrative account of news coverage with more formal characterizations of it using automated keyword searches (which yield more reliable data). I use stories from the New York Times, Washington Post, Washington Times, USA Today, CNN, Fox News, MSNBC, Newsweek, the National Review, and National Public Radio (NPR). These sources vary in both medium (newspaper, cable news, magazine, and public radio) and ideological orientation (Groseclose and Milyo 2005; Ho and Quinn 2008). Nonetheless, because all sources offer original reporting on the Supreme Court, I expect the factors and constraints influencing their coverage to be similar (see chapter 3). I conduct Lexis-Nexis keyword searches for the following terms: private property, eminent domain, Fifth Amendment, Takings Clause, Lingle, and Kelo. A total of 394 news articles and segments mentioned one or more of these terms between April 23 and July 22, 2005.6 To analyze the effects of the Supreme Court’s takings jurisprudence on coverage of property rights, I divide the coverage period into three phases: Phase 1, pre-Lingle (April 23–May 22); Phase 2, post-Lingle, pre-Kelo (May 23–June 22); and Phase 3, post-Kelo (June 23–July 22). Figure 4.1, which plots the number of property rights stories on a daily basis, makes clear the importance of Kelo in raising the salience of the issue. One observable implication of the dissensus dynamics model is that the press will frame an issue in different terms

depending on whether a relevant Court ruling is consensual. We would expect the media to employ different frames to characterize property rights after Lingle than after Kelo. And indeed, each decision has a distinct effect on the media’s characterization of Page 61 →the issue (see fig. 4.2). Stories discussing the Fifth Amendment decline from 36 percent of all property rights news in Phase 1 to 29 percent in Phase 2 and to 23 percent in Phase 3. Conversely, the post-Kelo world becomes one in which the eminent domain frame defines the issue (more than three-fifths of stories include its mention). Not surprisingly, the Court’s announcement of decisions also changes the shape of coverage, with discussions of Lingle peaking in Phase 2 (7 percent of stories) and those of Kelo in Phase 3 (22 percent of stories). Page 60 → Fig. 4.1. Volume of Property Rights Coverage, Daily Fig. 4.2. Percentage of Property Rights Stories Mentioning Various Issues. (Note: *p < .1, **p < .05, ***p < .01, two-tailed, indicating a change from previous phase.) Fig. 4.3. Frequency of Dissent Mentions in Coverage of Property Rights. (Note: *p < .1, **p < .05, ***p < .01, two-tailed, indicating a change from previous phase.) Given the considerable similarities between Lingle and Kelo, why might press coverage of government takings have changed so dramatically from April to July 2005? Two major differences in the cases involve their perceived victims and voting outcomes. We can explore the salience of these features as expressed by the content of media coverage across the three phases. To do so, I conduct searches for mentions of dissent and indicators of victims in property rights struggles. I use the Supreme Court’s written opinions in Lingle and Kelo as a guide for how victimhood may be framed, focusing on invasions or intrusions of personal property as well as mentions of vulnerable populations such as the poor, the weak, and minorities.7 As a percentage Page 62 →of all property rights coverage, mentions of victims occur least frequently in Phase 3, after Kelo limits the rights of homeowners to fight eminent domain. At the same time, mentions of judicial dissent alongside discussions of property rights spike in Phase 3, as figure 4.3 indicates. The media’s apparent attentiveness to judicial dissensus following the Kelo decision does not verify that the dissenters’ arguments dominated coverage. Nor does it explain why the case drew so much notice. But it does indicate that the press pays attention to dissent on the Court. Between the end of May and the end of June, the media’s coverage of property rights expanded dramatically, with discussion of the Fifth Amendment receding and that of eminent domain swelling. Against the backdrop of this significant change, Americans were also learning more frequently about judicial disagreements surrounding property rights.

Content of Coverage A number of explanations may account for how judicial dissensus sets apart coverage of Lingle and Kelo. One account suggests that the cases receive identical portrayals but that dissensus raises the salience of Kelo alone. This account runs contrary to the predictions of dissensus dynamics. And as we have already seen, it is also inaccurate, for the substance of property rights discussions changes alongside the volume of coverage across the three phases. Another explanation is that the relationship between press coverage and judicial dissensus is in this case spurious. Little evidence supports this explanation, given the fact that the Supreme Court decided by large majorities other property rights issues in its recent history, none of which led to a dramatic reversal in press coverage. Lingle and Kelo themselves differ minimally in the nature and direction of their decisions, occur within the same time frame and political climate, yet vary dramatically in their media treatments. A third explanation sees judicial dissensus as an important signal that only indirectly influences press coverage of an issue. In this scenario, consensual and dissensual decisions trigger different types of coverage, but the Court’s opinions do little to shape the content of this coverage. In the case of Kelo, the specific arguments raised by O’Connor and other dissenters would be much less relevant than the fact that they dissented in the first place. Finally, another explanation suggests that the press draws directly on the frames and arguments supplied by dissenters when shaping its coverage. For example, media outlets may choose to highlight criticisms of the majority’s definition of “public purpose” in Kelo precisely because the dissenters offer this argument. Page 63 →Evidence suggests that dissent in Kelo exerts both a direct and an indirect effect on coverage of property rights. The frames used by the dissenters have some effect on the content of media coverage, but the

impact of these frames is guided by their newsworthiness. O’Connor’s warning about the vulnerability of property rights in the wake of Kelo strongly influences coverage. Other frames track closely the debate between the majority and dissent, but some do not. For example, Thomas raises two key considerations—an originalist interpretation of the Constitution and the effects of eminent domain on minorities and the poor—but only the latter gains traction in the media. Based on this evidence, it appears that the specific frames used in written opinions gain traction to the extent that they help the press craft simple yet dramatic coverage. So while the voting outcome in Kelo allowed the press to offer a variety of counterframes to rebut the Court majority, reporters only drew on the most newsworthy arguments raised by the dissenters. Phase 1: Pre-Lingle Prior to Lingle, the national press offered a steady but small stream of property rights coverage. This coverage ranged from detailed examinations of eminent domain to appeals regarding property rights to reports on property law and courts; from invocations of American values to unrelated discussions of the Fifth Amendment.8 The modal story gives only passing mention to a vague notion of private property, with little discussion of the concept. A Washington Post story from May 20 is typical. The article discusses federal efforts to reshape the Endangered Species Act, noting a need to balance private property rights with habitat protection.9 A May 6 segment on Fox News illustrates even more clearly the offhand references to private property that occur in the pre-Lingle phase, with a guest on The Big Story joking that “there’s no controversy about Marxism, except for the sticky bit about no private property.”10 News coverage often reinforces the assumption that private property rights are sacrosanct. According to Fox News host Bill O’Reilly, “Seizing private property after an American dies is simply socialism.”11 Such straightforward invocations of property values make up a large portion of pre-Lingle coverage. Nearly half of all articles in this phase talk about private property and the average story includes 13 related words (i.e., home, right, own). The frequency of private property discussions here is not out of the ordinary, however, as this frame dominates coverage even after the Supreme Court’s rulings in Lingle and Kelo. It is more notable, in fact, that the press does little to flesh out its portrait of property rights in more detailed terms. Page 64 →Two other important frames play a large part in Phase 1 coverage. Discussions of eminent domain and the Fifth Amendment regularly take place. Both topics come up more frequently than in the post-Lingle phase (though Kelo returns the issue of eminent domain back to the media’s gaze). Most of the coverage of the Fifth Amendment during this period, however, has little to do with property rights and more to do with the Michael Jackson child molestation trial. (Cable news, in particular, paid careful attention to the witnesses’ frequent invocations of the right against self-incrimination.) Eminent domain is another matter. News organizations published 18 stories on the topic that deal with a range of projects and controversies. Most of these stories are impressively detailed. The Washington Post offered a portrait of the Kelo case.12 Other articles focus on projects using eminent domain in Brooklyn; Manhattan; Salina, N.Y.; and Washington, D.C.13 Consider one article, “Fighting the Power to Take Your Home,” which explores the process for condemning homes. The piece catalogs the emotional reactions of homeowners to eminent domain, characterizing the power as commonly used. Though the story raises some concerns about the tactic—and mentions the impending Kelo decision—it also depicts takings law as fairly straightforward. “Most people accept the practice of eminent domain, even if they consider it unpleasant and distasteful,” the article notes. According to one lawyer, “Without eminent domain, we’d hamstring the general public’s interest at the expense of one individual.”14 Despite the conventional wisdom that Kelo shone a spotlight on a little-known area of the law, then, eminent domain issues were not entirely obscure in the spring of 2005. In fact, fully one-quarter of the stories involving

property rights in the mainstream press discussed this government power. But in the pre-Lingle phase, the media depicted property rights in vague and uncontroversial terms. Private property was lauded and eminent domain recognized, but very few reports detailed dramatic struggles between states and property owners. Against this backdrop, the Supreme Court prepared to release its May 22 opinion broadening the government’s ability to conduct regulatory takings. Phase 2: Post-Lingle, Pre-Kelo The Court’s Lingle decision passed with little notice in spite of the fact that it violated a popular presumption about the imperviousness of property rights. Though existing models of framing suggest that culturally incongruent frames like the one espoused by the Court—that the government need Page 65 →not demonstrate a legitimate interest in taking property—should activate counterframes in the press (Entman 2004), the press did not frame Lingle as a violation of American values. In fact, the few stories that commented directly on the ruling expressed deference to the Court. According to dissensus dynamics, the justices fostered favorable coverage by achieving unanimity. The New York Times first reported on the decision on May 24 in a six-paragraph article, noting, “In a unanimous ruling, the court clarified its private property precedents.” The article draws heavily on O’Connor’s majority opinion in describing the ruling as lifting confusion in takings jurisprudence.15 Five days later, the Washington Post editorialized that the Court “unambiguously repudiated a dangerous doctrine it had articulated in 1980, a doctrine with horrid implications for environmental and other regulatory enforcement.” The Post relies on the language of the O’Connor’s opinion in its analysis and largely focuses on the ruling’s apparent legal implications. The story gives only passing mention to the rent cap on Chevron’s Hawaii filling stations that sparked the lawsuit.16 In another article, law professor Richard Garnett praises the Supreme Court’s ability to achieve unanimity, contending that despite the fact that people focus on controversy, the Court sometimes achieves consensus in difficult-to-decide cases.17 Garnett’s offhand characterization of Lingle as a complicated legal matter belies the Times and Post coverage portraying the decision as self-evident. It does, however, echo much of the legal literature on the case. In fact, there are three reasons that the ruling might have fostered greater attention and controversy in the media. First, the outcome of the case surprised some legal observers (Fletcher 2006), which might have drawn reporters’ and editors’ attention under the right conditions. Second, many scholars view the ruling as one of the most significant in takings law (Barros 2006; Baron 2007). They argue that by discarding the “substantially advances” test, the Court fundamentally altered three decades’ worth of limits on the takings power, with major ramifications for property owners. News organizations could have used the Lingle ruling as an opportunity to explore judicial limits on property rights. They could have further framed the case as a precursor to the Court’s consideration of “public use” requirements in Kelo. And finally, the decision appeared to violate the common assumptions about the secure nature of property rights. Some press outlets might have chosen to explore further the implications that the government does not need to feign a legitimate interest in taking property, that it can enact rent caps on business whenever it deems appropriate. Instead, Lingle passed with little fanfare. But favorable coverage of the ruling (which the Court helped to ensure by achieving consensus) left other Page 66 →media narratives about property rights undisturbed; they remained largely unchanged from the predecision phase. Once again, mentions of private property frequently occurred—in nearly 60 percent of the articles in the Phase 2 sample. Some invocations of the term were purely descriptive. At other times, reports emphasized respect for property rights. A Washington Times op-ed typifies such a story, arguing, “Without a responsible rule of law that protects private property rights, incentives and free markets, there can be no sustained development.”18 A guest on NPR depicts the violation of property rights as unethical.19 With the Michael Jackson testimony ending, references to non–Takings Clause portions of Fifth Amendment declined. Similarly, the media focused less on potential threats posed by eminent domain in the post-Lingle phase. As of June 22, 2005, the media had made little effort to portray property rights as under threat despite the Supreme

Court’s significant enhancement of the takings power in Lingle. Eminent domain and regulatory takings constituted a small minority of property rights news in Phase 2. Rather, most coverage continued to frame private property as a bedrock principle of American society, albeit one subject to reasonable limits imposed by the government and the judiciary. Phase 3: Post-Kelo The ruling in Kelo v. City of New London altered dramatically the media portrayal of property rights. Without question, Kelo’s influence can be traced in part to its violation of strong cultural norms and values regarding property (Sax 2005; Lee 2006; Nadler, Diamond, and Patton 2008). But this explanation is not enough on its own, for Lingle similarly curtailed protections for private property. Why, then, did Kelo change the way in which the media discussed property rights? Initial reports in the wake of a news event may be quite influential. In the case of judicial decisions, this influence is magnified. Legal rulings present difficulties to reporters who must trace a case history, interpret complex jargon, parse legal implications, and explore societal aftereffects in a compelling and entertaining manner (Greenhouse 1996). In most cases, a small cadre of reporters who have legal training and years of experience working at the Supreme Court shape initial coverage of a ruling (Davis 1994). Their reports help structure a press narrative over time. Kelo provides evidence for this pattern. The initial reports on the decision focus more on its legal meaning and offer more diverse viewpoints than many subsequent stories, but they also highlight many of the same arguments that recur in later coverage. Page 67 →In the first two days after Kelo’s announcement, more than 40 stories considered the case and its ramifications. Many news sources explained the ruling as the result of a broad interpretation of public use—that courts often read the Fifth Amendment to favor the government over individual property owners. The public use frame reflects positively on the ruling in Kelo, as it echoes closely the majority’s analysis. It further gives the impression that the types of takings endorsed by the majority were not altogether novel but rather constituted a continuation of a long-recognized, rarely used government power. Indeed, some reports emphasized that the Court deferred to lawmakers concerning eminent domain because the public helps constrain their interpretation of “public use.” According to CNN legal analyst Jeffrey Toobin, And basically what the majority said is, look, this belongs in the political system. If you don’t like how your mayor is exercising the power of eminent domain, vote him out of office, but the Constitution is no bar to it. We’re not going to get involved in micromanaging how mayors run their cities and how they exercise all their powers. The Constitution says—the literal words in the Constitution are “public use.” Eminent domain can be used for a public use. Under the majority’s view, public use includes some kinds of private development that benefits the community.20 These reinforcing frames—that the Constitution allows takings for public purposes, broadly defined, and that courts defer to legislative judgments about those takings—occurred with some regularity in both the initial coverage of Kelo and in Stevens’s majority opinion. Though it is not the place of the Court to adopt a new “bright-line rule” about the meaning of public use, according to Stevens, “nothing in our opinion precludes any State from placing further restrictions on its exercise of the takings power” (489). Had the media limited its coverage of Kelo to this argument, the portrait of the case would not have differed markedly from that of Lingle. Instead, however, a preponderance of the coverage framed the decision in less deferential terms (see table 4.3). Stories highlighted the Court’s redefinition of public use as public purpose, the hardships the ruling would impose on homeowners, the vulnerability of private property in a post-Kelo society, the expansion of government power, and the privileging of wealthy interests over those of the poor. Almost all of these frames were new to the discussion of private property since the spring of 2005. Many of these arguments originated in the opinions

released by the Court (particularly those of the Page 68 →dissenters)—that is, the redefinition of public use, the vulnerability of property, the expansion of government power, and the harm done to the poor. The media readily employed the language of dissenting opinions to make these arguments (although some of the most compelling language received a disproportionate share of the attention). Conversely, the powerful human-interest frame gained traction mainly because of the sympathetic nature of the victims in Kelo. All of these frames took shape in the days after the ruling and continued to guide the discussion of the case throughout Phase 3. The Hardships on Ordinary Homeowners A number of stories in Kelo’s aftermath employed a human-interest angle, exploring the effects of the ruling on the petitioners and other homeowners. Susette Kelo, in particular, offered a gripping personal story. “Protecting her Victorian dream house from the urban renewal bulldozer was always going to be an uphill fight,” USA Today notes. Kelo laments, “I am sick. Do they have any idea what they’ve done?”21 The media’s attention to sympathetic homeowners was not limited to the petitioners, however. A number of other reports explored the ramifications of the Court’s decision for a wide variety of property owners. For instance, Page 69 →CNN aired an interview with the son of another New London property owner who made plain the stakes surrounding takings: “The house, you know, is not a commodity to my father. My father came over from Italy in 1962 and, you know, land to him means that he’s rich and that he’s got all the gold in the world. You take that away from him, and you know, he has nothing.”22 Other reports followed suit, cataloging the battles of ordinary homeowners to maintain their property in the face of government power. For example, the use of eminent domain for a baseball stadium project in Washington, D.C., had been praised by newspapers in May 2005 but earned some unflattering press in “Half St. Citizens Continue to Fight Ballpark Bullies.”23 Table 4.3. Post-Kelo Framing of Property Rights Part of Dissenting Opinions?

Frame Hardships on ordinary No homeowners Redefinition Yes; of public use “The Court has erased the Public Use Clause from Our Constitution” (Thomas) Yes; Vulnerability “All private property is now vulnerable to being taken and transferred to another owner” of property (O’Connor) Expansion of Yes; government “An external, judicial check . . . is necessary if this constraint on government power is to power retain any meaning” (O’Connor) Privileging of wealthy interests

Yes; “These losses will fall disproportionately on poor communities” (Thomas)

The human-interest hardship frames can be traced most directly to the sympathetic nature of the victims in Kelo. These frames are powerful indeed and likely had a major impact on the public’s negative reactions to the decision (Nadler, Diamond, and Patton 2008). Nonetheless, they represent just one change in press coverage of private property post-Kelo and the only one directly traceable to the identity of the parties in the case. Other private property frames were related more directly to the Court’s voting coalitions. The Redefinition of Public Use as Public Purpose The day after the release of the Kelo decision, Matt Dery, the son of the petitioner, spoke to NPR, saying, “Yes, the Constitution does allow the taking of private property for public use, but redevelopment is not a public use.” Dery gave voice to an argument that had animated the history of Kelo and informed the analysis of

its dissenters but had seldom appeared in the media’s discussions of private property. Other commentators also took issue with the majority’s definition of public use.24 Not surprisingly, the most comprehensive analysis of the public use requirement came from reporters with extensive experience at Supreme Court. These included Linda Greenhouse at the New York Times, Charles Lane at the Washington Post, Guy Taylor at the Washington Times, Nina Totenberg at NPR, and Jeffrey Toobin at CNN. Their detailed coverage of the Court’s opinions regularly informed other pieces. Totenberg and Greenhouse put the dispute over the public use at the center of their analyses, with Greenhouse alternately quoting from Stevens’s broad reading of the clause and O’Connor’s objections. “The decision was a clear defeat for the long-term effort by Chief Justice Rehnquist and Justice Scalia to limit government control over private property.”25 The story also uses the powerful language of Page 70 →Thomas’s dissenting opinion: “The Court has erased the Public Use Clause from our Constitution.” Indeed, nearly all discussion of public use in the press can be traced to the Court’s controlling and dissenting opinions. Compared to other concepts, “public use” and its associated terms occurred with a high degree of frequency in Kelo (and in the dissenting opinions in particular). The media’s use of this frame in the postKelo phase followed from the fact that the dissenters decried the “redefinition” of public use. The Vulnerability of Private Property The idea that Kelo suddenly made private property vulnerable shattered the preruling myth of its inviolability. The tenuous hold that Americans had on their homes and other possessions became a central story line in the press. Consider the progression of print headlines regarding eminent domain: “The Limits of Property Rights” on June 24, “Your Castle No More” on June 27, “Uninformed Expropriation” on June 29, “Homeowners Lose on Property Rights” on July 1, “Your Land Is My Land” on July 5, “Home, Seized Home” on July 18.26 Many media outlets focused on the impact of Kelo on the ordinary homeowner, but others explored the its implications for the poor or even for large retailers.27 This coverage differs from the press characterization of Lingle’s effect on property rights. Why did Kelo suddenly cause reporters to suggest that the Court had attacked private property? The vulnerability frame comes directly from O’Connor’s dissenting opinion, where she argues, “All private property is now vulnerable to being taken and transferred to another private owner.” This powerful and arresting argument had a major effect on the way in which the press viewed the decision’s impact. Eight stories quoted this phrase directly, and many others built on the argument. In one of the earliest reports on the ruling, CNN warns, “You never know when your home or business is going to be targeted.”28 The frame retained its power throughout most of July 2005, with stories raising the vulnerability concern again and again. The influence of O’Connor’s dissent suggests that justices may be particularly powerful in shaping the responses to their opinions if they craft a compelling portrait of them in a way to engage the press and public. Justice Ruth Bader Ginsburg’s oral dissent in Ledbetter v. Goodyear Tire & Rubber Co. had a similar effect, leading to a backlash to the ruling and paving the way for the Lilly Ledbetter Act in Congress (Toobin 2012, 80–81). The media also noted an interesting twist: an effort to seize the home of Page 71 →Supreme Court Justice David Souter in response to Kelo. The plan recommended building a Lost Liberty Hotel to increase tourism and tax revenue on the site owned by “someone largely responsible for destroying the property rights of all Americans.”29 CNN, Fox News, MSNBC, the Washington Times, and USA Today covered the effort throughout July 2005. The Expansion of Government Power The theme of property vulnerability coupled nicely with two other arguments that regularly surfaced in press coverage of Kelo: the decision represents a dramatic expansion of government power, and it privileges the interests of the wealthy. The ideological divide between the justices became most salient when the government

power frame was used. NPR’s legal expert characterized the ruling as “the sort of expansive definition of government that economic conservatives generally object to.”30 The Washington Times paid careful attention to the matter. In its initial report on the ruling, it mentions the dissenters’ argument that the Court abandoned a long-standing limitation on government power. The accompanying editorial calls the decision a victory for big government. Three days later, the paper argued that calling the decision “corporatist or fascist is no mere epithet. It designates a system that maintains the veneer of property while political authorities have extensive powers to limit rights in the name of economic planning.”31 Guests on cable news programs regularly picked up on these arguments. The characterization of Kelo as sanctioning an all-powerful government, however, was misleading. Rather than endorsing government overreach, the Court majority made clear that its decision followed the Constitution and exhibited deference to the will of elected representatives. The decision portrayed the courts as a neutral arbiter in the fight over the constitutionally mandated power of eminent domain. Furthermore, according to the Court, the public sharply constrains lawmakers’ ability to abuse this power (Kelo v. City of New London, Hawaii Housing Authority v. Midkiff). Congress, too, failed to cooperate with the news portrayal of a unified largegovernment apparatus. A bipartisan group of representatives called Kelo a serious mistake and pledged to block federal funds for state and local governments using eminent domain.32 Though some news reports blamed the expansion of government power on a coalition of the Supreme Court’s five most liberal justices (who made up the majority in Kelo), this depiction failed to gain widespread traction. In fact, though the decision arguably increased government power, it did Page 72 →not fit soundly an ideological narrative. For example, the ostensibly liberal majority had sided not only with the government but also apparently with the interests of large corporations over those of ordinary Americans. And Thomas, perhaps the most conservative justice, used his dissent to discuss Kelo’s ramifications for minorities and the poor. The Privileging of Wealthy Interests Commentators demonstrated particular concern that the redefinition of public use as public purpose would make legislatures more likely to upgrade property by transferring it to businesses. The Court majority and dissenters debated precisely this point in their opinions. According to Thomas, Allowing the government to take property solely for public purposes is bad enough, but extending the concept of public purpose to accomplish any beneficial goal guarantees that these losses will fall disproportionately on poor communities.В .В .В . It encourages “those citizens with disproportionate influence and power in the political process, including large corporations and development firms” to victimize the weak. (Kelo v. City of New London, Thomas dissenting, 522) Though the media demonstrated less outright concern for the rights of the poor than those of “ordinary” homeowners in Kelo’s wake, the decision’s unequal impacts did not go unnoticed. In fact, many of the press’s claims about the use of eminent domain in blighted areas were quite dramatic: “This is only the beginning of fleecing the poor to give to the rich.” “Our new city-kings can take any property they choose—particularly if they take from the poor and give to the higher-tax paying rich.” “Several justices said this is a ruling for the wealthy and against the poor.” “This situation, it seems to me, is ripe for abuse by rich developers preying on the poor, the weak.”33

The Impact of Dissent Lingle and Kelo received dramatically different coverage despite their similarities. In both cases, the Court upheld a broad reading of the Takings Clause, making it easier for governments to seize private property for economic development without demonstrating a legitimate interest in doing so. Prior to these rulings, many news outlets portrayed property rights as a bedrock Page 73 →of American society. A few stories drew attention to eminent domain condemnations, depicting them as a rarely used tool for the good of society as a whole. Though Lingle did little to upend the general tenor of property rights coverage, the media used a frame of deference to characterize

the decision. But Kelo introduced a host of new frames for the coverage of property rights, and many of them raised concerns about government takings. Three major differences drove the divergent press reactions to the decisions. First, Kelo was a more inherently newsworthy case. Analysts found its popular resonance among its most remarkable features. But for reasons that have been less well understood, the substance of its coverage also differed dramatically from that of Lingle. Two factors played a central role in shaping the depiction of property rights in the cases. On the one hand, the nature of the victims fostered more skepticism about Kelo, as can be seen most directly in reports that discussed the hardships on ordinary homeowners that the decision would bring about. Americans learned for perhaps the first time that the Takings Clause could have an immediate, deleterious impact on their lives. The public’s negative reaction to the decision may have been largely animated by feelings of threat aroused by the prospect of losing one’s home (Nadler, Diamond, and Patton 2008). Yet the media shaped its coverage not only around the case’s victims but in response to the closely divided nature of its outcome. A five–four vote set Kelo apart from the remainder of the Supreme Court’s Takings Clause rulings. And absent Kelo’s dissenters (and two powerful dissenting opinions), many reporters might have continued to portray property seizures as rare, limited, and for the benefit of society overall. But the dissenters affected the coverage of property rights both indirectly (by drawing attention to the case and supplying fodder for critics who disagreed with the decision) and directly (by providing frames that resonated in the press). In fact, the media was roughly three times more likely to mention judicial disagreement in its property rights coverage post-Kelo. And four of the frames employed by the dissenters in their written opinions came to dominate coverage of the issue in June and July 2005. Figure 4.4 provides more evidence that the justices’ opinions directly influenced the content of coverage. To create this figure, I analyze the written opinions in Kelo to develop keyword searches for the most prevalent frames they employ (issue-specific frames, legal frames, originalism, deference, and so forth). These are plotted along the x-axis. The figure depicts the framing struggles in which the majority and dissent engaged (the light line) and changes in the media framing of property rights from pre- to postruling (the dark line). The two lines track each other closely, suggesting that the written Page 74 →opinions in Kelo had a direct as well as indirect effect on the media’s framing of property rights. Fig. 4.4. Framing of Property Rights in Kelo and in the News. (Note: Correlation between series is 0.67 with pvalue of 0.005.) And what of the nature of the Court’s voting coalitions? In the case of Kelo, little evidence suggests that the media paid special attention to partisan politics in its discussion of property rights in spite of the fact that the more liberal and conservative members of the Court divided over its outcome. One reason why the ideological rupture might have failed to resonate involves the unique nature of the decision, which, according to the dissenters, saw liberal justices favoring an economic development plan over the interests of the poor. To the extent that these actions were inconsistent with existing narratives about liberalism, the media gave them minimal attention. Indeed, after the ruling, many liberal and conservative politicians joined together in attacking the Court’s definition of property rights. A Post-Hoc Account Divergent press coverage of Lingle and Kelo provides evidence in favor of dissensus dynamics, but it also furnishes a more complete portrait of how the press uses negative frames to characterize dissensual decisions. Page 75 →To craft simple, accurate, and dramatic accounts of legal controversies, news organizations draw on the straightforward information that the Supreme Court communicates. The justices represent the highest level of elites in legal controversies, and, as other research shows, press coverage tends to highlight the range of perspectives and frames used by such elites (Bennett 1990; Bennett and Livingston 2003). So when a decision is consensual, reporters rely on the analysis of the justices alone. Reporters explain the Court’s reasoning and

highlight the outcome, portraying it as uncontroversial. But when a decision is conflictual, reporters use the opportunity to discuss multiple competing perspectives on the legal controversy. They explain the findings of the Court majority but also emphasize the criticisms raised by dissenters. But while dissent powerfully shapes press coverage, the effects of dissenting opinions vary. The specific arguments raised by dissenting justices receive coverage to the extent that they are relevant and resonant. Other criticisms that the dissenters fail to raise may nonetheless gain news coverage as a consequence of the press’s willingness to employ outside sources following dissensual decisions. The role of dissenters, then, is to cause reporters to offer competing perspectives in their coverage of Supreme Court rulings, based on newsworthy frames from inside and outside the courthouse. In 2005, the Supreme Court’s property rights jurisprudence helped to illustrate this process. The media’s reports on Lingle framed the decision as an unremarkable application of the Takings Clause, replete with recognition that the government at times possessed the power to sharply regulate and/or seize private property. Because the press made little effort to give voice to Lingle’s critics, the decision passed with minimal notice. It did little to change news coverage of property rights in the United States. But judicial division over Kelo allowed the press to highlight criticisms of the Court’s application of the Fifth Amendment. The media may have ignored Thomas’s textualist critique of the decision, but it seized on O’Connor’s contention that the Court had made all personal property vulnerable. Over time, the media built on this warning by emphasizing the human-interest consequences of the ruling. However, Kelo’s extraordinary resonance cannot be traced simply to divisions between the justices themselves. Rather, a confluence of factors—including O’Connor’s evocative warnings, the sympathetic nature of the victims in the case, and the willingness of other critics to attack the Court—raised alarm that the decision violated “bedrock” private property rights. Kelo continued to resonate. Before the decision, most national news coverage portrayed property rights in vague terms as a sacred part of the political system. One year later, however, discussions of the issue regularly Page 76 →explored the intricacies of government takings. More than 60 percent of the stories published on the issue between June 23 and July 22, 2006, continued to mention eminent domain controversies.34 Judicial dissent in Kelo had put forth an arresting account of property rights abuse that continued to shape coverage long after the decision was announced. In this sense, it was truly unique. But the case also illustrates how the media responds to judicial voting coalitions in covering the Court. Chapter 5 explores whether leading U.S. newspapers shape their coverage similarly, with an eye toward judicial votes across a range of high-profile cases from the past three decades.

Page 77 →

Chapter 5 “I Respectfully Dissent” A Test of Dissensus Dynamics in Newspaper Coverage According to one veteran print journalist, “A useful story about a Supreme Court decision, in my view, is necessarily interpretive.В .В .В . Readers also need to know the content of the decision, what the decision means, how the case got to the Court in the first place, what arguments were put to the Justices, what the decision tells us about the Court, and what happens next. Not all of these elements are necessary in each story, and not all of the questions can be answered in every case” (Greenhouse 1996, 1545). This perspective suggests a compelling alternative to the dissensus dynamics model—that newspapers follow few systemic rules in covering a wide range of judicial decisions. After all, each case features a unique set of facts, claimants, legal principles, and political implications. Can we really expect news organizations to use simple voting cues in shaping their coverage of diverse rulings? Do similar principles guide coverage of five–four decisions involving both abortion rights and tax law simply because of the votes of the justices? Do the same factors that shape coverage of the Court’s property rights jurisprudence also influence reporting on other decisions? To be sure, the media highlight distinct features of each Court decision. They employ different frames to discuss the various issues on which the justices rule. At the same time, however, journalists on the Supreme Court beat operate in a relatively stable environment. The Court has followed a similar set of procedures in handing down decisions throughout its history (Epstein and Walker 2010). It releases a fixed set of information that includes written opinions and the case syllabus (Davis 1994). Reporters are largely at the Page 78 →mercy of the Court, lacking knowledge about when it will rule and access to speak with the justices (Greenhouse 1996). Reporters remain under considerable constraints, required to simplify decisions in an entertaining and timely manner. As a result, many of the same ingredients shape journalism at the Supreme Court for various news organizations across cases and time.1 And some reporters who work on the Supreme Court beat confirm that they sketch their coverage using a consistent set of parameters (Davis 1994, 83). Dissensus dynamics predicts that we can trace the consistent features of press coverage to the constraints placed on journalists and news organizations. To the extent that divisions among the justices are observed, reporters become more likely to highlight criticisms of a ruling raised by both the dissenters and other elites. This chapter provides the first statistical tests of the model by exploring how three leading newspapers have covered highprofile Supreme Court decisions in recent decades. Newspaper coverage provides an interesting test case for dissensus dynamics. On the one hand, organizations like the New York Times provide some of the most comprehensive reporting on the Court. Because many of their judicial reporters have followed a large number of cases (Davis 1994), these reporters may more readily rely on formulas in shaping their coverage. On the other hand, these journalists also have unparalleled knowledge. They are well equipped to ignore simple cues such as voting signals, instead drawing on understanding of the law and on contacts with sources to sketch a unique portrait of each decision. To the extent that these journalists have expertise in judicial affairs, they may rely less on simple pieces of information in fashioning their coverage, making the newspaper context a difficult test of the theory. Nonetheless, evidence demonstrates that the New York Times, Washington Post, and Washington Times use judicial votes to guide their coverage of the Supreme Court, presenting more favorable portraits of rulings reached by large and ideologically diverse majority coalitions. This is some of the first evidence to demonstrate the systematic nature of reporting on the Supreme Court beat.

Testing Dissensus Dynamics in the Print Media This chapter presents a test of the dissensus dynamics model using newspaper coverage of judicial rulings taken

from a random sample of all high-salience decisions handed down by the Court between 1981 and 2008. The universe of high-salience cases is based on an index from Epstein and Segal Page 79 →(2000) that identifies rulings that received a front-page story in the New York Times the day following their official announcements.2 I focus on high-profile rulings because most Americans pay attention only to a limited number of salient judicial decisions. The sample consists of 88 cases.3 For each case, I conduct a Lexis-Nexis search of New York Times, Washington Post, and Washington Times4 coverage of the decision by using a keyword that includes either the name of the petitioner, or (in cases in which the petitioner had a common name, e.g., the United States) the name of the respondent. This selection of sources includes three leading national news organizations that cover the Court extensively, provide information to many other local papers, and help shape public understanding of judicial action. Furthermore, their editorial boards represent an ideologically broad spectrum of opinion (Groseclose and Milyo 2005; Peake 2007; Ho and Quinn 2008). I read and coded all articles (545 total) published within 30 days of a ruling and focused on the Court’s decision in the relevant case. Dependent Variable: Decision Support The chapter offers a broad conceptualization of how media coverage varies between more and less favorable across a wide range of Supreme Court rulings. The cost of doing so involves a loss of issue-specific detail. But the benefit involves a better understanding of how issue-independent features of judicial decisions impact their reception in the press. This may have important consequences for both the political communication literature, which largely overlooks the judiciary (e.g., Bennett 1990; Entman 2004), and for public opinion scholars who seek to understand more generally the conditions under which the Court can persuade (e.g., Hoekstra 1995; Johnson and Martin 1998). To characterize media coverage in broad terms, I focus on two key dimensions: frames used and tonal content. The former is defined as the extent to which each published article explains a decision, gives voice to justices in the majority coalition, and mentions supporters vis-Г -vis critics of the decision. The latter is defined as the extent to which each published article explicitly or implicitly expresses approval of a decision. These article level measures are indexed and aggregated across all coverage of a case. The ruling-centered measure of decision support captures the degree of favorable coverage afforded any single ruling across the entirety of its coverage. The measure has strong predictive and discriminate validity properties, although these are Page 80 →offset by less-than-ideal reliability properties. To remedy these shortcomings, I take additional steps described in appendixes A and B, and I confirm the robustness of all major empirical results with an additional, high-reliability measure of coverage content (discussed later in the chapter). In my analysis, I utilize OLS regression that includes robust standard errors since sampling weights are employed. However, there is also reason to expect that the DV may include unobserved variation at the lower and upper bounds of 0 and 1, in which case a tobit model would be most appropriate. Therefore, I confirm the robustness of all empirical results with tobit models; findings remain substantively unchanged using the two approaches. Independent Variables from Dissensus Dynamics Dissent

Using the Supreme Court Database (SCDB, Spaeth 2010), I focus on conflict between the justices in a given case, including the absolute size of the majority coalition (majority votes). Because majority and minority coalition sizes exhibit an almost perfectly inverse relationship, I expect that as the number of majority votes decreases so, too, will decision support in the press (the dissent hypothesis). I also consider whether media coverage responds to other indicators of dissent, including whether a decision is unanimous, with no dissenting justices, or whether it is decided by a five–four vote. Ideological Diversity

I expect that news outlets pay attention not only to the presence of dissent but also to the ideological makeup of the justices in the Court’s voting coalitions. Martin and Quinn (2002) and Bailey (2007; Bailey and Maltzman 2009) provide widely used ideal point estimates for every Supreme Court justice in every year of his or her tenure.5 I employ these estimates to characterize a case’s majority and minority voting coalitions based on the year it was decided. To measure ideological diversity, I take the absolute value of the distance between the most liberal and most conservative justices in the majority coalition, with the expectation that rulings with ideologically diverse majorities will garner more sympathetic coverage (the ideological diversity hypothesis).6 Page 81 →Independent Variables from Alternative Hypotheses Nonjudicial Elites

One important alternative hypothesis to dissensus dynamics states that political elites shape coverage of rulings with which they disagree. To test this alternative hypothesis, I compare the ideological preferences of leading political actors with the ideological substance of Court rulings using Bailey’s ideal point estimates, which are comparable across institutions and time. I construct the decision-president distance measure by taking the absolute value of the distance between the median member of a ruling’s majority coalition and the president. Similarly, the decision-House distance captures the absolute value of the ideological preference difference between the median members of the Court majority and the House of Representatives.7 These measures allow me to test whether other political elites’ disagreement with rulings shapes the media coverage afforded them. Political Context A second alternative hypothesis to dissensus dynamics suggests that the media offers more unfavorable coverage of rulings when it perceives interinstitutional conflict. So, for example, any high-profile decision handed down by a liberal Court is likely to receive negative coverage when conservatives control the other branches, since this scenario allows the press to highlight critics of the Court with ease. Using Bailey’s institutional ideology scores (2007; Bailey and Maltzman 2009), I create measures of the Court-president and Court-House distance, with larger values indicating interinstitutional preference differences. And I control for the presence of divided government when a decision is released. Issue Controversy and Interest Group Participation A third alternative explanation to dissensus dynamics is that certain cases and issues will be inherently divisive, garnering critical press attention even when Supreme Court justices reach a consensus in ruling. I employ a measure of controversy, adapted from Blake and Hacker (2010), based on the issues on the Court’s docket that receive disproportionate media coverage (BH controversy).8 A second measure of controversy is the presence of lower Page 82 →court dissenters in a case history. This measure, available in the SCDB, captures the existence of dissent in a lower court case being reviewed by the Supreme Court. The measure allows me to control for certain cases or issues that may be inherently controversial as they make their way through the judicial system. The measurement of interest group attention to cases is also essential to explore the determinants of coverage content. The most common and powerful way for interest groups to signal their interest in cases that come before courts involves filing “friend of the court” briefs, also known as amicus curiae (Epstein 1991; Caldeira and Wright 1998). These briefs are not neutral sources of expertise but rather pieces of advocacy for parties before the court, “urging the justices to rule in favor of one litigant over another” (Collins 2004, 808; see also Spriggs and Wahlbeck 1997, 341; Kearney and Merrill 2000, 841–42). As a result, there are reasons to expect that amicus participation demonstrates not only interest group attentiveness to a case but also precisely the type of strong advocacy perspectives that make these groups likely to respond to judicial decisions in the court of public opinion. I use a measure of the number of amicus curiae briefs filed in a given case to explore whether interest groups have the ability to shape the coverage afforded to high-profile Supreme Court decisions (amici n).9 Decision Characteristics

Finally, I also test whether specific characteristics of the ruling (beyond the voting outcome) affect coverage. An alternative hypothesis to dissensus dynamics is that the ideological characteristics of a decision affect coverage. For example, the press may be particularly unsympathetic to rulings decided by conservative majority coalitions, particularly when these judges employ textual analysis that lacks the simplicity and drama the media crave. To account for this possibility, I control for the majority median of a ruling using Bailey estimates to measure decision content. This measure of ideology is among the best available ways to characterize Supreme Court rulings (see Carrubba et al. 2012). I also control for whether a decision altered precedent or declared law unconstitutional. These measures are taken from the SCDB, whose coding is based on an analysis of the Court’s written opinions. I further control for end of term rulings (which I define as those announced in June or July) and those written by the chief justice (chief justice authorship). Both of these measures use data on rulings available from the SCDB.

Page 83 →A Summary Portrait of Supreme Court Coverage How do leading national newspapers portray high court rulings? The question seldom has been considered in detail, a problematic fact given important variance in coverage across cases. This variance occurs not only in the volume of reports (which has been explored previously) but also in content. I suggest that the decision support scale provides a meaningful look at the content of coverage. But first, a few words about salience. Newspapers cover landmark Supreme Court rulings with varying degrees of intensity. Some of the most prominent decisions saturate the news, with coverage most prevalent the day after a decision is announced. While papers publish fewer articles on a case over time, it is not uncommon for the most high-profile cases to be the subject of five or more stories in the month after the Court rules. At the other end of the spectrum, even low-impact decisions in the sample receive at a minimum front-page coverage in the New York Times. The average ruling in the sample garners about six unique print articles after its announcement. Figure 5.1 plots the volume of coverage for sampled cases, which has remained fairly constant since the early 1980s. Nonetheless, from time to time, monumental decisions lead to a spike in coverage. Kelo v. City of New London stands out as one such ruling: the 2005 decision caused newspapers to publish a host of stories that examined its implications for property rights in the United States. The press similarly paid careful attention to the abortion rights ruling in Webster v. Reproductive Health Services, printing nearly two stories per day on the case for an entire month. Coverage of Webster unfolds in accordance with a common pattern, where the press shifts over the course of time from a focus on decision mechanics to more analytic perspectives.10 What allows certain rulings to draw the attention of the press? Evidence suggests that decision salience stems in large part from newsworthiness. As the previous chapter demonstrates, media outlets focused on the most dramatic features of the Kelo ruling to shape their coverage. They allowed ordinary homeowners to discuss their personal anguish about the decision and emphasized O’Connor’s striking warning about the ruling’s consequences. In both Kelo and Webster, the press used the rulings to highlight human-interest perspectives, surprise outcomes on the Court, and controversy over the cases in American society. More germane to our purposes, however, is the content of coverage in high-profile cases. Do newspapers largely defer to the analysis set forth by the Court majority? Do they employ alternative frames that reflect poorly Page 84 →on rulings? The decision support scale provides a valid summary measure of content, showing wide variations in coverage of individual rulings. Fig. 5.1. Coverage of High-Salience Supreme Court Decisions, 1981–2008 The Spectrum of Coverage Consider a few Supreme Court rulings that meet with the most positive newspaper clippings. In U.S v. Winstar Corporation (1996, decision support = 1), the Court ruled that Congress violated contracts with savings and loans when it tightened accounting rules, ordering the government to pay damages. In its lead story on the decision, the Washington Post focused on the Court’s reasoning. It quoted a Washington attorney who portrayed the ruling

as self-evident, “a straight contract law case. [The justices] ruled that the government can’t renege on a contract.” According to the story, the government had no comment on the ruling. The Post’s editorial board viewed the ruling as “Costly but Correct.”11 Rumsfeld v. Forum for Academic and Institutional Rights (2006) garners similarly supportive coverage (decision support = 0.944). In the case, the justices upheld federal financing cuts for universities that restrict access to Page 85 →military recruiters. Notably, newspapers found few sources to criticize the decision. Even gay rights groups, which opposed the military’s “don’t ask, don’t tell” policy, saw a silver lining: the ruling would build momentum for a repeal of discriminatory policies. Editorial boards offered similar praise for MillerEl v. Dretke (2005, decision support = 0.875), “an important ruling that reiterates to all courts the importance of keeping discrimination out of jury selection” and a “strike against bias.”12 How does this type of coverage contrast with less favorable reports on a ruling? A small number of judicial decisions raised the ire of print media outlets. They included R.A.V. v. City of St. Paul (1992, decision support = 0.313), Bennis v. Michigan (1996, decision support = 0.194), and Kelo (2005, decision support = 0.098). Unfavorable coverage frequently took the form of editorials that offered direct criticism of the Court.13 But even straightforward news stories portrayed these rulings in an unfavorable light by focusing on arguments raised by the dissenting justices and by outside sources.14 Table 5.1 provides summary statistics regarding coverage of high-profile rulings. On average, decisions receive neutral to favorable news notices, with reporters quoting from or explaining rulings to readers. Coverage of the Court’s landmark decision involving gun rights in D.C. v. Heller is emblematic of how newspapers offer a variety of perspectives on a ruling while remaining largely deferential to the justices’ judgment. The New York Times, Washington Post, and Washington Times published 20 stories on the case. One editorial in the Post offered strong criticism of the Court, calling the decision a “misguided” form of judicial overreach. A number of other stories praised the ruling as a wise and reasonable interpretation of the Second Amendment.15 The vast majority of coverage, however, did not take a strong stance on the ruling but instead deferred to the Court majority while mentioning its critics. So, for example, in the New York Times’s 1,800-word lead story on the ruling, the first six paragraphs describe the legal reasoning behind the decision. The next six paragraphs highlight a dispute between the majority and dissenting coalitions over the text and history of the Second Amendment. The article goes on to review the history of gun rights in the United States and concludes with more detail about how the Court interpreted the “operative clause” of the Second Amendment.16 Despite the controversial issue at the center of D.C. v. Heller, then, its coverage approximated that of the average case in the sample, featuring a diverse set of perspectives counterbalanced by deference to the majority’s written opinion. What factors influence how newspapers cover important Supreme Court rulings? Why do the media portray many rulings in uncontroversial terms Page 86 →while denouncing the justices at other times? And when are Supreme Court rulings most vulnerable to the type of press censure that characterized Kelo?

Results and Analysis A simple model of decision support explores the effects of a fractured Court, which, in accordance with the dissensus dynamics model, should lead to less favorable coverage of rulings. Figure 5.2 shows the actual sample means of decision support by the number of votes in a majority coalition. It suggests that the justices can indeed earn favorable coverage for their decisions when they assemble large majority coalitions. Whereas roughly 80 percent of the coverage of unanimously decided cases presents them in a positive light, less than 60 percent of articles involving closely divided decisions are similarly supportive. Furthermore, the Court majority’s ability to dominate press coverage depends directly on its size. Decisions reached by nine-member majorities receive more support than those of eight-member majorities, which receive more support than those reached by seven justices, and so on. There is nothing in unanimity that changes dramatically the media’s responsiveness to coalition size. Supreme Court journalism appears unique in the sense that reporters respond not to the presence of conflict (i.e., they do not treat unanimous decisions differently than others) but rather to the extent of dissent.

They also, it seems, respond to the nature of conflict. Figure 5.3 plots the relationship between ideological diversity in a ruling (the distance between the most conservative and most liberal members of the majority) and support for it in the news. Broad and ideologically diverse majority coalitions receive the most favorable types of coverage, as expected. But not surprisingly, a strong correlation exists between the size and ideological diversity of majority coalitions (r = .605), raising the question of whether press support for rulings more directly follows from one or the other. Does the Court majority set the terms of coverage in eight–onePage 88 → rulings because reporters defer to the judgments of a large majority coalition or because they defer to the judgments of an ideologically broad spectrum of the Court? To what extent can we even distinguish between these characteristics? Table 5.1. Selected Summary Statistics of Newspaper Coverage Variable Mean SD Min Max Obs Decision support 0.67 0.21 0.10 1 88 Frame dominance 0.71 0.23 0.11 1 88 Supportive tone 0.63 0.21 0.09 1 88 Judicial activism 0.10 0.22 0 1 88 Stories per case 6.19 6.49 1 51 88 Page 87 →Fig. 5.2. Majority Votes and Decision Support Fig. 5.3. Ideological Diversity and Decision Support Table 5.2 illustrates a basic test of dissensus dynamics, where the press pays attention only to the voting signals sent by the justices. The table provides preliminary support for both the dissent and ideological diversity hypotheses, suggesting that the larger and broader the majority voting coalition, the more the Court is able to garner supportive coverage for its decisions. These results hold across a number of indicators of decision support, including most surprisingly whether news outlets include accusations of judicial activism.17 Large majority coalitions act as a bulwark against that charge, even after we control for the substance of their decisions (whether they alter precedent or declare laws unconstitutional). The media’s attentiveness to judicial voting signals indicates that the Court may insulate itself from criticism when it reaches consensus. What other factors may influence the manner in which the press frames judicial outcomes? Perhaps reporters pay attention not only to the signals sent by the Court but also to those offered by other elites. Perhaps certain issues on the Court’s docket will generate critical press attention regardless of a ruling. I test the most powerful alternative explanations to dissensus dynamics: whether elite disagreement, political context, issue area and interest group activity, or ruling characteristics affect coverage (table 5.3). There are reasons to expect that the media might pay attention to each of these features. For example, the press might depend on other elites to determine which Page 89 →legal rulings deserve unfavorable attention (Entman 2004). One might expect that the president affects media support for decisions depending on the extent to which he agrees with them. Presidents may praise the Court when it rules in accordance with their ideological preferences and upbraid it when they differ. They may more willingly criticize the Court when it has a reputation for ignoring their wishes. Or, given the popular standing of the Supreme Court and the risks that criticism entails, they may remain silent. Can leading political figures diminish the persuasive power of the Court with their criticism? To test this possibility, I look at what happens when the Court comes into conflict with political elites. Models 1 and 2 confirm that elite disagreement with the Court does little to alter the media coverage it receives. Conversely, elite disagreement within the Court continues to matter. Table 5.2. Basic Models of Favorable Coverage (1) (2) (3) (4) Decision support Supportive tone Frame dominance Judicial activism

0.032* (0.019) 0.091** Ideological diversity (0.040) R-squared 0.227

0.024 (0.018) 0.094** (0.042) 0.189

0.039* (0.021) 0.088** (0.042) 0.221

в€’0.034* (0.017) 0.053 (0.037) 0.273

N

88

88

88

Majority votes

88

Note: Results are OLS coefficient estimates with robust standard errors in parentheses. Model 4 includes controls for alters precedent and declares unconstitutional (estimates not reported). Similar results with a tobit model. *p < .1, **p < .05, ***p < .01. But perhaps criticism of the Court will follow depending on the cases it chooses to decide. Certain cases, such as those involving abortion rights, are likely to be inherently divisive, thereby allowing reporters to seek out critics regardless of the outcome on the Court. Again, the data do not support this alternative explanation: there is no indication that issue controversy shapes press coverage of the Court. Controlling for issue area (Model 3), attention to the case, and presence of conflict in a case history does little to diminish the impact of majority coalition size and diversity. All else equal, the presence of one additional vote in the majority coalition leads to coverage that is roughly 3.3 percent more supportive of the decision.18 If the additional vote also adds to the coalition’s ideological diversity (i.e., the justice who casts it is more conservative/liberal than any other majority justice), news coverage becomes still more favorable. Indeed, strong evidence indicates that press coverage of the Court responds directly to the signals sent by the justices alone. Most important, large majority coalitions effectively frame decisions by garnering supportive media coverage, regardless of the issues on which they rule, the substance of their decisions, and the opinions of other elites. Across all models, the presence of one additional majority vote generates coverage that is between 2.9 percent and 5.3 percent more supportive. To evaluate the significance of this finding, consider that the press portrays otherwise identical decisions that are reached by five- and nine-member majorities in very different terms. The latter decision will receive up to 21 percent more positive coverage. Americans who read about these two decisions will come away with very different portraits of them. Consider again the case of D.C. v. Heller. Of the twenty articles published, approximately one in four portrayed the five–four decision in unfavorable terms. Had the justices achieved unanimity while reaching an identical decision, direct criticism of it might have entirely disappearedPage 91 → from the press. In its place, readers would have likely read editorials praising the ruling. Page 90 → Table 5.3. Dissensus Dynamics: Decision Support in Newspapers, 1981–2008 (1) (2) (3) (4) (5) 0.045** 0.029 0.033* 0.051** 0.053** Majority votes (0.021) (0.019) (0.019) (0.023) (0.021) 0.067 0.108** 0.089** 0.039 0.047 Ideological diversity (0.050) (0.042) (0.040) (0.049) (0.049) Nonjudicial Elites 0.022 в€’0.046 Majority median–president distance — — — (0.025) (0.095) в€’0.099** 0.028 Majority median–House distance — — — (0.049) (0.726) Political Context 0.127 0.378 Court–pres distance — — — (0.130) (0.464)

0.011 — (0.128) −0.013 Divided government — — (0.065) Issue Controversy and Interest Group Participation Court–House distance

—

— —

−0.022 — (0.043) −0.024 — (0.043) 0.010 — (0.017)

BH controversy

—

—

Lower court dissenters

—

—

Amici n

—

—

Majority median

—

—

—

Alters precedent

—

—

—

Declares law unconstitutional

—

—

—

Chief justice authorship

—

—

—

End term

—

—

—

Coverage n

—

—

—

R-squared N

0.270 88

0.243 88

0.195 87

0.060 (0.562) 0.008 (0.080) в€’0.018 (0.046) в€’0.027 (0.047) 0.033 (0.021)

Decision Characteristics в€’0.091*в€’0.098 (0.046) (0.778) в€’0.075 в€’0.066 (0.113) (0.162) в€’0.020 в€’0.037 (0.068) (0.074) 0.029 в€’0.008 (0.072) (0.080) 0.011 в€’0.042 (0.045) (0.062) в€’0.004 в€’0.004 (0.004) (0.005) 0.294 0.348 88 87

Note: DV is decision support. Results are OLS coefficient estimates with robust standard errors in parentheses. Similar results with a tobit model. *p < .1, **p < .05, ***p < .01. It is unclear whether ideologically diverse majority coalitions can similarly shape press coverage. Ideological diversity has uneven effects across the range of models presented in table 5.3. This finding may result, in part, from multicollinearity in the more fully specified models. Majority votes and ideological diversity are closely related (r = 0.605), particularly in the most consensual decisions. This suggests that the diversity cue is most informative to the press in closely divided rulings. To test this effect, I rerun Model 3 separately for five–four decisions and all others, with the expectation that ideological diversity most powerfully affects media coverage when dissenters are numerous. Figure 5.4 presents the estimated coefficients and confidence intervals on ideological diversity in these cases, providing support for the diversity hypothesis in the case of five–four rulings. The effect of ideological diversity is significant solely in closely divided decisions, meaning that newspapers frame five–four rulings in more favorable terms only when their majorities consist of ideologically broad coalitions of justices. The coverage of Kelo can be used to provide an example of the magnitude of effects. As we have seen, the media presented the ruling in a negative light, warning about the expansion of government power and the dangers it posed to ordinary homeowners. Many of these arguments were drawn directly from the written opinions of the dissenting justices. As a result, Kelo received the most unfavorable newspaper reports of any case in the sample (decision support = 0.098). The press greeted the ruling, reached by five of the Court’s most liberal jurists,

with tremendous skepticism. What does dissensus dynamics tell us about how the voting outcome in Kelo contributed to its coverage? I use coefficient estimates from Model 3 to explore this question. In figure 5.5, I first graph the number of favorable and unfavorable stories that were published on the decision. Then, using the coefficient estimates, I predict how this breakdown would have changed in four instances: when one additional justice joins the majority, when all four dissenters join the majority, when the next most conservative justice replaces a member of the majority, and when the most conservative justice replaces a member of the majority. All of these shifts have modest yet statistically significant effects on Kelo’s coverage. All told, changes in the voting outcome alone shift coverage of the maligned decision in a noticeably more positive direction. As the press shapes its coverage of high-profile Supreme Court rulings, these results demonstrate the attention paid to the justices themselves. Most Page 93 →important, jurists can strongly signal the presence of consensus by forming large majority coalitions, which, in turn, positively affect the coverage rulings receive. When dissenters become more numerous, however, press attention becomes more critical. But the media pays attention to another important signal in closely divided decisions: the makeup of the Court’s voting coalitions. An ideologically broad spectrum of justices can limit the most aggressive press censure that often accompanies five–four rulings. When ideological division accompanies dissent, however, the Court loses its ability effectively to frame its decisions. Page 92 → Fig. 5.4. Effect of Ideological Diversity on Decision Support, by Majority Coalition Size. (Note: Shows the estimated coefficients and 95 percent confidence intervals of the effect of ideological diversity on decision support from Model 3, run separately for five–four rulings [n = 30] and all others [n = 59].) Fig. 5.5. Actual Coverage of Kelo Figure 5.6. Scenarios of Predicted Coverage Based on Who Joins Majority (Note: Charts show favorable articles [decision support ≥ .5] in light and unfavorable articles in dark. Predicted coverage based on coefficient estimates from Model 3.) Robustness of Results To explore further the robustness of these conclusions, I use a second measure of favorable coverage—the extent to which the press gives voice to the Page 94 →opinion of the Court in its coverage (majority reasoning). The measure is strong on reliability, arising out of an automated coding procedure using the program Wcopyfind to account for overlap between Court opinions and news coverage (Bloomfield 2012). The software is designed to find matching text across documents. It enables me to quantify the extent to which newspapers draw on language from a majority opinion to shape their coverage of a case. Simply put, the majority reasoning measure captures the percentage of language from the controlling opinion that appears in news coverage of a case.19 Approximately 5 percent of the writing from a majority opinion appears in the news for the average decision, with a standard deviation of 3.5 percent.20 As expected, broad substantive consistencies underlie the portrait of Supreme Court media coverage using the decision support and majority reasoning measures. To test this, I use the model specified in table 5.3, Model 5, with majority reasoning as the dependent variable. For either operationalization of the DV, there is no evidence that elite disagreement, political context, or controversial issues generate unfavorable press coverage of Supreme Court decisions. But we again find strong evidence that the absolute size of the majority coalition determines the amount of deferential coverage given to any decision (p < .01). Conversely, the effect of ideological diversity is significant at the .05 level and in the direction opposite to expectation, a fact that calls into question whether the ideological makeup of justices in the majority has any effect on the content of news coverage. It provides, at a minimum, no evidence that leading national newspapers are more likely to quote from controlling opinions when they are the by-product of an ideologically disparate coalition of justices. Still, the decision support and majority reasoning indicators—two distinct ways to measure the tenor of news reports surrounding high-profile Supreme Court decisions—suggest that the size of the majority coalition has the most consistent effect on coverage. In cases where dissent is limited, the media offer more support for a decision

and more readily quote from the Court’s written opinion. Favorable coverage and majority opinion citations are also similarly unresponsive to a number of other factors such as issue controversy and elite disagreement. But evidence across all models is weak that the ideological makeup of the Court’s voting blocs has an influence on media coverage. And one important distinction exists between the shape of coverage depending on whether the decision support or majority reasoning indicator is used. The chief justice’s authorship of an opinion makes the press more likely to quote from that opinion though no more likely to project support for it. These results paint a portrait of judicial journalism Page 95 →that navigates a constraint-filled environment with a reliance on the voting signals sent by the justices.

Discussion While national newspapers cover extensively some of the most important Supreme Court rulings, the nature of their coverage varies dramatically from case to case. Rather than defer to the judgments of the Court majority, reporters shape their coverage in response to the size and alignment of the majority and dissenting coalitions. Newspaper coverage of the Court has characteristics similar to that in other areas of American politics, particularly in its reliance on elite sources (Bennett 1990; Entman 2004). But in legal controversies, the media defer to the justices themselves at the expense of all other elites. Given the Court’s popularity, it is unlikely that political actors have an incentive to criticize decisions reached by a large majority. And given the constraints faced by reporters, they, too, rely on the signals sent by the justices. Most directly, the size of the majority voting coalition influences strongly whether a decision meets with favorable coverage. As more justices join the Court majority, the arguments against its decision become both less numerous and less powerful, leading to more sympathetic coverage. In closely divided cases, conversely, media outlets more willingly highlight criticisms of a decision. How do reporters translate voting signals into a diverse array of coverage? Evidence from the previous chapter suggests that judicial dissent has both a direct effect on coverage by supplying criticism of a ruling and an indirect effect by broadening the scope of conflict, thereby allowing outside sources to frame a ruling in negative terms. Coverage of D.C. v. Heller shows these mechanisms in action. Initial reports on the decision include prominent discussions of dissent, allowing journalists effectively to navigate the constraints associated with work on the Supreme Court beat.21 By exploring internal disagreements on the Court, reporters illuminate the meaning and scope of legal disputes in a manner that showcases conflict to keep readers interested. Reporters may also use a frame of ideological conflict when discussing rulings over which the justices disagree. But dissent on the Court has an indirect effect as well—broadening the range of sources that reporters to consult to help shape their coverage. In its coverage of Heller, the Washington Post published a story on “new skirmishes” between advocacy groups that the ruling would engender. The report solicits reactions from presidential candidates John McCain and Barack Obama, Page 96 →members of Congress, mayors, historians, law professors, the National Rifle Association, and Democratic and Republican campaign strategists. Many of these sources raised considerations that the Supreme Court had not addressed. Pollsters debated which party would benefit more from the ruling. Mayor Richard M. Daley of Chicago asked, “Does this lead to everyone having a gun in our society?”22 In broadening the sources it consulted, the Post provided a portrait of Heller that considers its negative consequences more fully than did the justices themselves. This chapter provides the first comprehensive test of dissensus dynamics across a range of decisions. Though there are reasons to expect that leading print journalists may eschew formulaic coverage given their considerable knowledge about the law, evidence suggests that the size of the majority coalition has the strongest influence on positive coverage afforded a decision. Though experienced journalists disagree about whether they rely on a template to cover the Court (Davis 1994; Greenhouse 1996), evidence from this chapter shows that they cannot help but respond to judicial voting signals. Does press attentiveness to voting coalitions hold across other forms of media? In the next chapter, I provide a test of the theory on cable news outlets.

Page 97 →

Chapter 6 Dissent, Debate, and Depictions of Decisions on Cable News On the morning of June 28, 2012, Americans who watched cable news saw a stunning, complex, and ultimately erroneous account of the Supreme Court’s Affordable Care Act (ACA) ruling. Minutes after the Court released the decision, a CNN reporter noted, “It appears as if the Supreme Court justices have struck down the individual mandate, the centerpiece of the health care legislation.” The network quickly posted onscreen two banner headlines: “Supreme Ct. Kills Individual Mandate” and “Individual Mandate Struck Down.” Shortly thereafter, the station switched gears, with its anchor cautioning that the Court’s action “would” be dramatic but more information was to follow. On Fox News, the story was similar. A news anchor reported that the mandate was unconstitutional, with a banner headline noting the same. Quickly, though, Fox began to backtrack. Another anchor stepped in to suggest that the justices had upheld the Affordable Care Act law as an exercise of the taxing power (though the erroneous headline remained onscreen). CNN then updated its banner headline: “Supreme Court Rules on Obamacare.” The network continued to portray the decision incorrectly but added qualifications to its analysis. After nearly ten minutes of confusion, both CNN and Fox News confirmed that the Court had upheld the ACA’s individual mandate (Goldstein 2012). These developments help to crystallize the unique challenges facing cable news networks in their coverage of the Supreme Court. While networks strive for accuracy, they place a greater premium on instantaneous reporting than do print outlets. They also may focus more on the sensational and dramatic in their coverage (Graber 2001). And the structure of cable news Page 98 →reporting differs in important ways from newspaper coverage. In place of edited copy written by a single journalist, cable news networks feature a variety of anchors, reporters, and guests (with varying degrees of expertise) in their Supreme Court coverage. Because of these characteristics, cable news channels are even more likely to rely on simple cues like judicial voting signals in shaping their reports. But no existing research examines the basis of Supreme Court coverage on cable news channels, even though many Americans learn about judicial politics from television broadcasts. Because of the singular nature of the cable news environment, televised stories may have complex effects on American attitudes toward the law. The rapidly shifting and unclear reports on the ACA ruling, for example, provide a stark contrast to the straightforward information presented to subjects in experimental studies of policy agreement (i.e., Hoekstra 1995; Bartels and Mutz 2009). How might differences in the informational environment affect American responses to high-profile judicial decisions? This chapter examines the dissensus dynamics model in light of cable news reports on the Supreme Court since 1995.

Cable News Coverage of Supreme Court Rulings Compared to newspaper accounts, cable news coverage of the Supreme Court tends to be less technical, less precise, and less structured. But cable news channels more readily emphasize drama in their reports and they more willingly feature variegated perspectives offered by multiple on-camera personalities. Consider how the following cable news discussions of Kelo v. City of New London (2005) employ speculation in place of legal analysis: “How do you justify this kind of opinion? I mean a lot of people, lay people are going to look at this and they’re going to say wait a second, so some mayor wants to build a new waterfront and he can just say all right, you know what, I’m going to take that home and that home and that home and that home, and I’m going to get rid of them.” “This situation, it seems to me, is ripe for abuse by rich developers preying on the poor, the weak and the unpolitically connected. Why shouldn’t people who are powerless fear that their land is

going to be taken?” Page 99 →“I think it’s an example of Chief Justice Rehnquist, who may be leaving the Court shortly, not having had his way on several important issues. And this is an issue he has believed in for a long time, limiting the power of government, limiting eminent domain. But he’s in the minority. He never persuaded his colleagues. And 5–4 is the same as unanimous. It’s the law of the land. So this gives local politicians, local government authorities, more or less carte blanche to condemn property and give it to other private people if they believe it’s in longer term interests of the community.” “The Supreme Court is made up of appointees from both administrations. And in this particular instance, not that it always works this way, it was the liberals that allowed government in, not the conservative members of the court. And it has nothing to do with the Bush administration. He, to date, has not appointed anybody on the Supreme Court. So this is—you could look back to all the presidents who had the opportunity to make appointees and blame each and every one of them and that includes President Clinton, President Reagan and so on.”1 Many of the stories involving Kelo used dialogue between on-air personalities to further illuminate the implications of the case. For example, CNN aired a prerecorded report on the ruling from correspondent Brian Todd before cutting to anchor Wolf Blitzer, who prodded for further clarifications. “How did this little community, Brian, in Connecticut get this big case?” he asked.2 Discussions like these—which employ nontechnical language to provide context—are common in cable news coverage. A number of reports also feature more animated arguments about the legal merits, political dimensions, and human-interest implications of judicial rulings. To be sure, then, there are important stylistic and substantive differences that distinguish print and cable news reporting on the U.S. Supreme Court. And indeed, research suggests that broadcast journalism is more likely to emphasize drama and controversy (Slotnick and Segal 1998), focus on human-interest perspectives (Postman 1985; Graber 2001), emphasize partisan considerations (Bennett and Iyengar 2008; Iyengar and Hahn 2009), and eschew context for episodic information (Iyengar 1987, 1991). But even though cable and newspaper organizations produce content that differs in style, their reporters face a similar set of constraints as they shape coverage of the Supreme Court. One important distinction that sets apart Page 100 →newspapers and cable news channels—the use of unedited interviews with elites, which are broadcast quite frequently on television but not often published in print outlets—does not materialize when the Court is involved. Because they lack direct access to the justices, print and broadcast reporters have the same legal information at their disposal—the written opinions of the Court. And cable news stations have goals similar to those of print media: to emphasize drama and conflict, present accurate yet simple information, and report in a timely fashion (Forgette and Morris 2006; Bennett 2007). Given these priorities, we would expect cable news to closely adhere to the predictions of dissensus dynamics, with judicial voting signals proving instrumental. All else equal, coverage is more likely to be favorable and deferential to the Court in cases when a large majority coalition forms. The ideological makeup of the Court majority provides a similarly consequential signal, which cable news programs may emphasize, given their affinity for partisan conflict (Iyengar and Hahn 2009). Dissensus dynamics can also offer predictions about an important feature of the cable news environment: televised incivility (Forgette and Morris 2006; Brooks and Geer 2007; Mutz 2007; Sobieraj and Berry 2011). Mutz and Reeves (2005) define this phenomenon as political conflict in the form of “particularly tense and heated exchanges” onscreen (3). Consider the following example from the MSNBC program The Situation, as on-air personalities debate the merits of Kelo: Guest 1: I think [the ruling] benefits the liberal wing of the Court, simply because it was done on the

basis of, the stores will generate more money for government to spend than your home will. Guest 2: You can take a partisan angle on it, but there is another side to it, too, which is— (Crosstalk) Host: Well, it’s not partisan. It’s ideological. Guest 2: Well, to say this is a liberal problemВ .В .В . it’s not necessarily a left-right split. On this case, the one, I think, silver lining here is that states can take action to protect homeowners if they want to. And I think states will. (Crosstalk) Host: I think what makes it a liberal issue, in effect—and a lot of liberals I know are appalled by this—but in the case of the Court, is that it’s a victory over the group over the individual. The Court is essentially saying, these people can get bent in favor of the greater good. (Crosstalk) Page 101 →Host: And that’s a scary precedent. (Crosstalk)3 These unstructured discussions, which allow guests regularly to interrupt one another, add a singular dimension to how the media covers judicial politics. They are so qualitatively distinct from the types of reports in print media that they deserve attention. They tend to take shape based on a few simple pieces of information (e.g., ideology of a majority coalition on the Court), which enables guests who lack legal expertise to join the debate. Furthermore, they may have important implications for the Supreme Court: though the institution depends on diffuse support to function effectively, a growing body of research suggests that aggressive rhetoric diminishes citizens’ trust in the political process (Mutz and Reeves 2005; Forgette and Morris 2006; Brooks and Geer 2007). Much as news organizations rely on judicial voting signals to guide the extent of favorable coverage they afford a ruling, they similarly use voting signals to determine when they feature incivility in their reports. The reasons for this are twofold. First, cable news producers and program hosts are more likely to feature conflictual voices on-air when the justices signal internal discord with their votes. So in response to judicial dissensus, producers invite incivility through their selection of guests. At the same time, on-air personalities have more material to animate their arguments when the justices disagree. They may draw directly on the disputes that the majority and dissenting coalitions highlight in their analyses. Or guests may disagree aggressively about the meaning and implications of a decision by making inferences based on the voting behavior of the justices. In the preceding exchange, for example, ideological conflict on the Supreme Court allows guests to tussle over whether the Kelo ruling represents a liberal expansion of government power. Televised Incivility Hypothesis: Large and ideologically diverse majority coalitions diminish televised incivility in cable news coverage. The model of dissensus dynamics suggests that reportorial and editorial motivations interact with context to determine the nature of Court-related cable news coverage. To impart knowledge and provide entertainment, cable news organizations rely on the information that elites convey. With respect to the judicial branch, this information takes on a specific form: the voting outcomes on the bench.

Page 102 →Testing Dissensus Dynamics on Cable News Channels To examine dissensus dynamics in the cable news environment, I utilize a simple random sample of all high-

salience decisions handed down by the Court between 1995 and 2008 (see chapter 5). The sample consists of 48 cases, 11 of which receive no coverage from cable news stations (and are thus excluded from regression analyses). For each case identified, I review CNN, Fox News, and MSNBC coverage of the decision by using Lexis-Nexis keyword searches for the name of the petitioner, the name of the respondent, and a term relevant to the case. I read and coded all stories focusing on the Court’s decision in the relevant case that were broadcast within 30 days of the ruling (359 stories total). I then take the indicators for each story variable and average them across all broadcasts that dealt with a given decision. The resulting score supplies a measure of the overall tenor of the coverage for each variable surrounding a single high-salience Court ruling. The independent and dependent variables, which I review here, are described in more detail in chapter 5 and in appendixes A and B. Dependent Variables Decision Support

The measurement of decision support uses the same procedures and standards described in chapter 5, with each broadcast coded for tone and framing content. Krippendorff’s alpha reliability scores between the author and a blind coder for supportive tone and frame dominance were .74 and .62, respectively. As before, I improve reliability by combing the two author-coded measures into a scale of decision support (Cronbach’s alpha of .87). I also improve the efficiency of estimates by averaging story-level decision support over the entire spectrum of a ruling’s coverage, which allows me to provide a summary measure of content for any given case. The decision support scale maintains convergent, predictive, and discriminate validity in the cable news sample. For more information, see appendix B. As in chapter 5, I present the results of OLS regressions here and confirm their robustness with a tobit model. Page 103 →Televised Incivility To investigate the unique impact of the television medium, stories are scored as to whether they feature televised incivility (Mutz and Reeves 2005). In these instances, two or more figures engage in contentious debates, interacting in a less-than-cordial manner. In appendix A, I provide a descriptive codebook that formalizes the coding standards for this variable based on the definition provided by Mutz and Reeves. Percent agreement between the author and the blind coder is .76 for this measure, with Krippendorff’s alpha of .49. To improve the efficiency of estimates, I again aggregate incivility scores across the entirety of a ruling’s coverage. The televised incivility measure has desirable validity properties as well. In particular, its discriminate validity allows us to effectively distinguish it from the measure of decision support. As we might expect, incivility on cable news programs is associated with decreases in decision support (r = в€’.406), but adding the incivility measure to the decision support scale weakens its reliability. Later in the chapter, I provide more evidence about the relationship between the two distinct constructs of decision support and televised incivility. Independent Variables To account for alternative explanations, I employ the independent variables described in chapter 5, which allow me to test the alternative hypothesis that factors outside judicial voting outcomes influence cable news coverage. In particular, I explore whether the preferences of nonjudicial elites, the political context, the issues before the Court and interest group participation in cases, or the specific characteristics of decisions affect the content of cable news reports.

A Summary Portrait of Supreme Court Coverage on Cable News The content analysis provides novel insight into how cable news programs portray judicial decisions, as seen in the summary statistics in table 6.1 and figure 6.1. First, the analysis suggests that cable news channels pay careful attention to the Supreme Court. The typical high-salience case in the sample receives nearly eight unique segments worth of coverage. To be sure, some of these stories are brief, but many others use extensive legal

analysis to explore Page 104 →a decision’s implications. Eleven of the 48 cases sampled receive no news coverage, while the Court’s affirmative action ruling in Grutter v. Bollinger drew the most notice (63 stories). The Spectrum of Coverage A broad range of Supreme Court rulings meet with deferential press coverage, despite the fact that cable news outlets readily emphasize conflict (Forgette and Morris 2006). For example, coverage of Kimbrough v. U.S. (2007; decision support = 1) portrayed the ruling as one of common sense and judicial modesty. According to one reporter, “As you knowВ .В .В . there is a disparity between the sentences for crack and powder cocaine, and the justices basically said, look, judges can take that disparity into consideration when they are sentencing.”4 The informal language and simple analysis is emblematic of televised Court coverage, but stories on Kimbrough nonetheless gave the viewer the impression that the justices exercised wise judgment. In another case, Burlington Industries v. Ellerth (1998; decision support = 0.83), the Supreme Court ruled that companies can be held liable if they do not exercise reasonable care to prevent sexual harassment in the workplace. Cable news stations broadcast 23 stories on the decision, many of which speculated about its implications for sexual harassment suits against President Bill Clinton. The ruling “drew immediate praise from women’s groups” and would create a “better workplace.” One guest on a news program argued, “One of the advantages of having the Supreme Court make a pronouncement is it’s somewhat uniform throughout the country.”5 The vast majority of reports involving Burlington Industries follow this template: praising the decision and speculating about its implications. At other times, cable channels highlight conflict and raise substantial criticism of rulings, questioning the legal findings and political motivations Page 105 →of the Court majority. In so doing, the programs most commonly rely on information provided by the dissenting justices, the responses of political actors, perspectives from attentive interest groups, and the editorial judgments of on-air personalities. Consider coverage of a ruling concerning the Establishment Clause of the First Amendment. Much as it did in Kelo, the media focused on a compelling argument raised by the dissenters in the case. Using language from a dissenting opinion by Justice Clarence Thomas, CNN reported that the ruling “bristle[d] with hostility” to any form of religion. The network also invited dialogue between proponents and opponents of religion in public life, which had the effect of portraying the Court as party to an ongoing political dispute.6 Table 6.1. Selected Summary Statistics of Cable News Coverage, 1995–2008 Variable Mean SD Min Max Obs Decision support 0.68 0.27 0 1 37 Frame dominance 0.69 0.30 0 1 37 Supportive tone 0.67 0.28 0 1 37 Judicial activism 0.07 Televised incivility 0.14 Stories per case 7.58

0.15 0 0.17 0 12.33 0

0.67 37 0.6 37 63 48

Fig. 6.1. Cable News Coverage of High-Salience Supreme Court Decisions, 1995–2008 Unfavorable depictions of a ruling take hold most often during the course of on-air arguments (televised incivility). Decisions in U.S. v. Playboy (First Amendment law), Atkins v. Virginia (cruel and unusual punishment), and Grutter v. Bollinger (affirmative action) received some of the most extensive uncivil coverage in the sample. On Fox News, discussions of the Playboy case used the type of rhetoric that characterizes televised incivility, with guests denouncing one another in succession: Page 106 →Guest 1: I say shame on the Supreme Court Host: You’re misrepresenting the case. This case was about an attempt to protect constitutionally

protected speech. Guest 1: Alan, that’s your opinion. Host: No, it’s not my opinion. That’s the Court’s opinion. Guest 2: The Supreme Court’s opinion. Host: That’s not my opinion. Guest 1: Well, I’ll tell you, I agree. The Supreme Court 200—100 years ago felt that African Americans weren’t persons. Guest 2: We’re not talking about the Supreme Court 200 years ago. Guest 1: 50 years ago—50 years ago, all of the materialВ .В .В . Host: That’s not the issueВ .В .В . Guest 1:В .В .В . all of the material on these cable sex channels would have been considered obscene and unprotected by the First Amendment. Host: You’re misrepresenting—you’re misrepresenting the case, Bob. Guest 1: How did we get from there to a place where you can exposeВ .В .В . (Crosstalk) Guest 2: You’re not even on the same page. You’re not even in the same century.7 As is common in these types of exchanges, talking heads frequently stop one another midsentence, speak over one another, and direct insults at one another (Mutz and Reeves 2005). Many such debates reflect poorly on the work of the Supreme Court, not only by heightening popular distrust of the political process but more directly by attacking the decision-making capabilities of the justices themselves. At times, however, uncivil exchanges are limited in focus; while on-air personalities argue with one another, they may betray respect for the Court and its rulings. CNN featured similarly vigorous debate over the Playboy case, but while guests expressed different views about censorship, they deferred to the findings of the Court about acceptable legal standards of judgment.8 Partisanship and Ideology on Cable News Channels Important differences in style thus distinguish coverage of the U.S. Supreme Court across print and broadcast media. In particular, the simple yet heated exchanges that appear prominently on CNN and other cable channels have no equivalent in the New York Times or the Washington Post. But what about partisan rhetoric on cable news? Existing evidence suggests that broadcast Page 107 →journalism involving the Court might be particularly likely to explore the political and ideological implications of its rulings (Bennett and Iyengar 2008; Iyengar and Hahn 2009). Fig. 6.2. Effect of Majority Median on Decision Support, by Network. (Note: Graph shows estimated coefficients and 95 percent confidence intervals for the effect of majority median on decision support by network, with positive coefficients indicating increased support for conservative rulings.) But there is no evidence to this effect. Indeed, cable news networks do not cover rulings differently based on their apparent ideological content. In figure 6.2, I plot the estimated effect of majority median (a measure of a ruling’s ideological content; Carrubba et al. 2012) on decision support across the three cable news networks. Though Fox News is widely presumed to be more politically conservative than MSNBC, no discernable

differences exist in how the networks treat liberal versus conservative rulings. These findings suggest not that no important distinctions exist between the cable news networks in their coverage of the Court but rather that the networks are unlikely to afford favorable coverage to rulings simply on the basis of ideological content. As we will see later, the ideological cues that the Court offers have a consistently small influence on broadcast coverage when compared with print stories. With this general picture of cable news’ depiction of judicial decisions in hand, we turn now to the central question of this analysis: To what extent Page 108 →do features of Supreme Court rulings determine the nature of cable news coverage given them?

Results and Analysis Does cable news coverage become more unfavorable when a judicial decision features multiple dissenters and narrow ideological coalitions? There are reasons to expect some similarities between print and cable sources in their coverage. Most important, reporters for both receive identical signals on decision day: the case syllabus, majority, concurring, and dissenting opinions. While journalists may seek comment from other sources, official communication from the justices plays the most important role in informing news coverage because it emanates from precisely the type of legitimate elite voices that news reporters covet (Bennett 1990; Entman 2004). Table 6.2 examines decision support and televised incivility on the cable news networks. I begin with simple models of favorable coverage (operationalized with the decision support indicator), controlling for the opinions of nonjudicial elites, the political context, the issues under consideration, and characteristics of the decision itself (Models 1–4). Given the appetite for controversy on cable news channels, the strongest alternative account to dissensus dynamics suggests that broadcast journalism depicts judicial Page 109 →rulings in a more unfavorable light when they concern contentious issues (irrespective of how the Court rules). There is a minimal amount of evidence for the issue controversy explanation (in Model 3 alone, one indicator of issue controversy—judicial disagreement at the lower court level in a case’s history—significantly decreases decision support). But even then, large majority coalitions continue to increase the supportive coverage granted a decision. Table 6.2. Dissensus Dynamics on Cable News, 1995–2008 Decision Support Televised Incivility (1) (2) (3) (4) (5) (1) (2) (3) (4) (5) Majority 0.10** 0.09* 0.10** 0.13*** 0.11* –0.07** –0.08** –0.08** –0.06** –0.07* votes (0.05) (0.05) (0.04) (0.05) (0.06) (0.03) (0.02) (0.03) (0.03) (0.04) Ideological в€’0.16в€’0.13в€’0.08–0.35** в€’0.220.09 0.12* 0.14* 0.10 0.12 diversity (0.13) (0.12) (0.11) (0.13) (0.14) (0.07) (0.07) (0.07) (0.07) (0.09) Controls forВ .В .В . Nonjudicial x elites Political x context Issue controversy Decision character R-squared 0.17 0.18 N 37 37

x

x

x x

0.24 37

x x

x x

x

0.34 37

0.62 37

x x

0.32 37

0.31 37

Note: Results are OLS coefficient estimates with standard errors in parentheses.

0.16 37

x x

x

0.60 37

0.65 37

*p < .1, **p < .05, ***p < .01. Indeed, across Models 1–4, there is strong evidence that majority coalition size influences coverage content. On average, an additional justice joining the majority coalition generates cable news reports that are approximately 10 percent more supportive of a ruling. Model 5 presents a fully specified account of decision support that controls for all alternative explanations. Once again, it shows evidence that broadcast journalists craft coverage in response to the voting outcome on the Supreme Court. Dissensus dynamics is particularly effective at accounting for coverage of Kimbrough (seven justices in the majority coalition, decision support = 1) and of Lorillard Tobacco Company v. Reilly (five votes in the majority coalition, decision support = 0.75). Despite the fact that the former case dealt with a more newsworthy controversy (which allowed the media to seize on the racial implications of disparities in drug sentencing) it also garnered more supportive coverage than the latter. CNN’s coverage of Lorillard featured objections to it from one of the parties to the case, attorney general Tom Reilly of Massachusetts, who noted, “Every day that goes by that [tobacco companies] are able to target their advertising toward kids like this is a bad day.”9 Because cable channels did not view the Lorillard ruling as particularly noteworthy (they broadcast only three stories on it), they did not give detailed explanations about the Court’s written opinions. But the presence of four dissenters signaled to CNN the need to broaden the range of debate about the ruling to include Reilly (even if the channel did not use the arguments raised by the dissenters). These reports contrast with the deference shown by the network to the seven–two ruling in Kimbrough. Conversely, the cases that receive a minimal amount of coverage on cable news networks present the most problems for the full model (Model 5). Two of the cases with the worst fitted values—Bush v. Vera (1995) and Smiley v. Citibank (2000)—garnered one story apiece. In the case of the former, CNN framed the decision favorably in spite of a five–four vote on the Court. Smiley met with criticism in the single cable news report to cover it despite the Court’s unanimous verdict. The model’s shortcomings with respect to lower-salience cases like Smiley are to be expected given the fact that the entirety of the coverage rests on more limited—and thus more volatile—editorial decisions. Page 110 →Majority coalition size performs similarly well in accounting for the presence of televised incivility (Models 6–10). All else equal, the presence of an additional vote in the majority decreases cable news reliance on on-air confrontations by about 6–8 percentage points. To understand the significance of this effect, consider once again reporting on Kelo v. City of New London. Cable news outlets broadcast 30 stories involving the decision, with one-third of them featuring the type of televised incivility that grabs viewers’ attention. These reports used aggressive language (i.e., “ridiculous”) and featured guests interrupting one another frequently. Had the Court reached a unanimous ruling in Kelo, however, we would expect uncivil coverage to have declined markedly. Instead of broadcasting ten segments with incivility present, cable news channels would have substantially reduced this coverage to three or fewer reports, according to the models I present here. Televised Incivility and Decision Support Statistical evidence demonstrates that large majority coalitions significantly increase the favorable coverage afforded a Supreme Court ruling and significantly decrease the use of intemperate rhetoric on cable news programming. But we still have an incomplete sense of the relationship between these two coverage characteristics. To what extent can we differentiate between incivility and decision support? According to scholars, unfavorable messaging can be differentiated from incivility based on style and tone (Mutz and Reeves 2005; Mutz 2007). “Some comments can, in fact, be quite critical of an opponent, and still not earn a classification as вЂuncivil.’ Incivility requires going an extra step; that is, adding inflammatory comments that add little in the way of substance to the discussion” (Brooks and Geer 2007, 4–5). In cable news coverage of the Supreme Court, uncivil exchanges feature aggressive rhetoric by on-air personalities, while unfavorable coverage directs criticism at judicial rulings. These characteristics are not indistinguishable. For decisions that receive mostly unfavorable coverage (decision support < 0.5), incivility is present in about one in four news reports. This figure drops to one in ten for rulings that generate mostly positive coverage.

One account of incivility and decision support, then, sees intemperate on-air discussions as contributing to unfavorable coverage. After the Court rules, a cable news organization looks to the voting signals sent by the justices to shape coverage. Dissent both directly and indirectly fosters unsympathetic portraits of a ruling. It allows news reporters to emphasize arguments Page 111 →raised by the dissenting justices to the extent that they are compelling. It also causes journalists to seek out other sources to criticize a ruling. At the same time, dissent affects the production of news. Cable channels view closely divided rulings as worthy of debate since the justices themselves disagreed over the case. To emphasize this conflict, networks become more likely to use a broadcast format where anchors and guests can express differences of opinion. Programs like Hardball and Crossfire, in particular, might be more likely to cover dissensual rulings since these programs, by design, foster incivility. As on-air debates transpire, critics of the Court are more likely to emerge. Furthermore, they are empowered to attack rulings using the arguments the dissenting justices have published. This account suggests that televised incivility mediates the relationship between dissent and negative coverage of a ruling. I use Baron and Kenny’s (1986) mediation analysis to explore this possibility. Their test consists of estimating the effect of the independent variable on the mediator (which should be significant), estimating the effect of the IV on the dependent variable (which should be significant), and (3) estimating the effects of the IV and mediator on the DV (the mediator’s effect should be significant and larger). Collinearity is a particular concern given the structure of the test and the variables employed here (MacKinnon, Fairchild, and Fritz 2007; Hayes 2009). For this reason, I focus exclusively on three variables: decision support (the DV), televised incivility (the mediator), and majority votes (the IV). I exclude other independent variables since majority votes has the only significant effect across all coverage models in table 6.2. The results of the test, presented in figure 6.3, suggest a strong relationship between majority votes and decision support that is mediated by the presence of incivility on cable news networks. To understand how uncivil debate mediates the relationship between majority coalition size and press coverage, recall the case of U.S. v. Playboy. In response to the Court’s five–four ruling that struck down regulations on explicit programming as a restriction on speech, the press took notice, broadcasting five reports on the decision within days of its announcement. These reports mixed criticism and implicit support for the majority’s holdings. In news-oriented segments, reporters used a measured tone and interspersed background information, quotations from the majority and dissenters, and sound bites from activists. On CNN, reporter Charles Bierbauer read from Anthony Kennedy’s opinion, quoted Stephen Breyer’s dissent, and included taped reactions from the Family Research Council, National Cable Television Association, and Playboy Enterprises. On Fox News’s Hannity and Colmes, however, contentious debates drove the discussion. The hosts Page 112 →and guests argued over First Amendment principles, the definition of obscenity, and the role of federal government. Some of these arguments echoed those made by the Court majority and dissenters. Other portions of the debate diverged from the justices’ analyses, as when one guest on the program levied a charge of activism: “The Supreme Court does not recognize any authority over itself. The Supreme Court lives in the arrogance of power.”10 Fig. 6.3. Voting Coalitions and the Content of Cable News Coverage The case illustrates the process linking judicial dissensus to cable news coverage. At times, the press draws directly on the justices’ arguments to support or critique a decision. But dissensus also entices the media to feature televised incivility. Networks broadcast three segments on U.S. v. Playboy featuring guests who could reasonably have been expected to vehemently disagree with one another. In the Hannity and Colmes segment, the guests included the president of Morality in Media and the vice president of Penthouse magazine. The debate between these guests shaped the coverage of the case, calling into question the Court’s ruling, referencing both original arguments and those espoused by the justices, and levying accusations of judicial recklessness. Distinguishing between incivility and negativity in this manner provides one of the first pieces of evidence that helps to disentangle producers’ and reporters’ responses to judicial decisions. It suggests why journalists, editors (for print publications), and producers (for cable news networks) may react similarly to judicial voting signals. Because reporters rely on sources to add texture to their coverage, they can raise criticisms of the Court

most easily when they give voice to dissenting justices. But editors and producers are more likely to prioritize reporting that will attract an audience with drama and conflict. Though they intend whenever possible to highlight schisms on the bench, they can most readily do so in closely divided rulings. Page 113 →The first support for this implication of dissensus dynamics comes from cable news networks, where we can partially disentangle the work of producers (who invite guests on-air to foster intemperate debates) and reporters (who describe a case outcome using sources). Both of these groups respond to dissent on the Supreme Court. Cable News and the Effect of Ideological Diversity in the Court Majority Cable news channels pay almost no attention to the ideological makeup of the Court majority. Contrary to the predictions of dissensus dynamics, no evidence indicates that more diverse majority coalitions attract more favorable coverage for rulings. This can be seen in table 6.2, where the estimated effect of ideological diversity on coverage is significant in only three of the ten models presented and in the wrong direction. In Model 4, for example, broad voting coalitions actually decrease positive coverage of rulings. To consider more fully the impact of ideological diversity, I estimate its effect on coverage for only five–four rulings, where we would expect the press to pay the most careful attention to it. Once again, diverse majorities have no impact on cable news reporting, even in the most closely divided decisions. How can we account for this null effect? One explanation is that cable news networks are likely to feature guests, speaking in real time, who lack expertise about the law.11 It is difficult for on-air personalities to make inferences based on the ideology of the justices if they do not closely follow the Court (Davis 1994). And in fact, the justices often behave in a manner that obscures their ideological beliefs: reaching unanimity, couching their opinions in the language of the law, deferring to precedent, distancing themselves from the party of the president who appointed them, and denying that personal values play a role in their decisions (Davis 1994; Epstein, Segal, and Spaeth 2001; Bailey and Maltzman 2008). For these reasons, guests on cable news networks sometimes struggle to characterize rulings along an ideological spectrum. Consider a debate broadcast on MSNBC about Kelo. One guest depicted the decision as liberal because it would generate money for the government (though the guest does not reference the makeup of the majority coalition on the Court). The news anchor clarified this perspective, arguing the decision was ideological but “not partisan.” Finally, another guest asserted, without offering any justification, that the decision could not be portrayed along an ideological spectrum.12 Given the incomplete depictions of Kelo each figure offers, it is not surprising that these pundits failed to reach agreement about its ideological content.

Page 114 →Discussion Cable news organizations offer a singular portrait of Supreme Court jurisprudence. They confound expectations in a number of ways. Though research shows that cable programs tend to focus on the sensational and dramatic, they offer ample coverage of the law and the Court. In fact, networks broadcast upward of 20 distinct reports on the most newsworthy of cases. Though research demonstrates that cable channels emphasize partisan conflict in their coverage, no clear differences exist in how various networks portray conservative (or liberal) Supreme Court rulings—cable channels pay little attention to the makeup of voting coalitions. And despite the fact that the Court follows procedures to raise respect for and compliance with its decisions, television networks regularly use colloquial language and aggressive rhetoric in their discussions of the law. In fact, the stylistic content of cable news coverage merits more attention than it has received to this point, particularly because many print stories about the Supreme Court contrast markedly with information found on channels such as CNN and Fox News. Scholars know little about whether informal and entertainment-oriented discussions of the Court increase interest in the law or have detrimental effects on public knowledge. And most of the research involving televised incivility focuses on campaign contexts (Mutz and Reeves 2005; Brooks and Geer 2007) despite the fact that the Court depends on political trust to a greater extent than do other actors. Might

intemperate arguments on cable news networks damage the legitimacy of the judicial institution? But despite the singular aspects of the cable news environment, broadcast journalists respond, as expected, to the size of the majority coalition on the Supreme Court when crafting their coverage of rulings. Voting outcomes have a direct effect by supplying journalists with arguments from the justices’ written opinions and an indirect effect by motivating reporters to seek out outside sources for reactions to dissensual rulings. But the effect of judicial votes is further mediated by uncivil debates on cable news, which producers invite more readily in response to divided decisions. In the course of these debates, on-air personalities use belligerent language against one another and regularly denounce the Court’s decision-making capabilities. These findings provide more support for the model of dissensus dynamics, attesting to the important link between judicial votes and media coverage across time and medium. Page 115 →A Review of Dissensus Dynamics Taken altogether, the evidence provided in chapters 3–6 provides a novel account of how coverage takes shape in response to judicial rulings. It demonstrates significant cross-case variation in how the press depicts decisions and records several important findings. Given the constraints journalists face on the Supreme Court beat, they look to judicial voting signals when crafting their coverage of rulings. Reporters portray decisions unfavorably, including criticism and counterframes to the ones offered by the majority, as the dissenting coalition increases in size. This effect occurs similarly across mediums, news outlets, and time. The size of a judicial voting coalition signals to journalists whether to accept the frames offered by the majority or to highlight conflict in coverage. In dissensual decisions, written dissenting opinions receive coverage to the extent that they are newsworthy (the direct effect). Dissenting votes also cause reporters to seek out other sources, including critics who may challenge a decision by using arguments not raised in the written opinions of the justices (the indirect effect). There is little consistent evidence in favor of the ideological diversity hypothesis; ideologically disparate coalitions of justices appear to receive no more favorable coverage for their rulings. Future work on this point would be well served by examining the effects of ideological diversity in five–four decisions alone (see chapter 4). Important distinctions exist between cable news and print media. Because cable news outlets more readily feature those with little legal expertise in their coverage of the Supreme Court, they are poorly equipped to discern the ideological signals sent by the justices. Coverage produced by cable news also differs in style from that of newspapers, employing colloquy and speculation more readily. More important, cable news produces content featuring aggressive rhetoric and debate in response to decisions that divide the justices. These findings fashion a new portrait of the manner in which the media frames high-profile Supreme Court decisions. They help to demonstrate the importance of taking seriously the missing link between the Court and the public, since the media rarely describe the majority’s analysis faithfully Page 116 →and without comment. Rather, the press balances competing objectives—simplicity, accuracy, drama, and conflict—against one another, leading to wide variations in coverage across cases. This coverage often presents a range of perspectives, both supportive and critical of a ruling, even as the institution itself maintains its reservoir of goodwill among the public. The scenario whereby a trusted institution is criticized for upholding unsound policy offers intriguing tensions. The theory of policy agreement suggests that the Court can build popular consensus when it releases high-profile decisions. Does this theory hold up when we add a skeptical press to the equation? When a ruling creates the perception, for example, that it renders the private property of every American vulnerable to government seizure at a moment’s notice? Such questions are not entirely rhetorical, because the reputation of the Court is

powerful indeed. To fully evaluate the press’s importance in shaping attitudes toward the judiciary and its rulings, one must take seriously the range of factors that influence popular attitudes about them. The second part of this book turns its attention to these factors. The following chapters assess how complex media environments, which reference arguments raised by both majority and dissenting justices and by both supporters and critics of the Court, influence public opinion about landmark judicial rulings.

Page 117 →

Part Two Public Responses in Complex Media Environments

Page 118 → Page 119 →

Chapter 7 In the Court of Public Opinion After the drama surrounding the Patient Protection and Affordable Care Act (ACA) played out inside the walls of the Supreme Court building, a more public second act opened. On June 29, 2012, front pages across the country trumpeted the Obamacare decision with attention-grabbing headlines. In the Boston Globe: “Upheld.” In the Washington Times: “Roberts Court Stuns Nation.” In the Denver Post: “Justices Rule Mandate Just.” In the Tampa Bay Times: “The Law Stands.” Other papers employed more comprehensive headlines that situated the decision in a broader political context. The Washington Post’s subheading noted the surprising vote of Roberts and the decision to characterize the ACA as a tax. The New York Times was descriptive, stating, “Justices, by 5–4, Uphold Health Care Law; Roberts in Majority; Victory for Obama.” The Wall Street Journal said simply, “Court Backs Obama on Health Law.” These headlines described a clear and unmistakable reality: the U.S. Supreme Court deemed the ACA constitutional. But below the fold, the portrait was muddled. Articles balanced context with criticism of the decision, including the legal objections raised by dissenting justices. Juxtaposition of the Court endorsement with multifaceted coverage may have had any number of effects on the many ordinary Americans who paid careful attention to the case. In the coming weeks, Americans grappled with their attitudes in the wake of a Supreme Court decision that brought only a measure of finality to the health care debate.

Does the Media Matter? Contrary to common assumptions, press coverage of Supreme Court rulings is extensive and multifaceted. However, in spite of the institution’s considerablePage 120 → expertise and popular legitimacy, the media do not evince deference to many of its decisions. Indeed, though some portraits of the institution emphasize myth (Casey 1974), and though they at times describe its rulings in straightforward and deferential terms (Hoesktra 1995; Zink, Spriggs, and Scott 2009), they more commonly introduce criticism of Court decision making. Dissent on the bench subjects the institution to heated rhetoric that belies its image as the country’s “republican schoolmaster.” How do Americans react when the decisions of a revered political institution are criticized? Two possibilities exist, with distinct implications for our understanding of the ability of the judiciary to persuade. On the one hand, the specifics of coverage may matter relatively little. Given the trust that Americans place in the Court, they may defer to its judgments and become more likely to support the policies it upholds regardless of the manner in which the media depicts a decision. This scenario—where the legitimacy of the institution carries singular persuasive weight in shaping opinion—has animated optimism about the Court’s role as a consensus builder. On the other hand, the institution’s power to persuade may represent only a piece of the public response puzzle, and the content of news coverage may have a systematic influence on public opinion. An important part of the currency that is judicial legitimacy is the belief that the institution exercises principled discretion and employs unbiased decision-making procedures (Tyler 1990, 2006; Baird and Gangl 2006). Both of these core tenets may be undermined when critics question the bases of high-profile decisions. At the same time, the specific information highlighted in the media may influence Americans’ evaluation of judicial decisions even if it fails to raise doubts about the decision-making procedures the Court employs. The remaining chapters examine media coverage’s ability to influence public opinion about Supreme Court decisions and the mechanisms through which coverage matters. The chapters focus on making connections between the theories of public response and the reality of how the media portray the institution. Ultimately, the remainder of the volume provides a portrait of an institution retaining a wealth of popular trust that nonetheless loses persuasive currency when criticism of its decisions arises.

Media Effects and Policy Approval: Competing Hypotheses Institutional endorsements—like those offered by the Court when it renders a judgment on public policy—act as cues that have the power to shape public opinion. A cue is defined as a simple piece of signaling information that Page 121 →allows people to make inferences without the need for detailed knowledge (Eagly and Chaiken 1993; Rucker and Petty 2006; Druckman et al. 2010). Its value is heuristic in nature: it enables individuals to minimize cognitive costs in forming and maintaining sensible policy-relevant attitudes (Sniderman, Brody, and Tetlock 1991; Lupia and McCubbins 1998).1 A landmark study by Lupia (1994) demonstrates that uninformed voters can draw on endorsement information when popular initiatives are on the ballot. Doing so allows these individuals to make choices similar to those made by better-informed voters. But not all cues are created equal. Research demonstrates that the credibility of cuing sources plays a central role in the effectiveness of the cue. For highly engaged individuals, the credibility of the source serves as persuasive evidence in favor of (or contrary to) the cue (Mondak 1990). For low-knowledge individuals, the trustworthiness of sources acts as a heuristic to minimize cognitive effort (Lupia 1994). Source credibility encompasses a wide array of characteristics—including honesty, likeability, legitimacy, and trustworthiness—that people may employ to distinguish between cue givers (Brady and Sniderman 1985; Sobel 1985; Mondak 1990; Bartels and Mutz 2009). There are many reasons why Supreme Court rulings can function as effective cues. These rulings provide a clear signal through which Americans can make inferences about the constitutionality, wisdom, and value of public policy.2 Their value as heuristics is heightened because the justices typically resolve complex, multidimensional controversies. Americans may be particularly likely to minimize cognitive costs as they form opinions on these matters. And at the same time, popular trust in the Court is high and is manifest in the belief that its justices exercise principled decision making, which people value in legal disputes (Tyler 1990, 2006; Baird and Gangl 2006; Woodson, Gibson, and Lodge 2011). The Court’s credibility makes it the quintessential example of an institution with the ability to exercise opinion leadership (Mondak 1994). Evidence verifies that Supreme Court endorsements of policies can function as effective cues. Drawing on the elaboration likelihood model of attitude formation (Petty and Cacioppo 1983; Petty, Cacioppo, and Schumann 1983; Fiske and Taylor 1984; Chaiken 1987), Mondak (1990, 1992) demonstrates that a Court decision serves as a heuristic for evaluating policy when message recipients put forth little cognitive effort. The same decision acts as persuasive evidence when effort increases. More recent evidence suggests that systematic and peripheral processing of Court endorsements are not mutually exclusive; rather, institutional persuasion takes place by Page 122 →enhancing the degree and direction of issue-relevant thought (Bartels and Mutz 2009). Given that Court decisions serve as potent cues, it is not unreasonable to think that the media might hold comparatively little sway over public attitudes. After all, Americans typically place less trust in the press and other political actors than they do in the Court, which may make judicial endorsements exceptionally valuable. In fact, some evidence suggests that media coverage has very little impact on policy legitimacy (Mondak 1994). Most other research on the Court and opinion does not take seriously the idea that media effects may circumscribe the institution’s persuasive power; rather, scholars typically explain the failure of the institution to persuade as a function of the low salience of its rulings (e.g., Marshall 1989). Nonetheless, it is misguided to assume that a Supreme Court decision overrides all other considerations; there are a number of reasons to expect media coverage to have an important effect on policy approval, even in the face of a clear Court endorsement. One important factor that explains why media effects matter after the release of a ruling comes from research on information processing and public opinion. Evidence suggests that the effects of cues from even high-credibility sources (like the Court) depend on a variety of factors: the relevance of the issue to the receiver, the receiver’s motivation and sophistication, and the strength of available arguments, among others. Studies show that low-knowledge segments of the population may rely most heavily on cues to form their attitudes (e.g., Lupia 1994). The effect is similar among individuals who lack the desire to engage in effortful processing, which occurs most frequently when issues have little personal relevance (e.g., Fiske and Taylor 1984; Chaiken 1987). Both of these scenarios demonstrate real world limits to the persuasive power of a Court

endorsement, since judicial decisions rarely achieve the prominence of something like a presidential campaign and because many rulings have little personal relevance to Americans outside of the local communities they affect (Hoekstra 2000, 2003). To this end, Mondak (1994) distinguishes between policy legitimacy, which the media has little power to influence, and policy agreement, which is shaped by the content of press coverage. Other relevant work verifies that two-sided information environments alter the shape of public opinion. Much of this evidence comes from the literature on framing, which demonstrates that the presence of multiple perspectives diminishes the power of a frame relative to the uncontested condition (Riker 1995; Druckman 2004; Chong and Druckman 2007a, 2007b, 2012). This may result from the fact that framing struggles stimulate conscious deliberation and effortful processing among some individuals, leadingPage 123 → them to focus on “the substantive merits of a frame in judging its persuasiveness” (Chong and Druckman 2007a, 109). Research on frames and cues in tandem is more limited (Chong and Druckman 2007a) but suggests that frames may have more influence than cues on opinion construction by having initial unconscious effects on the evaluation of an issue (Druckman et al. 2010). Another line of research demonstrates that the specific information to which people are exposed has systematic effects on the considerations they call to mind when expressing their opinions (Zaller 1992; Zaller and Feldman 1995). As nearly all work on policy agreement employs a single-message environment—with the link between ruling cue and public response direct, clear, and uncontested—the findings from research on framing effects and competitive information environments present a challenge for the external validity of existing experimental research (McDermott 2002). Indeed, Mondak (1990) demonstrates that high court rulings are more persuasive when accompanied by powerful supporting information. But supporting information may not always effectively make the case for a Court ruling in the arena of public opinion. The justices may indeed be mindful of the way in which the public receives and interprets their decisions, but this popular appeal is but one goal among many. Critics of the Court, however, may focus exclusively on public persuasion after the announcement of a ruling. A study conducted by Clawson and Waltenburg (2003) shows the influence of frames in the wake of the Court’s decision in Adarand v. Pena. Framing the affirmative action ruling as reinforcing the principle of equal treatment increased support for it among white respondents, while framing it as a dramatic setback for African Americans did not. Another study makes a similar point about public reactions to the 2005 property rights decisions: “The Court was far more effective in Lingle than in Kelo in engaging directly with public unease about the relationship between government and private property” (Baron 2007, 616). Furthermore, because the justices speak once and then go silent, critical perspectives can over time come to dominate popular discussion. Thus, though scholars have paid little attention to how Americans navigate a complex information environment in which they have to make sense of endorsement cues from the Supreme Court alongside various framing perspectives, media effects can rob of the institution of its persuasive clout. There are strong reasons to anticipate that the framing of legal and political issues will have a significant effect on policy approval even after people are made aware of a relevant Court decision. The persuasiveness of cues—like a Court decision to uphold a policy—depends on a variety of factors, and competitive information environments diminish the impact of even strong frames and Page 124 →clear cues. The Supreme Court’s passive media stance further circumscribes its influence—while the institution retains credibility, it pays relatively little attention to framing issues for and engaging in popular dialogue. Media Effects Hypothesis: Distinct issue frames can alternately increase or depress approval for a policy the Court has endorsed, even after people are made aware of the institution’s ruling. Though the media effects hypothesis comports with research on framing, cuing, and attitude formation, it adds a new dimension to the study of the Court and public opinion. Literature in this vein has typically viewed the low salience of judicial rulings as the major obstacle to policy agreement. Recently, however, scholars have begun to recognize that reactions to even high-profile rulings can vary widely (Persily 2008; Egan and Citrin 2009). A systematic portrait of Court influence demands movement away from case-specific explanations to a broader framework through which scholars can assess the relative weight of a judicial endorsement when it is counterbalanced by the types of multifaceted media coverage that Court rulings attract. Moving toward a

perspective that takes media effects seriously—and away from a focus on case salience as the primary determinant of opinion response—has broader implications as well, since scholars have little sense of how cues and frames shape opinion more generally (Chong and Druckman 2007a; Druckman et al. 2010). If media effects circumscribe approval for specific policies endorsed by the Court, do they also diminish loyalty toward the institution? The answer is likely no, given the durable nature of Court support. According to Easton (1965), legitimacy allows individuals to accept outputs that they dislike, and evidence shows that short-term outcomes have very little effect on popular trust in the institution (Gibson, Caldeira, and Baird 1998; Gibson, Caldeira, and Spence 2005). Even the ruling in Bush v. Gore, one of the Court’s most politicized and controversial, failed to damage its reputation (Gibson, Caldeira, and Spence 2003b). (Grosskopf and Mondak [1998], conversely, argue that the institution loses legitimacy when it involves itself in high-profile political controversies.) The balance of evidence provides reason to be skeptical that the media may damage the Court’s standing when they depict rulings in unfavorable terms. The framing of judicial decisions may well alter the bases of opinion about specific policies, but it does little to upend the dimensions through which Americans evaluate the judiciary. Indeed, most people’s orientation toward the Court develops early in life and remains intact, reinforced by Page 125 →the symbolism associated with the institution (Gibson and Caldeira 2009a; Woodson, Gibson, and Lodge 2011). And when the press chronicles attacks on the judiciary, people become uneasy, since they dislike strong criticism directed at the institution (e.g., Caldeira 1987). Institutional Loyalty Hypothesis: Specific criticism of a Court ruling will have no effect on diffuse support for the institution. The institutional loyalty hypothesis suggests not that the standing of the judiciary will remain intact throughout all controversies but rather that simple media effects alone are not enough to diminish loyalty toward the institution. Chapter 9 explores how the justices’ involvement in the most high-profile political disputes—about which Americans hold the strongest opinions—can damage their popular standing. It provides evidence that the Court’s decision in the Affordable Care Act case—much like its involvement in the battle about flag burning and the First Amendment (Grosskopf and Mondak 1998)—led to a short-term evaporation of trust. The Effect of Judicial Legitimacy on Policy Approval in a Complex Media Environment How do diffuse and specific support relate to one another when Americans learn about a judicial decision? If the framing of a ruling can diminish (or enhance) approval for a Court-endorsed policy, it is essential to consider the possible mechanisms that link loyalty and policy approval. The media effects and institutional loyalty hypotheses point to a possible tension between approval for the specific output of the Court, which the media may limit depending on the content of its case-specific coverage, and trust in the institution itself, which remains unscathed by media effects. How might Americans reconcile these contradictory attitudes? What is the effect of legitimacy on policy approval in a competitive media environment? How does this process play out at the individual level? These questions strike at the heart of the policy agreement process. But the relationship between institutional trust and support for Court decisions—the central relationship underlying the theory of policy agreement—remains understudied at the microlevel. Consider an individual who expresses a high degree of loyalty toward the Supreme Court, trusting it to reach decisions that are best for the country as a whole. What happens when this individual reads about a seemingly dangerous five–four ruling like Kelo? How might she react compared with Page 126 →another who views the Court as untrustworthy? On the one hand, institutional legitimacy may function as a type of political capital, causing the high-trust individual to express more support for the decision relative to the low-trust individual in spite of her alarm about it (Mondak 1992). This scenario suggests that the judiciary can draw from its reservoir of support when its decisions come under attack and thereby increase support for those decisions. Unfavorable media coverage would be most damaging to the Court when it reaches individuals who already express a low level of

trust in the institution. On the other hand, the process may be reversed entirely. Trust in the institution may have a comparatively small effect on policy approval in a competitive media environment. Rather, individuals who evince both strong and weak loyalty to the Court should express similar levels of approval for rulings when they come under attack. Figure 7.1 details these competing mechanisms through which media effects may influence policy agreement. Fig. 7.1. Microlevel Relationship between Institutional Trust and Policy Approval Behind this latter process is a relevance mechanism. Evidence indicates that people evaluate the relevance of messages when forming their attitudes (Chong and Druckman 2007a). After the announcement of a decision, the press employs a variety of perspectives that directly address the policy controversy resolved by the Court. Because media effects—and frames in particular—alter the considerations that individuals use to evaluate issues (Druckman 2001; Chong and Druckman 2007a), criticism of Court decisions paves the way for Americans to evaluate the merits of the criticism yet makes the endorsement cue irrelevant. Though the Court will not suffer a reputational cost when its decisions are depicted in unfavorable terms, it may lose the ability to increase policy approval based on its reputation alone. Page 127 →Relevance Hypothesis: Specific criticism of a Court ruling will limit the institution’s persuasive power by divorcing diffuse and specific support from one another. For a further example, consider the case of Kelo more fully. Chapter 4 as well as other research (Nadler, Diamond, and Patton 2008) demonstrate that the ruling was lambasted in the press and disliked by the public. Major media outlets emphasized five important frames that reflected poorly on the decision: (1) it would place hardships on ordinary Americans, (2) it rendered private property vulnerable, (3) it upended the Public Use Clause, (4) it expanded government power, and (5) it privileged the wealthy. Absent from this formulation was a serious call to discredit the Court. The press instead focused on the somewhat comical effort of activists to condemn the home of Justice David Souter as “just deserts” for the decision. It was unsurprising that, in Kelo’s wake, the public felt a newfound aversion to eminent domain. But people had little reason to rethink their long-standing trust in the Court, since most coverage focused on specifics of the property rights controversy. Few if any articles suggested that the justices utilized an unprincipled or biased decision-making approach. Even the alternate reality where the institution’s legitimacy was seriously threatened (by, for example, a congressional push to limit its jurisdiction) might not have led to an evaporation of trust in it, given the fact that any discussion of the Court might make the institution the beneficiary of positivity bias (Gibson, Caldeira, and Spence 2003b).

Upcoming Chapters and Contributions The remainder of the book explores the effects of Supreme Court decisions on public opinion, with attention to the role played by the media. These chapters add a new but essential wrinkle to the study of public reactions, demonstrating why careful attention to the press (which animated part 1 of the book) is necessary to evaluate the nature of public attitudes toward rulings and the Court’s ability to build consensus. The analysis offered here provides some of the first evidence that a salient cue from the judiciary is only as powerful as the context in which the media situates it. The study of media effects and the Court begins with a controlled experimental setting before proceeding to a more complex reality. Chapter 8 describes two experiments that demonstrate the importance of media contextualization. The experiments show that the cue of a Court ruling has only limited power: when it is situated alongside arguments from low-credibility Page 128 →“critics” of a ruling, the justices’ persuasive power diminishes. Furthermore, the framing of policy positions across a variety of issues has an impact on public opinion that is similar in magnitude to that from the cue of a Court endorsement. Depending on the frame, the judiciary’s influence can be amplified or diminished. Chapter 8 also provides the first evidence that institutional loyalty loses its potency to increase policy approval in competitive media environments. The results

suggest the need for careful attention to the framing of Court decisions to trace their impact on public opinion. They also demonstrate how the justices, by employing powerful framing perspectives, may be influential in shaping popular responses to their rulings. Chapter 9 traces reactions to the ACA ruling. The case is a singular one, given the unique judicial majority that coalesced, the complex coverage that discussed constitutional principles alongside political considerations, and the strong preexisting attitudes that had formed, largely along partisan lines, even before the Court ruled. The chapter employs one of the most powerful research designs yet in the study of public opinion and the Court. Panel data help to track the evolution of attitudes from more than 400 respondents between June and July 2012. Related evidence from an embedded experiment shows that the specific considerations to which people were exposed had a clear effect on the opinions about the ACA. In particular, the opinion of Chief Justice John Roberts, which portrayed the law as a type of tax, had differential effects on respondents depending on their levels of income. The case is also notable for its distinctive popular impact. The ruling increased support for the national health care law, but this increase was concentrated among Democrats alone, while the views of Republicans remained almost entirely unchanged. Most surprising, evidence shows that views toward the law exerted a causal influence on support for the Court, suggesting that on rare occasions, attitudes toward a policy may be stronger than attitudes toward the institution. The second half of this book demonstrates the importance of examining elite competition, framing, and other media effects in the study of policy agreement. And it ultimately completes the portrait of the Court-press-public link, showing how the study of media coverage sharply alters our understanding of the judiciary’s power to persuade.

Page 129 →

Chapter 8 Cues v. Frames Elite Competition and Media Effects after the Supreme Court Rules The U.S. Supreme Court is, for the majority of Americans, an institution worthy of considerable support. Large majorities favor protecting the Court’s jurisdiction and its right to interpret the Constitution in a countermajoritarian fashion. Large majorities trust it to render decisions in the best interest of the country as a whole (Gibson, Caldeira, and Spence 2003b). But what happens when this reservoir of goodwill is tested through unflattering portrayals of Court rulings? Consider the question in light of the Kelo decision. In this case, many Americans were forced to weigh their trust in the institution against what the media depicted as an unwise and potentially dangerous interpretation of the Public Use Clause of the Fifth Amendment. Ordinary Americans did not react particularly charitably toward the ruling, a fact that casts doubt on the Court’s ability to create consensus in favor of its preferred policy through its stamp of legitimacy.1 But only a limited amount of evidence demonstrates that the manner in which the media framed the ruling was responsible for popular reactions to it (Nadler, Diamond, and Patton 2008). It is also possible that the ruling left public attitudes unchanged but only raised the salience of an issue about which many Americans had already formed strong opinions. Or perhaps the Court indeed increased support for eminent domain seizures but did so only in a very limited manner, with the vast majority of Americans continuing to oppose them. Did the ways in which the press framed the ruling alter the nature of public opinion about Kelo? Did the media’s decision to highlight the evocativePage 130 → language used by the Court’s dissenters undermine the institution’s power to persuade? This chapter answers these questions more generally by evaluating the relative persuasive weight of a Court endorsement against that of frames employed by the press. It does so with a set of experiments that pose new questions about the complex relationships among the judiciary, media, and public. The first experiment focuses on three important questions: Does a critical portrayal of a Supreme Court ruling diminish its support relative to a positive portrayal (which existing research commonly employs)? Can such frames prove effective when they come from relatively unknown sources and do not reference particularly sophisticated arguments? And finally, how do framing struggles mute the impact of Court endorsements? A second experiment—evaluating the persuasive value of Court endorsements and media frames across four disparate issues—builds on the answers to these questions.

Study 1: Framing Effects and Opinion about a Religious Expression Ruling A total of 165 subjects from Amazon’s Mechanical Turk (MTurk) pool of workers took part in a study about “political issues in the news.” Internet subject pools are playing an increasingly influential role in political science and related disciplines. MTurk in particular has proven a valuable tool for recruiting large, diverse subject pools in a timely and cost-effective manner. MTurk allows access to a pool of “workers” who can take part in any number of tasks available on the website. Many workers participate in tasks of their choosing in return for nominal compensation. MTurk samples commonly meet or exceed the diversity and representativeness found in other Internet samples (Buhrmester, Kwang, and Gosling 2011), and researchers have recently replicated published experiments in political science and psychology using pools of MTurk subjects (Horton, Rand, and Zeckhauser 2011; Berinsky, Huber, and Lenz 2012). Leading journals are increasingly publishing results based on MTurk studies (e.g., Arceneaux 2012; Crawford and Pilanski 2012; Doherty 2013; Dowling and Wichowsky 2013; Huber and Paris 2013). Subjects were advised that the researcher was interested in learning their opinions on a range of policy matters and directed to read a story about a current event from the Washington Post. In actuality, subjects had been randomly

assigned to one of three conditions—to read a positively framed story about a Supreme Court decision ensuring public funding of some types of Page 131 →religious expression, a negatively framed story about this same decision, or an unrelated article. This design allows me to explore the relative power of cues and frames as they relate to the high court. The stories in the treatment conditions had a strong measure of external validity, adapted from actual newspaper coverage of Rosenberger v. Virginia (1995), a case where the Court ruled that the University of Virginia was constitutionally bound by the First Amendment to fund a publication by a Christian student group as it did other secular publications. Treatment articles describe the decision identically across conditions, including the cue of a Court endorsement. They differ only as to whether they frame the decision in positive (“praise”) or negative (“criticism”) terms. In the positive frame condition, subjects read arguments that the Court protected the Constitution and put an end to the double standards in the funding of religious student groups; in the negative frame condition, they read that the Court dishonored the Constitution and created a new double standard (see appendix E). The design of these treatments is intentionally very conservative. Subjects received identical descriptions of the ruling across conditions. The positive and negative frames use the same generic arguments to praise and criticize the Court (e.g., that it protected/dishonored the Constitution). This ensures that differences in the strength of frames do not contribute to the results presented here. The design allows me to explore the most basic critique of the Court necessary to undo policy agreement. By describing a decision concerning religious expression, an issue on which most people hold strong opinions, I further expect that frames will have a smaller effect on mass opinion than they would on a less salient topic (Brickman and Peterson 2006). After reading the article, subjects responded to a survey intended to measure demographic characteristics, political knowledge, trust in the Court, opinions about religion, and attitudes toward the decision. Seventy-five males and ninety females took part in the study. A majority of subjects were white (83 percent) and over the age of forty (55 percent). They were, in general, well educated, knowledgeable about politics, and secular. Ninety-nine subjects (60 percent) had received a bachelor’s degree or higher, approximately half expressed a preference for the Democratic Party, and the median subject self-categorized as “moderate.” Thirty-two percent of subjects expressed no religious preference; of those who did express a preference, a plurality identified themselves as Mainline Protestants. And nearly half of all subjects indicated that they never attend religious services apart from occasional weddings, baptisms, and funerals. Page 132 →Measures Policy agreement is a complex phenomenon that must be measured with care. It may indicate an increase in public support for the issue position that the Court endorses. With respect to the case here, subjects respond about whether to allow public funding for religious publications. Agreement may also result in an increase in support for the constitutional interpretation offered by the Court, which here privileges free speech considerations over the separation of church and state. It may involve support for a ruling of the Court, regardless of the substance of the ruling. I also explore whether subjects differentiate between their “personal opinions” and their views about the Constitution when considering a ruling of the Supreme Court. All told, I utilize a four-item decision approval scale that taps the complex nature of reactions to the Court’s religious expression ruling.2 The six-item institutional legitimacy scale (scored 0–1) is used to measure diffuse support for the Court. It remains a staple of the literature, tapping the willingness to abolish the Court and limit its jurisdiction, trust in Court decision making, beliefs that the Court favors some groups over others and is too politicized, and views about the Court’s role as arbiter of the Constitution (see chapter 2; Gibson, Caldeira, and Spence 2003b). Results We begin by looking at the effects of framing struggles on support for the religious expression decision. As expected, the frames used to describe the decision influence opinion about it. Subjects who read about a positively

framed Court ruling express significantly more support for allowing the government to fund religious publications than those in the control condition (Mdiff = 0.16, p = 0.007, two-tailed). But subjects confronted with a negative decision frame opposite a Court endorsement are not persuaded by the institution (Mdiff = в€’0.05,p = 0.42, twotailed). Figure 8.1 shows the estimated effects of framing on support for the Court’s ruling. The significant effect of a positive frame on decision support is not surprising, as the treatment in this condition adheres closely to those in existing research on legitimation. It is relatively uncontroversial that on some issues, a positively framed judicial endorsement can legitimate policies. But the impact of a specific negative frame has not yet been explored, and under these conditions, the Court’s ability to legitimate policies diminishes. This result provides some robust support for the media effects hypothesis Page 133 →given the conservatism of the treatments applied. Again, all information presented to subjects in the two conditions is equivalent, with the exception of whether it is used to praise or criticize the ruling. It does not appear that it is necessary for negative frames to employ strong arguments to lessen the Court’s persuasive capacity (in spite of its institutional reputation). Fig. 8.1. Effect of Framing Condition on Decision Approval. (Note: Bars are OLS coefficient estimates of the effect of framing condition on decision approval, relative to control group. Lines represent 95 percent confidence intervals. Results robust to the inclusion of other variables.) But what of this reputation? In general, subjects rate the Court somewhat favorably on the institutional legitimacy scale (Mcontrol = 0.562), a result in line with other research on legitimacy (Gibson, Caldeira, and Spence 2003b). But evidence shows that people credit the institution when presented with a positively framed ruling; subjects in this condition view the Court as more legitimate (Mdiff = 0.073, p = 0.018, two-tailed). This result follows other research that suggests a positivity bias—people credit the institution for decisions of which they approve but do not penalize it when they disapprove of a ruling (Gibson and Caldeira 2009a, 2009b; but see Grosskopf and Mondak 1998). And as expected, the Court’s standing does not suffer, relative to the control group, when respondents read about a ruling framed negatively (Mdiff = 0.038, p = 0.253, two-tailed). Page 134 →The fact that the Court’s legitimacy does not suffer when its decision is attacked (institutional loyalty hypothesis) leads to the question of how specific and diffuse attitudes involving the judiciary relate during framing struggles. If we assume that no differences exist in the diffuse and specific support expressed by respondents in the control and negative treatment groups, two competing explanations exist. People may defer to the Court more when its rulings come under criticism (to the extent that they view it favorably), or they may defer to it less. In the former scenario, a strong correlation should exist between diffuse and specific support in the negative framing condition alone: those who view the Court most positively should applaud the ruling, while those who like it least should oppose the ruling when criticized. In the latter scenario, the pattern would change, with no correlation between diffuse and specific support when a negative frame is present. This is indeed the case. Table 8.1 and figure 8.2 show strong support for the relevance hypothesis. Diffuse and specific support are closely related in the positive framing condition but unrelated in the negative framing scenario. This is shown across the three models of decision approval presented in table 8.1. While church attendance has a significant effect on support for the religious expression ruling regardless of condition, attitudes toward the Court are significant Page 135 →only in the positive framing condition. In this case, positive views about the institution increase support for its ruling. But when the ruling is framed in negative terms, perceptions of institutional legitimacy no longer affect support for its output. Table 8.1. Framing Effects, Institutional Legitimacy, and the Basis of Decision Approval Decision Approval Control Positive Frame Negative Frame

в€’0.138 0.714*** (0.210) (0.235) 0.130 0.203** Protestant (0.082) (0.076) 0.419*** 0.282* Frequent church attendance (0.150) (0.154) Institutional legitimacy

onservatism Age Female R-squared N

0.223* 0.530*** (0.132) (0.144) в€’0.151 в€’0.200 (0.111) (0.133) 0.028 0.140* (0.066) (0.079) 0.409 0.451 58 54

в€’0.251 (0.249) 0.052 (0.123) 0.330** (0.156) 0.204 (0.172) 0.028 (0.156) 0.177* (0.093) 0.339 48

Note: Results are standardized regression coefficients. The effect of institutional legitimacy is substantively similar across a variety of specifications. *p < .1, **p < .05, ***p < .01. Fig. 8.2. The Effect of Diffuse Support on Decision Approval, by Framing Condition. (Note: Symbols represent the intragroup mean of decision approval relative to the control group. Among those expressing the most support for the Court, decision approval is significantly higher in the positive frame condition than in the control condition (p = 0.001, two-tailed. No other intragroup relationship is significant.) As figure 8.2 shows, this effect is driven by those who view the Court most favorably: they defer to it much more when its ruling is praised but then quickly eschew deference when a framing struggle takes place. Conversely, little correlation exists between framing condition and decision support for those respondents who rate the Court least favorably. These tests provide some of the first evidence of the importance of framing struggles surrounding high court decisions. They suggest that policy agreement may prove harder for the justices to achieve when the press frames their rulings negatively. The tests also suggest that powerful source cues do little to lessen framing effects. Though the Court’s diffuse support remains intact, its ability to persuade changes markedly. Whereas one might assume that judicial credibility proves most persuasive when the Court is under attack, the evidence here shows the opposite pattern—people essentially ignore their attitudes about the Court when formulating opinions about a negatively framed ruling. Though the theory of policy agreement depends Page 136 →on the link between support for the institution and support for its output, diffuse and specific support have little relationship once a judicial ruling receives criticism. Study 1 raises important questions about the potency of judicial influence. It also offers an explanation for why policy agreement appears prevalent in the laboratory but not in the real world. Nonetheless, questions remain about whether frames can similarly diminish Court influence across a range of legal controversies.

Study 2: Framing Four Legal Controversies Study 1 demonstrates that negative frames can attenuate policy agreement, even when they oppose the potent cue offered by a Supreme Court decision. Study 2 further explores these results. First, it traces relative effects of frames and cues with a design that allows for variation along both dimensions. (In Study 1, variations in frames constituted the treatments, while the cue of a Court endorsement was present across both treatment conditions.) The approach used in Study 2 helps to disentangle the interactive effects of frames and cues when they are consistent (i.e., a positive frame and cue present) and inconsistent (i.e., a positive frame and no cue present).

Second, Study 2 explores whether earlier results are issue-dependent. Was the Court’s ability to persuade in Study 1 circumscribed by people’s strong preexisting attitudes about religious expression? Did the salience of the issue influence the power of a judicial endorsement (see Hoekstra 1995)? To better answer these questions, Study 2 features four diverse legal controversies. Subjects were recruited, again through MTurk, to participate in a survey about their opinions on a variety of political issues: gun rights, religious expression (the separation of church and state), government takings of private property, and term limits for members of Congress.3 These controversies were selected because they vary across a number of potentially important dimensions. To the extent that salience matters in how people react to judicial decisions (Hoekstra 1995), these issues may prove instructive, varying from relatively high profile (gun rights) to relatively low profile (term limits). The issues also implicate different dimensions of opinion. For example, subjects may be more likely to defer to the Court regarding the constitutionality of term limits than they would about a matter involving church-state separation, about which they may hold strong a priori opinions. Finally, the decisions vary in their ideological implications, which may affect how subjects view them (Egan and Citrin 2009). One ruling is more consistent with politicalPage 137 → conservatism (where the Court struck down gun control legislation), one is more consistent with liberalism (where the Court banned the display of the Ten Commandments in public places), and two have ambiguous ideological characteristics. Subjects were randomly assigned to either a control group, where they answered a variety of questions about these and other issues, or a treatment group, where they were first tasked with reading a newspaper article about one of the rulings. These articles varied as to whether they included the cue of a Supreme Court endorsement and whether they used a positive or negative issue frame.4 The four distinct articles per issue included: positive frame plus endorsement cue, positive frame only (no cue), negative frame plus endorsement cue, and negative frame only (no cue). All told, the experiment was a four-issue (gun rights, religious expression, takings, term limits) by two-frame (positive, negative) by two-cue (present, absent) design, with a control group. The issue of gun rights provides a fuller illustration of the experimental design. The positive frame was defined as a position consistent with a Supreme Court ruling on the topic. The positive frame article highlighted arguments in favor of gun rights: federal firearms regulations infringed on local authority and Second Amendment protections. The negative frame article took an antigun position, emphasizing the safety of schoolchildren and Congress’s authority to regulate guns under the Commerce Clause. Frames were drawn from actual media coverage of a relevant Court ruling. Articles that included a cue mentioned that the Supreme Court had recently ruled in favor of gun rights. So subjects in the treatment groups were randomly assigned to read one of four articles: Positive frame plus cue: arguments in favor of gun rights with information about a pro-gun-rights Supreme Court decision Positive frame: arguments in favor of gun rights Negative frame plus cue: arguments against gun rights with information about a pro-gun-rights Supreme Court decision Negative frame: arguments against gun rights This design allows for a more careful examination of the relative effects of cues and frames, which often occur in concert with one another in coverage of the Supreme Court. Full versions of stimuli are available in appendix F. A total of 595 subjects took part in the study, 346 males and 249 females. About three-quarters of the subjects were white, with Asians (9 percent) and blacks (7 percent) as the next-largest groups. Subjects were also Page 138 →fairly well informed about politics. About 58 percent of subjects expressed a preference for the Democratic Party and characterized themselves as liberal. The average subject spent about seven minutes participating in the study and was compensated fifty cents. Measures

After reading these articles, subjects responded to a series of questions regarding their attitudes on gun rights, religious expression, eminent domain, term limits, and the Supreme Court. For each issue, a four-item decision approval scale (0–1) taps various dimensions of approval. These items mimic the design of questions used to construct the decision approval scale in Study 1. They tap into various attitudes surrounding the issue in question.5 Scales used to create the key dependent variables had sound properties. Reliability coefficients for the decision approval scales were 0.81 for gun rights, 0.71 for religious expression, 0.67 for takings law, and 0.69 for term limits. For institutional legitimacy, reliability was 0.78. All scales generated a single significant underlying factor.6 Results How might Americans respond to a complex informational environment as they learn about legal controversies? This study focuses on two features of the environment that are most relevant to the study of policy agreement: framing context and cuing information. To test expectations, we begin with a look at approval across experimental conditions for policies involving gun rights, religious expression, eminent domain, and congressional term limits. In each case, we can compare the opinions of subjects in a treatment group with those in the control group. Figure 8.3 presents mean differences in decision approval across these groups. Unsurprisingly, frames and cues have the most powerful effects on opinion when they align with one another in a positive direction. For example, subjects who read a story that frames limits on religious expression in positive terms (i.e., “The government must protect the rights of all Americans, no matter their religion”) and mentions a Supreme Court ruling on the matter express significantly more support for church-state separation than those in the control group (Mdiff = 0.139, p = 0.062, two-tailed). Similarly, a positive frame coupled with a Court endorsement of gun rights leads to a Page 139 →significant increase in approval (Mdiff = 0.185, p = 0.030, twotailed). These effects are as expected because, in both scenarios, the considerations that subjects bring to mind when expressing their attitudes have a consistent effect on their opinions (Zaller 1992; Druckman et al. 2010). Fig. 8.3. Decision Approval Relative to Control Group. (*p < .1, **p < .05, ***p < .01, two-tailed [versus control group].) More interesting, however, is what transpires when frames and cues are at odds with one another. In particular, coverage of the U.S. Supreme Court regularly approximates that presented to subjects in the negative frame plus SC cue condition, especially when dissenters are numerous. In these instances, as we have seen in the case of Kelo, the media regularly portray a ruling (the cue) in unfavorable terms (the frame). According to existing research, the Court’s credibility may allow it to persuade despite the negative framing context. Indeed, “the Court is more influential than Congress in using its institutional credibility to move opinion, and it can do so fairly unconditionally, regardless of people’s sophistication levels, levels of issuerelevant thinking, or the presence of issue-relevant arguments” (Bartels and Mutz 2009, 259). But accounts such as these seem to offer an inflated assessment of the Supreme Court’s ability to persuade. Study 2 provides additional evidence that opinion change is minimal when frames and cues are at odds. Across the eight contrast conditions (cue plus negative frame, no cue plus positive frame), opinion changes in a significantPage 140 → positive direction only in the case of takings law. In all other conditions, the differences between policy support in the control and treatment groups are statistically indistinguishable. In fact, the pattern of responses remains relatively issue-independent. There is no evidence that positive frames absent a Court endorsement are more powerful for some controversies than others. There is no evidence that the Court is more influential on low-salience issues such as term limits for congressmen. These results, particularly in the negative frame plus SC cue conditions, demonstrate limits on the Court’s ability to persuade. They echo the findings from Study 1. According to the relevance hypothesis, negative frames divorce diffuse support for the institution from specific

support for its output. Trust in the Court remains stable after criticism of a ruling, but cues lose their potency. This limiting effect may result in part from relevance, since issue-specific frames provide information that directly relates to a policy issue, whereas cues speak only indirectly to policy considerations. When an endorsement cue is the most relevant information present, it has the most powerful effect on opinion (policy agreement). But the cue loses influence when a frame raises more pertinent considerations. Diffuse support for the institution remains stable, but the Court loses its ability to persuade. Figure 8.4 provides evidence that diffuse support has little effect on policy agreement when framing struggles take place. For all four issues, subjects express significantly more support for a policy when presented with a positive frame plus cue compared with a negative frame plus cue. But across all four issues, frames have no effect on support for the Court. Consider the matter of gun rights. Subjects who read arguments in favor of gun rights and learned about a Supreme Court ruling with the same substantive effect (the positive frame plus cue group) are more supportive of these rights than those who read about the same ruling coupled with counterarguments (the negative frame plus cue group). But subjects in both conditions express similar levels of trust in the Court. In fact, both groups of subjects respond similarly even when asked to assess whether the justices “can usually be trusted to make decisions that are right for the country as a whole” (Mdiff = .018, p = .784, two-tailed). Nonetheless, this trust does not lead to policy agreement.7 Taken together, the results of Studies 1 and 2 suggest the important role the frames play in shaping public responses to judicial rulings.

Discussion Does the content of news coverage influence popular reactions to Supreme Court decisions? Though numerous studies demonstrate how a Court endorsementPage 141 → legitimates policies in even controversial cases, the actual persuasive power of the Court seems more circumscribed. This chapter demonstrates that media effects have similar importance to a Court endorsement cue, with support for a ruling dependent on the presence of positive frames. Critics need not introduce any novel considerations or reference sophisticated arguments to effectively neutralize the Court. Reports that simply “criticize” a decision using the same generic arguments that can be employed to praise it can limit Court influence. Fig. 8.4. Relative Approval by Framing Condition, with Court Cue Present. (Note: Positive values indicate greater approval in the positive framing condition. *p < .1, **p < .05, ***p < .01, one-tailed.) As the press frames its policy output in critical terms, the Court loses the ability to convince. The institution’s diminished persuasion powers stem not from a loss of legitimacy—as subjects are reticent about formal censures of the Court—but rather from an unwillingness to defer to it. As the framing environment becomes more unfavorable toward the judiciary, it can face the prospect of decision-specific disapproval, if not a dramatic erosion of support. These findings hold across a range of controversies with which the Supreme Court has involved itself in recent decades. Still, this account does little to suggest what drives backlash against a ruling (where people express significantly less support for a policy after the Court endorses it), which future research would do well to explore. Backlash may arise only when strong negative frames dominate the media environment (Chong and Druckman 2007b). A single news article may offer only Page 142 →a limited impetus for backlash; for many other issues, a sustained media campaign may be necessary for people to reject strongly a Supreme Court decision. For example, the public’s overwhelmingly negative reaction to Kelo v. City of New London occurred while the press lavished attention (much of it unfavorable) on the ruling. This chapter also suggests that more work is needed to explore the relative effects of frames and cues. It demonstrates that both matter to an extent: particularly when frames and cues align with one another, they powerfully affect opinion. But the cue examined here—a ruling by the U.S. Supreme Court—is relatively powerful, given the credibility of the institution. Do cues remain as potent with a relatively anonymous source? When individuals believe the source to lack credibility? One might imagine a more comprehensive study of how endorsements affect political campaigns depending on the identity of the endorser. But the study of frames and

cues must move beyond the campaign context to consider other communication environments. For now, we can say with confidence that the Supreme Court’s ability to persuade is more limited than previously believed. It is circumscribed by the types of framing struggles that regularly characterize coverage of the judiciary, which cause Americans to eschew the endorsements the Supreme Court offers. Chapter 9 offers the most detailed look yet at the Court-press-public dynamic in the aftermath of the landmark Obamacare ruling. The findings demonstrate the continued influence of the press, even for an issue about which many Americans have already formed relatively stable opinions. More important, new evidence demonstrates the limits of the power of legitimacy when the Court involves itself in political controversies about which Americans hold strong attitudes.

Page 143 →

Chapter 9 The Supreme Court on Trial Public Opinion and the Obamacare Ruling The Obamacare decision was a landmark case in the study of the institution and public opinion, unprecedented in stature and complexity. It ranks among the most high-profile, widely anticipated, politically consequential judicial decisions ever released. But while Americans paid careful attention to the Court’s pronouncement, they were not ripe for persuasion—far from it. In fact, many people had long before formed strong and durable positions on the Affordable Care Act (if not on each of its individual components), and partisan and ideological differences over the law were extraordinarily pronounced. The justices themselves added intrigue to the mix, with a verdict in multiple parts, each reached by unique coalitions of justices. In this chapter, we turn our attention for one final time to this decision—to its media coverage and the popular reactions to it. Making meaningful inferences about the decision’s complex effects requires a powerful research design. A panel study of popular opinion pre- and postruling serves as such a design. The data help to address three of the most vexing shortcomings in existing research on the Court and public opinion: an absence of comparable preand postdecision measures of opinion (see Marshall 1989), a lack of data on individual-level opinion change (e.g., Franklin and Kosaki 1989), and the failure to use indicators of diffuse support when assessing loyalty to the judiciary (see Gibson, Caldeira, and Spence 2003b). As one might imagine, the press plays a central role in the Obamacare controversy. Most major media outlets eschewed coverage deferential to the Court for a more complex portrait of constitutional disputes, vote swings, and political implications. Did such coverage affect public opinion about Page 144 →the decision? Evidence suggests the answer to this question is yes. But other patterns emerge as well: characterizations of the ACA as a tax had a systematic influence on popular views about it and the ruling led to a small but very real increase in approval for the health care law among Democrats alone. The panel study also facilitates new conclusions about the relationship between institutional legitimacy and policy approval. Scholars debate whether the Court damages (Mondak 1992; Mondak and Smithey 1997; Grosskopf and Mondak 1998; Stoutenborough and Haider-Markel 2008) or enhances (Gibson, Caldeira, and Spence 2003b; Gibson and Caldeira 2009a, 2009b) its standing when it becomes enmeshed in political controversies. A related question—regarding the microlevel relationship between diffuse support for the Court and approval of its decisions—has received little attention. This is problematic for the theory of policy agreement, which suggests that institutional trust influences decision support at the individual level, an account that has rarely been interrogated with any microdata. Surprisingly, evidence suggests that the Obamacare ruling caused Americans to alter their views of the Court to accord with their views of the health care law—not, as one might expect, vice versa. This pattern runs contrary to existing theories of Court influence (e.g., Gibson, Caldeira, and Spence 2003b), suggesting that in select controversies, policy preferences can be quite powerful in shaping specific attitudes toward the Court.

Attitude Change in the Wake of the Health Care Case The Court’s influence on policy approval comes about through both symbolic and persuasive means (Mondak 1994; Bartels and Mutz 2009). Following important decisions, Americans learn about an association between a policy outcome and an institution in which they place trust. This simple endorsement—where the institution confers its symbolic stamp of legitimacy—can have a positive influence on views of the policy (Gibson, Caldeira, and Spence 2003b). But the extent of attitude change also depends on context. Rulings have greater impact when they are high salience (Murphy and Tanenhaus 1968) and have the appearance of procedural fairness (Baird and Gangl 2006; Zink, Sala, and Spriggs 2009). Based on these criteria, the Obamacare decision appeared poised to increase support for the health care law. The case’s high salience made clear the link between credible institution and policy outcome. The opinion, penned by Republican appointee John Roberts, had an aura

of procedural fairness. And there is another reason to expect increase in support for the health care law in Page 145 →the wake of the decision: Supreme Court rulings are more influential when they introduce new information into public discourse (Johnson and Martin 1998; Brickman and Peterson 2006). Information introduced by the Court (and in the media) becomes accessible as considerations for Americans expressing their policy attitudes (Zaller 1992; Zaller and Feldman 1992; Brickman and Peterson 2006). Given the case’s high profile, the public was quite familiar with Obamacare, with the exception of a novel consideration: the law represented a type of tax. Before the ruling, President Obama had explicitly disavowed the tax characterization, while opponents of the law had eschewed this argument in favor of more aggressive claims about its purportedly unconstitutional sweep. Only after the Court upheld the law as an exercise of Congress’s taxing power did the tax argument come to animate public discourse (National Federation of Independent Business v. Sebelius [2012], 43–44).1 But while the characteristics of the Obamacare decision—its high salience, appearance of procedural fairness, and redefinition of the law—placed dynamic pressure on public opinion, dramatic changes in opinions about the ACA were unlikely to follow. Indeed, public opinion about the ACA overall was well defined prior to the Court’s ruling. Approval and disapproval of the law remained relatively static between March 2010 and June 2012 (Campbell and Persily 2012). Such attitude crystallization circumscribes the Court’s ability to persuade (Johnson and Martin 1998; Brickman and Peterson 2006). However, the decision may have slightly altered views about the law across different subgroups of the population (see Franklin and Kosaki 1989). Because Democrats and Republicans were sharply divided over the health care law (Campbell and Persily 2012), partisan divergence in opinion response became more likely. By raising the salience of the health care law anew, the decision may have reignited critical discussions of it among Republicans and positive discussions of it among Democrats, leading to a more sharp crystallization of opinion differences. Indeed, even in the most controversial of cases, the “informational and evaluative biases of different social groups lead to different patterns of response to public decisions” (Franklin and Kosaki 1989, 763). All told, we can expect the controversy over Obamacare to engender a complex public response that provides insight about the various theories of Court influence that scholars have proposed in recent decades. The ruling placed countervailing pressures on public opinion. On the one hand, its high salience and nonpartisan appearance opened the door for an increase in approval for the law. The Court, with the help of the national news media, introduced a relatively novel consideration into what had been a well-worn Page 146 →public debate over health care policy. On the other hand, a fractured and ossified public greeted the decision. Most Americans had formed relatively strong and stable opinions about Obamacare long before June 2012. And any attitude changes that occurred after the ruling might have further increased partisan polarization. Relating Policy Approval and Diffuse Support The strength of health-care-related attitudes and the high profile of the health care case lead to another question: whether the decision might have concurrently reshaped opinions about the Court and the law. This is a more theoretically fraught puzzle that can be addressed in two ways. First, do rulings such as the Obamacare one damage or enhance the Court’s standing? Debate on this question concerns both mechanisms and measurement. Mondak (1992), Grosskopf and Mondak (1998), and Stoutenborough and Haider-Markel (2008) find that the institution expends political capital when it becomes involved in controversies, diminishing popular trust in it. But Gibson and Caldeira argue that when the Court is in the public eye, exposure to symbols of judicial legitimacy actually increases support for the institution (Gibson, Caldeira, and Spence 2003b; Gibson and Caldeira 2009a, 2009b). Furthermore, they argue that diffuse support for the Court is a more appropriate indicator of its standing (Gibson, Caldeira, and Spence 2003b, 537). Chapter 8 of this volume provides no evidence that the media may depress the institution’s legitimacy based simply on the manner in which it frames rulings. Second, do Americans shape their views on health care policy based on the level of esteem they assign to the Court? This is a crucial question for the theory of policy agreement, which posits a microlevel mechanism: individuals with high levels of trust in the Supreme Court should express more support for the policies it endorses. But only one study offers evidence of the causal process at the individual level (Gibson, Caldeira, and Spence 2003b). It is far more common for scholars indirectly to link diffuse support and policy-specific support by

assuming the link between legitimacy and policy approval has a microlevel underpinning (e.g., Hoekstra 1995; Bartels and Mutz 2009). But chapter 8 casts doubt on this mechanism by showing that diffuse and specific support have little relation to one another when the Court comes under attack in the media. The Obamacare decision thus offers a further opportunity to examine how diffuse support for the Court influences an individual’s views about Page 147 →policy. It also raises questions about whether existing theories of legitimation lack a crucial variable—what we might call attitude strength. There is a vast literature in psychology on attitude strength. This is the extent to which attitudes exhibit durability, impactfulness, or both (Krosnick et al. 1993; Krosnick and Petty 2014). These features are observable, causal indicators of attitude strength; in other words, evidence of durability and impact is evidence of strong attitudes. There are two manifestations of each feature. Durable attitudes display persistence, remaining unchanged over a long period of time, and an ability to resist attacks. Impactful attitudes have the ability to influence information processing and judgments and to shape behavior (Krosnick and Petty 2014). This framework provides another lens through which we can view the debate regarding institutional support. Those who take the view that judicial esteem is unlikely to decline in the short term are suggesting that it meets the criteria of persistence and resistance. Similarly, the ability of the Court to draw on its loyalty to legitimate controversial policies suggests that institutional support can influence other judgments. These are precisely the characteristics that describe strong attitudes, even though legitimacy scholars do not use the term with regularity. What happens to our understanding of Court influence when we view institutional support not as a singular aspect of public opinion but simply as a relatively strong attitude (which much evidence suggests that it is)? Once again, the literature on attitudinal changes is helpful. According to congruity theory, people resist imbalance in their attitudes when an assertion is made that links a source and a policy (Osgood and Tannenbaum 1955). Because a person will hold attitudes toward both the source and the concept, an assertion linking the two will frequently (though not necessarily) give rise to attitudinal incongruence. In this scenario, it is unlikely for people to maintain sharply divergent attitudes toward the two; rather, people will be inclined to reshape their attitudes toward one or both to avoid attitudinal imbalance (Osgood and Tannenbaum 1955; Osgood 1960; Feather 1964; Tannenbaum, McCauley, and Norris 1966; Tannenbaum 1967). The nature of this change depends on attitude strength, since powerful attitudes are more resistant to transformation. The health care decision fits neatly into this framework, providing Americans with a clear association between source (the Court) and policy position (its decision to uphold the law). Consider the dilemma facing an individual who expresses some support for the Court as an institution but opposes the health care law. To be sure, the Court’s decision to uphold might result in attitudinal contradictions (“I oppose the policy but support the institutionPage 148 → that upheld it”), but it is possible that certain individuals can indeed hold apparently contradictory attitudes by making a sophisticated distinction between processes and outcomes (see Gibson and Caldeira 2011 for a discussion of Americans’ sophisticated views about the Court). But even this possibility becomes more remote as attitudes become stronger. Instead, congruity theory provides evidence that attitudinal changes toward source, policy, or both may follow in many circumstances. More specifically, we would expect institutional support to dominate policy attitudes in cases where people felt strongly about the Court but more ambivalent about the ACA. These individuals would express increased approval of the law after the ruling. But for staunch opponents of the law, we would expect the opposite process at work: they would update their attitudes about the Court to accord with their opinions about the law to the extent that their preferences about the latter were stronger.2 Is institutional support simply a very potent attitude, or is it somehow unique? Once again, this question comes into sharper focus in only the most unusual cases, where views about specific policies may prove stronger (more impactful and persistent) than views about judicial institutions. In these cases, we might observe the influence of policy approval on institutional support. And indeed, the health care case is just such an exceptional case. Attitudes regarding the ACA appeared quite strong at the time of the ruling, as evidenced by their durability to that point and their basis in “high partisan passions” (Persily, Metzger, and Morrison 2013, 2; Campbell and Persily 2012). Given the strength of opinions about the ACA in 2012, the decision represents a best case for

observing such effects. Might the relatively powerful views that many Americans held regarding the ACA have influenced their strong but possibly less potent views about the Court itself? This question ultimately can be adjudicated only by evidence.

Exploring Opinion Change about the ACA This study combines the benefits of both laboratory and real-world studies of Court influence by employing a panel design. Respondents were recruited through MTurk in June 2012. They agreed to take part in a survey about their attitudes on political issues relating to the 2012 elections, answering an initial set of questions as well as a follow-up survey approximately one month later.3 Respondents were required to verify age and location in the United States to participate. To generate high-quality data, I once again restricted the respondent pool to experienced MTurk workers who had earned Page 149 →approval in more than 95 percent of the studies in which they had participated. Respondents were required to demonstrate their attentiveness to the survey as they responded. Each participant received fifty cents for the first part of the study. The MTurk interface was ideal for generating a large subject pool and ensuring timely responses (Berinsky, Huber, and Lenz 2012), two essential components in a study of the short-term effects of the health care ruling. The first part of the panel was administered between June 13 and 27. Respondents answered a variety of demographic questions and items about their views on politics in general and the 2012 election in particular. In early July, I messaged all participants in the first phase of the study and encouraged them to take part in the follow-up, offering each one dollar for taking part in the second survey. After five days, I sent another message to those respondents who had not yet completed the follow-up. All told, 71 percent of the respondents from the first wave participated in the postdecision survey, which was administered between July 3 and July 22. In this second survey, respondents answered questions that enabled me to verify their prior participation. They then answered a variety of questions about the health care ruling, other Supreme Court decisions, and their media usage over the previous month. Because the postruling questionnaire was not identical to the preruling one, it is possible that question-order effects influence responses even on identical items (Zaller and Feldman 1992). Nonetheless, by analyzing identical pre- and postruling items, this study goes a step further than much of the existing research on actual Court rulings, which often relies on comparable yet distinct items administered across different surveys (e.g., Marshall 1989; Gibson, Caldeira, and Spence 2003b) or on postruling items alone (e.g., Grosskopf and Mondak 1998; Nadler, Diamond, and Patton 2008). Dependent Variables Research on policy agreement explores how Supreme Court rulings alter approval for controversial policies. I assess whether respondents changed their overall assessments of the health care law with an approve/disapprove item: Obamacare approval change captures the within respondent change in approval from the pre- to postdecision phases. I also test whether respondents distinguish between their approval of the law and their views about its constitutionality, asking, “Regardless of your personal opinion, do you think that it is constitutional or unconstitutional for the federal government to Page 150 →require all Americans to have health insurance, or do you not know enough to say?” Obamacare constitutional change captures within-subject change on this item. All change variables described here range from в€’1 to 1, with a score of 0 indicating no within-subject change from pre- to postruling, 1 indicating change from strong disapproval to strong approval, and в€’1 indicating a shift in the opposite direction. An ideal measure of institutional support would employ the six-item institutional loyalty scale developed by Gibson, Caldeira, and Spence (2005). As they note, this scale does a much better job of capturing enduring support for the Court than do standard confidence items, since there exists a “strong influence of contemporary events on replies to the confidence question” (363). However, the use of the full six-item scale in what was otherwise presented to respondents as a survey about politics and the 2012 election risked sensitizing them to the ruling.4 Furthermore, use of the full scale would have increased the likelihood that they would guess at the study’s purposes, thus introducing bias into the results.5 Left with this dilemma, I judged it unwise to

administer the entirety of the institutional loyalty scale in the preruling phase. This judgment was informed by these empirical concerns, of course, as well as by practical ones. There now exists a wealth of survey evidence drawing on the institutional loyalty scale (e.g., Gibson and Caldeira 2009a) but no panel studies of opinion change in response to decisions. The risk of sensitizing respondents—which would call into question whether we are learning about the effects of the ruling or the effects of the survey—were simply too great to include the full loyalty scale. Instead, I use perhaps the best single-item indicator of institutional support: a feeling thermometer, which measures “highly general” attitudes toward the Court, including “some elements of loyalty” (Gibson, Caldeira, and Spence 2003a, 363). I also employ a single item from institutional loyalty scale: whether respondents judged the Court to be too mixed up in politics.6 I measure changes in these assessments from pre- to postdecision with the variables SC temperature change and apolitical Court change. Independent Variables In the predecision survey, respondents answered questions about their gender, race, age, family income, party identification, ideology, plans to vote in the 2012 presidential election, and preferred presidential candidate. From these items, I create the variables that are included in the analysis here: age and income (scaled 0–1) and dummies for female, black, Hispanic, and plan Page 151 →to vote. I construct measures of party identification (on a seven-point party identification scale, 0–1, using the variable Democrat) and liberalism (on the same sevenpoint scale). I also construct a 0–1 scale measure of political knowledge from five items: whether respondents correctly identified the size of the congressional majority required to override a presidential veto, the office held by Vice President Joe Biden, the office held by Chief Justice John Roberts, the party holding the most seats in Congress, and the relative conservatism of the two major political parties. In the postdecision survey, I asked respondents a number of other questions about their awareness of the Supreme Court’s health care ruling. I create a measure of news usage by asking respondents about the frequency with which they learned about the news on television, on radio, from the Internet, and in newspapers and magazines (news, scaled 0–1, with 1 indicating a “very frequent” response to every medium). Finally, I asked respondents to estimate the frequency with which they had discussed political affairs with family, friends, or acquaintances in the preceding month (discuss politics, scaled 0–1). Subject Characteristics More than 400 respondents took part in both phases of the study. Of these respondents, 58 percent identified as female, and a plurality were between the ages of 30 and 39. The racial makeup of the population was 81 percent white, 11 percent black, 2 percent Hispanic, and 5 percent other. Workers on MTurk are commonly more liberal than the general population, a pattern that holds here. The median respondent leaned toward the Democratic Party and characterized herself as slightly liberal. The respondents were also highly knowledgeable about and interested in politics: 84 percent of them planned to vote in fall presidential election, and half of them answered correctly all political knowledge items. One particular concern for any panel study is dropout. The dropout rate here is relatively low: about 29 percent. But if respondent characteristics have a systematic effect on dropout, this could bias results. I run a logit regression to predict dropout from a variety of respondent characteristics. The results of this regression are presented in table 9.1. It appears that only one factor has a strong influence on a respondent’s decision to leave the study: the least knowledgeable respondents were less likely to participate in the postdecision survey. All other respondent characteristics have little effect on the decision to drop out. Later, I present evidence to show that for respondentsPage 152 → who participated in the follow-up study, political knowledge did not impact their views of the ruling. Validity of Results Since respondents do not constitute a random sample of the population, we must take care in making generalizations. Nonetheless, there are reasons to expect the MTurk sample to generalize at least as well as would

other nonrandom respondent pools. And there are indications that the results presented here are broadly consistent with the shape of public opinion in more representative surveys of the population. For example, Campbell and Persily (2012) review surveys on the Obamacare ruling from Quinnipiac, Kaiser, Pew, the New York Times/CBS, and the Washington Post/ABC, and their Page 153 →characterizations of public attitudes closely approximate the data I present here. In particular, the predecision phase featured “substantial intergroup differences in support for the legislation,” with the largest difference across partisan groups, and “opposition to the health care law [that] exceeded support” (Campbell and Persily 2012, 10–12). After the ruling, “the Court’s job approval fellВ .В .В . most notably amongВ .В .В . Conservatives, Republicans” (19). The decision “had a small positive effect on ACA support,” but “the greatest change in attitudes occurred among Democrats” (15). Each of these results is borne out in my data. Table 9.1. Panel Dropout Dropout 0.26 Female (0.02) 0.14 Age (.36) в€’0.08 Black (0.41) 0.25 Hispanic (0.54) в€’0.60 Democrat (0.61) в€’0.19 Liberalism (0.63) в€’0.25 Plan to vote (0.41) 0.61 Obamacare approval (preruling) (0.43) 0.05 SC diffuse support (preruling) (0.38) в€’1.39*** Knowledge (0.53) N 537 Log-likelihood в€’273.5 Note: Results are estimated effects from a logit model with standard errors in parentheses. *p < .1, **p < .05, ***p < .01 More specifically, I administered an identical item to that used in New York Times/CBS News polls from June 7 and July 18, 2012. Of the self-identified Democrats expressing an opinion on the health care law in the preruling poll, 72 percent strongly or somewhat approved and 29 percent strongly or somewhat disapproved. The comparable figures from study respondents were 79 percent and 22 percent. Among Republicans, approval of the law broke down as follows: 6 percent strongly or somewhat approved, and 94 percent strongly or somewhat disapproved in both the Times/CBS poll and the respondent population. These results attest to a strong consistency in attitudes between study respondents and more representative samples of the population. These similarities, however, belie the many advantages to using the MTurk population. The data is well suited to address existing controversies in the literature on Supreme Court influence, providing directly comparable

indicators of pre- and postruling opinion that are measured within individuals across time. This allows us to disentangle causality much more fully than the literature has been able to do. At the same time, some indicators—particularly of diffuse support for the Court—are precisely the ones at the center of a dispute in the literature (Grosskopf and Mondak 1998; Gibson, Caldeira, and Spence 2003b; Stoutenborough and Haider-Markel 2008), but they do not appear in standard opinion surveys like the ones reviewed by Campbell and Persily (2012).

Results Before the Supreme Court’s health care ruling, respondents were divided in their views about the ACA. Roughly 45 percent of respondents indicated they disapproved of it (with nearly two-thirds expressing “strong” disapproval), and 55 percent indicated they approved (with only one-third expressing “strong” approval). When asked about their views regarding the Page 154 →law’s constitutionality, respondents were more unsympathetic: about 63 percent judged the law to be unconstitutional. Despite the Democratic bias in the sample, then, opinion of the ACA was not particularly favorable in June 2012, with doubts about its constitutionality and misgivings about the law as a whole. The Effect of the Decision on Opinion In the postdecision phase, respondents evinced clear knowledge about the Supreme Court decision, which is to be expected given its extraordinarily high profile and the political awareness of respondents. Roughly 90 percent indicated that they had heard about the ruling, and nearly the same number correctly identified that the Court upheld the law. What effect did the decision have on attitudes toward Obamacare? Figure 9.1 presents pre- and postdecision approval for the ACA. It demonstrates a small increase in approval for the law, concentrated in the percentage of subjects expressing strong approval for it. But we can explore the effect of the ruling more rigorously by looking at within-respondent attitude change. In figure 9.2, I depict the mean of Obamacare approval change and Obamacare constitutional change (within-respondent) along the results of a paired t-test. These results demonstrate that though the effect of the Supreme Court ruling on attitudes was substantively modest, it was real and statistically significant. The average respondent expressed an increased approval of the law (Mapproval change = 0.03, p = 0.09) and was more likely to rate it as constitutional (Mconstitutional change = 0.11, p = 0.001) after the ruling. These general patterns, however, obscure group-level variation. Evidence indicates that Democrats and liberals responded positively to the ruling, while Republicans and conservatives were not persuaded by it. (In fact, the Court decision proved so unpersuasive that not a single Republican identifier altered his or her assessment of the ACA’s constitutionality.) Figure 9.2 shows a significant increase in approval for the health care law and views about its constitutionality solely among self-identified Democrats. In table 9.2, I explore changes in support for Obamacare from pre- to postruling more fully. In Models 1 and 2, I control for demographic characteristics and stable predispositions that include partisan attitudes and diffuse support for the Supreme Court. These models explain very little about how the ruling altered attitudes toward health care in the United States. In Model 3, I control for predecision opinion about the ACA, and in Model 4 I include interactive effects. Doing so provides a significant increase in fit, Page 157 →making plain some surprising patterns. Both preruling opinion and predispositions shaped reactions to the ruling, but in different directions. The law’s strongest supporters are less persuaded by the ruling, which can be seen in the negative coefficients on Obamacare approval (preruling), while among those respondents who disliked the ACA before the decision, support for it increases.7 At the same time, we see further evidence that reactions differ based on partisanship and ideology in Models 3 and 4, with increases in health care approval concentrated among liberals. Page 155 →Fig. 9.1. Approval for the ACA Fig. 9.2. Effect of Ruling on ACA Approval. (Note: Bars represent 95 percent confidence intervals on Obamacare approval change and Obamacare constitutional change using a onetailed t-test. No Republican changed his/her assessment of the law’s constitutionality postruling.) Page 156 → Table 9.2. Changes in ACA Approval

(1)

(2)

(3)

(4)

Demographic characteristics 0.01 в€’0.01 в€’0.02 в€’0.03 Female (0.04) (0.04) (0.04) (0.04) в€’0.06 в€’0.06 в€’0.06 в€’0.06 Age (0.06) (0.06) (0.06) (0.06) в€’0.13*в€’0.14*в€’0.21***в€’0.23** Income (0.07) (0.07) (0.07) (0.08) Black Hispanic

в€’0.01 (0.06) в€’0.06 (0.12)

0.01 (0.06) в€’0.05 (0.12)

0.03 (0.08) в€’0.01 (0.11)

0.01 (0.08) в€’0.03 (0.12)

в€’0.03 (0.10) в€’0.13 (0.11) 0.01 (0.08) 0.00 (0.00) 0.00 (0.09)

0.25** (0.11) 0.18* (0.10) в€’0.08 0.08 0.00 (0.00) 0.09 (0.11)

0.32 (0.24) 0.21* 0.11 в€’0.17 (0.14) 0.00 (0.00) 0.07 (0.11)

Preruling predispositions Democrat

—

Liberalism

—

SC diffuse support

—

SC specific support

—

Knowledge

—

Preruling opinions —

—

Obamacare constitutional —

—

Obamacare approval

Interactive effects Democrat *approval Democrat *constitutional Democrat *apolitical Court Democrat *SC temp R-squared N

в€’0.42***в€’0.30** (0.09) (0.15) 0.09 0.03 (0.06) (0.10)

—

—

—

—

—

—

—

—

—

—

—

—

0.04 139

0.09 137

0.31 110

0.35 (0.36) в€’0.21 (0.31) в€’0.19 (0.29) 0.00 (0.00) 0.33 110

Note: Results are OLS coefficient estimates on Obamacare approval change with standard errors in parentheses. Results are substantively similar using an ordered logit model. *p < .1, **p < .05, ***p < .01. A number of potential (but largely unsatisfactory) explanations exist for the divergent partisan responses witnessed thus far. According to the most straightforward explanation, Democrats may have been more sympathetic to the ruling because they held more favorable attitudes toward the Court. This is a simple account of

policy agreement—to the extent that Americans view the Court as trustworthy and credible, they are likely to defer to its judgment when they learn about its decisions. But this account has no support here. There is no evidence that Democrats were more likely to be persuaded by the Court given their attitudes toward it; indeed, Republicans placed significantly more trust in the judiciary prior to the Obamacare ruling. To test this, I compare party differences across two items: a feeling thermometer rating of the Court and a more stable measure of diffuse support asking subjects to what extent they view the institution as too involved in political disputes. Using either measure, Republicans expressed greater preruling support for the Supreme Court than did Democrats (Mdiff SC temp = 16.05, p = 0.00, two-tailed; Mdiff apolitical Court = 0.21, p = 0.00, two-tailed). Franklin and Kosaki (1989), in their pioneering work on Roe v. Wade, suggest another explanation for partisan differences: a structural response to rulings. The structural response account accords with our expectation of increased polarization following the decision. According to the authors, we should observe divergent opinion responses when “between-group variance is high” (763). This is the case here, with significant predecision differences between Democrats and Republicans in approval for Obamacare. But in two respects, the structural effect does not materialize. There is no evidence that Republicans changed their attitudes toward the ACA in general. And there is no evidence that Republicans embedded most strongly in their social environment reacted to the decision in particularly negative terms. To test this core tenet of the structural response hypothesis, I first add a main effect term for discuss politics and an interactive term for Republican* discuss politics to Model 4 from table 9.2 (not presented). We would expect the coefficient on the interactive term to be negative and statistically significant if Franklin and Kosaki’s account were true, as this result would indicate Page 158 →that Republicans expressed low levels of support for the health care law after discussing the ruling in detail. But the estimated coefficient is not significant (p = .21). Another way to test the structural response explanation involves looking at only the strongest Republican identifiers, who should be most strongly embedded in their partisan environment. Among those who identified as Republicans or strong Republicans, no evidence shows that opinions about Obamacare became more negative in the wake of the Supreme Court’s ruling (Mapproval change = .03, p = .58, two-tailed). Another explanation for unchanged Republican attitudes suggests that the structural response hypothesis cannot describe reactions when a large portion of the respondents already express the lowest levels of support possible for the health care law. How, then, could Republicans take any more of an unfavorable view of it in the postdecision survey? But if we restrict our analysis to those respondents who did not strongly disapprove of the ACA, attitudes toward it remain unchanged after the ruling (Mapproval change = .02, p = .36, two-tailed). This pattern of opinion stability can further be seen in figure 9.1, where the percentage of respondents who “somewhat disapproved” of the ACA remains constant from pre- to postruling. There exist no strong explanations of why Democrats and Republicans had different patterns of opinion response to the health care decision. Some of the most likely accounts—differences in views about the Court, differences in structural environment—fall short. However, evidence clearly shows that the Court’s decision increased support for the health care law and that this increase was concentrated among only Democrats, a group that had been more ambivalent toward the law prior to the decision. The pattern provides partial support for our expectation of polarization: while Democratic approval of the ACA increased postdecision, Republicans failed to become more antagonistic toward the law. At the same time, the result further underscores the Court’s limited ability to persuade in even the most high-profile cases, as can be seen in the reluctance of Republicans to alter their support for the ACA. Table 9.2 also suggests that the tax consideration influenced reactions to the ruling. All of the models indicate that the wealthy responded unfavorably to the Court’s ruling, expressing less approval of the health care law after the justices upheld it (as can be seen in the coefficient estimates on income). The impact of the tax consideration is, of course, not the only way we can make sense of the connection between personal income and ACA, but it is the most readily plausible. These results suggest that the binary decision offered by the Supreme Court does not directly affect attitudes; rather, specific considerations raised by the Court and in the press matter (Johnson Page 159 →and Martin 1998; Clawson and Waltenburg 2003; Brickman and Peterson 2006).

The Impact of News Considerations: An Embedded Experiment The design of my study enables a more full exploration, using an embedded experiment, of how considerations and media coverage influenced support for Obamacare. I make comparisons across respondents who were randomly assigned to one of three conditions. One group—the control group—includes those respondents discussed previously in the chapter, who answered to pre- and postruling surveys without viewing any further material. Another group—the legal treatment group—was first asked to watch an actual segment from Fox News that used legal principles and Court-related images while offering criticism of the Obamacare decision. The segment shows symbols associated with the Supreme Court, including an image of Chief Justice John Roberts and a shot of the justices, in robes, talking with one another. A third group—the political treatment—was asked to watch a news segment from Fox News that characterizes the ruling by raising political considerations. The segment shows Obama defending the ACA as tax-neutral. The treatment and control groups otherwise responded to identical surveys in phases 1 and 2. Nonetheless, in balancing control and external validity, I lean heavily toward the latter, in contrast to chapter 8 (Campbell and Stanley 1966; Aronson et al. 1990; see appendix G for a further description of stimuli). In table 9.3, I present a model of Obamacare approval change for each of the three conditions. These results make plain an important fact: though the mean differences in opinion change are similar across conditions, its bases differ. In the control condition, three factors affect how respondents’ views of the ACA change: their income levels, partisan/ideological beliefs, and preruling opinions of the law. These patterns are somewhat unremarkable given the types of information to which Americans were exposed about the health care case. Consider the finding that income affects how approval of the law changed from pre- to postdecision. Many respondents in the control condition were likely learning of the depiction of the ACA as a tax for the first time in late June, which affected their attitudes conditional on their own personal economic situations. What happens when respondents are exposed to different sets of considerations in the experiment? For respondents who watched the law-focused Page 160 →news segment, political predispositions and income lose their influence. It is unclear why income no longer has a significant effect, though the stimulus may not have made tax considerations salient enough to matter. However, the diminishing impact of partisanship can be directly traced to the apolitical stimulus. Respondents who viewed the news segment were asked to consider the legal conclusions the Court reached, not the case’s political implications; in turn, political factors fail to influence these respondents’ attitudes toward the ACA. Table 9.3. Explaining Change in ACA Approval by Condition Control Legal Political Demographic characteristics Female Age Income Black Hispanic Preruling predispositions

в€’0.01 0.05** (0.02) (0.03) в€’0.03 в€’0.02 (0.03) (0.04) в€’0.10***в€’0.06 (0.04) (0.05) 0.01 0.05 (0.04) (0.05) в€’0.01 в€’0.03 (0.06) (0.06)

в€’0.01 (0.02) в€’0.06* (0.03) 0.11*** (0.04) в€’0.02 (0.04) 0.05 (0.06)

Democrat Liberalism Apolitical Court SC temperature Knowledge

0.12** (0.06) 0.09* (0.05) в€’0.04 (0.04)

0.10 (0.07) 0.07 (0.07) 0.01 (0.05)

0.13** (0.05) 0.10* (0.06) в€’0.12*** 0.04

0.08 (0.05) 0.04 (0.05)

в€’0.02 (0.08) 0.03 (0.07)

0.16*** (0.06) в€’0.06 (0.06)

Preruling opinions в€’0.21***в€’0.31***в€’0.27*** (0.04) 0.06 (0.05) 0.05 0.06 0.00 Obamacare constitutional (0.03) (0.04) (0.03) R-squared 0.31 0.28 0.37 N 110 112 117 Obamacare approval

Note: Results are OLS coefficient estimates on Obamacare approval change with standard errors in parentheses. Results are substantively similar using an ordered logit model. *p < .1, **p < .05, ***p < .01. Page 161 →The stimulus in the legal condition also enables us to consider the effect of legal imagery on support for Supreme Court rulings. Some recent research suggests that the persuasiveness of the judiciary stems in part from the symbols with which it is associated (Woodson, Gibson, and Lodge 2011), but there is no evidence here of such an effect. In the legal condition, respondents viewed pictures of the justices in their robes and listened to a fairly sober criticism of the ruling’s legal basis. But these images did nothing to increase diffuse or specific support for the Court; furthermore, Court support had an insignificant effect on Obamacare approval change in this condition. The results suggest that the judiciary’s ability to draw on its legitimacy and increase support for its output may be more limited in real-world contexts, where depictions of legal symbols occur in the midst of complex discussions about the Court and the law. Attitudes take shape in the political condition in response to an entirely different set of factors. As is to be expected, partisanship and ideology structure opinion about Obamacare, since the political news segment highlights disagreements between President Obama and Republicans over the ACA. When the press highlights these distinctions, Democrats and liberals express significantly more support for the health care law in the postdecision phase. The effect of income on Obamacare approval change in the political condition also warrants an exploration. In contrast to the control condition, the coefficient on income is statically significant and positive, indicating that wealthy respondents were more supportive of Obamacare after watching the politically focused news segment. At first glance, this effect is counterintuitive—why would we see the opposite pattern across conditions? But recall that respondents in the political condition were directly exposed to Obama’s argument that the ACA is not a tax. So wealthy respondents in the political condition directly reject the tax argument. Conversely, wealthy respondents in the control condition were more likely to view the law as a tax given the balance of real-world coverage to which they were exposed after the Court’s ruling. The effects of attitudes regarding the Supreme Court are more complex in this condition. Respondents who rated the Court warmly on a feeling thermometer prior to the ruling expressed significantly more support for Obamacare, but respondents who viewed the Court as apolitical expressed significantly less support for the law. Why might these two related attitudes have countervailing effects on opinion? The explanation lies in

respondents’ careful distinction between specific and diffuse support for the judiciary. Specific support—SC temperature—has no formal relationship to the view that the Court is not involved in the political process. Respondents may rate the Court favorably yet still view it as an actor that involves itself in political Page 162 →controversies. These respondents see nothing distasteful about the Court resolving the conflict over Obamacare; in fact, respondents who view the Court favorably express more support for the law after the judiciary upholds it. But diffuse support is different. Respondents that believe the Court is apolitical are expressing a belief about its proper institutional responsibilities. The Court, they believe, does not (and should not) involve itself in political disputes. When presented with evidence to the contrary (the political news segment), these respondents reject the decision reached by the Court. In fact, as the coefficient on apolitical Court shows, they express significantly less support for Obamacare after the justices uphold it. In sum, the results from the embedded experiment suggest that the effect of news coverage on opinion is minimal on average; there are no statistically significant differences in Obamacare approval change across conditions. But news coverage does alter the basis of approval change, since an important link exists between the considerations to which people are exposed and the opinions they express (Zaller 1992; Brickman and Peterson 2006). In the control condition, partisanship and views about taxation structure attitude change toward the ACA. When respondents are exposed to legal frames about the rulings, the effect of partisanship predictably disappears. But somewhat counterintuitively, legal frames and imagery do little to make the Court more persuasive. Exposure to political coverage, however, reinvigorates the partisan and ideological bases of support for the ACA while leading to backlash among respondents who view the Court as unwisely entering the political arena. Most interestingly, the political news segment allows Obama to effectively rebut the taxation argument in a manner so effective as to suggest that the Court’s persuasiveness on this point depends on the information conveyed by the media to the public. The Law and the Court: Surprising Patterns of Influence Did the Court damage or enhance its standing with its health care ruling? According to the negativity bias account, we should expect the former, as the institution expended political capital by becoming involved in the controversy (Mondak 1992; Grosskopf and Mondak 1998). According to the positivity bias account, conversely, Americans should express more trust in the Court after the decision, given their exposure to symbols of judicial legitimacy (Gibson, Caldeira, and Spence 2003b). But here, using an item from Gibson, Caldeira, and Spence’s preferred measure of diffuse support, there is no evidence that the Court’s standing either suffered or improved after the Page 163 →decision (Mapolitical Court change = в€’0.02,p = 0.53, two-tailed). In fact, attitudes toward the Court appear to diverge along partisan lines, with Democrats placing increased trust in it (Mapolitical Court change D = 0.06, p = 0.03, two-tailed) and Republicans moving more strongly in the opposite direction (Mapolitical Court change R = в€’0.14,p = 0.00, two-tailed).8 The result indicates that changes in attitudes toward the Court may be best understood not as a general trend in the population but rather a specific reaction that can differ across subgroups depending on their policy-related attitudes. But perhaps the most important pattern in opinion reactions to the health care ruling is the Supreme Court’s inability to draw on its diffuse or specific support to persuade. At no time do we find evidence that preruling support for the judiciary leads to approval of its policy output. This can be seen in the estimated coefficients on apolitical Court and SC temperature presented in table 9.2, none of which is significant. Nor do attitudes toward the Supreme Court matter among certain groups—for example, there is no evidence that the Court persuaded only those Republicans who viewed it in favorable terms. In table 9.4, I test an alternative explanation: people’s attitudes toward health care influenced diffuse and specific support for the judiciary. The panel data enable me to do so in a manner that disentangles causality by using attitudes toward the law before the ruling to explore changes in Court-related attitudes after it. The effect is clear, unmistakable, and somewhat unexpected: attitudes toward the ACA significantly affect evaluations of the Supreme Court (and not vice versa). The pattern holds for specific support for the institution (Model 1) and even

for diffuse support for it (Model 3). The importance of these findings should not be understated. Existing evidence suggests that diffuse support for the judiciary influences attitudes about its rulings, and not vice versa. This occurs even when the Court becomes involved in some of the most fraught political controversies, like the 2000 presidential election dispute (Gibson, Caldeira, and Spence 2003b). But the panel data here provide the clearest evidence of causality, which runs from issue-specific attitudes to seemingly more stable views about the institution. Can we rescue the causal story of Court influence in another way? Perhaps respondents are quite sophisticated in their understanding of constitutional law and use their views about the Constitution to inform their attitudes toward the Supreme Court. This would be a conditional hypothesis of judicial influence: as long as Americans believe that the Court interprets the Constitution with accuracy and fairness, they are willing to defer to its judgments. So in the health care case, respondents might have lost confidence Page 164 →in the Court not because of their issue-specific attitudes but rather because they viewed its legal reasoning as inadequate. There is no evidence for this perspective. To test it, I include two potential explanatory factors of opinion change about the Court in Models 2 and 4: preruling approval of Obamacare and preruling views about its constitutionality. The estimated effect of approval alone is significant, indicating that respondents used their specific approval for the ACA to evaluate the Court, regardless of their views about the law’s constitutionality. Table 9.4. The Influence of Health Care Opinion on Attitude Change about the Court SC Temperature Apolitical Court (1) (2) (3) (4) Demographic characteristics в€’0.05 в€’0.05 0.02 0.01 Female (0.03) (0.04) (0.05) (0.06) 0.04 0.04 в€’0.03 в€’0.01 Age (0.06) (0.06) (0.09) (0.10) в€’0.01 в€’0.02 0.08 0.08 Income (0.07) (0.08) (0.10) (0.12) 0.04 0.06 в€’0.05 0.04 Black (0.06) (0.08) (0.09) (0.13) в€’0.03 в€’0.03 в€’0.15 в€’0.09 Hispanic (0.11) (0.12) (0.18) (0.18) Preruling predispositions 0.11 0.06 (0.12) (0.16) 0.02 в€’0.08 (0.11) (0.16) в€’0.06 в€’0.08 (0.11) (0.13)

0.19 (0.18) в€’0.14 (0.17) 0.07 (0.17)

0.31*** 0.36*** 0.34*** (0.06) (0.09) (0.10) −0.05 Obamacare constitutional — — (0.06)

0.37*** (0.14) в€’0.10 (0.09)

Democrat Liberalism Knowledge

0.04 (0.10) 0.03 (0.10) в€’0.02 (0.09)

Preruling opinions Obamacare approval

R-squared

0.32

0.36

0.15

0.17

N

136

109

136

109

Note: Dependent variables are SC temperature change and apolitical Court change. Results are OLS coefficient estimates with standard errors in parentheses. Results are substantively similar using an ordered logit model. *p < .1, **p < .05, ***p < .01. Fig. 9.3. Changes in Institutional Support by Strength of Preruling ACA Attitudes. (Note: Changes in views of an apolitical Court with error bars representing 95 percent confidence intervals from one-tailed t-tests.) Page 165 →The evidence paints a remarkable portrait that has not previously been seen in the literature.9 But it is instructive to consider how individual attitudes take shape. The attitude strength account posits that strong policy preferences alter the relationship between diffuse and specific support. According to this explanation, the Obamacare decision is unique because it is a rare case in which people’s preferences toward the policy were more durable than their loyalty toward the Court. And indeed, some evidence shows that the strength of policy preferences influenced attitudes toward the Court. In figure 9.3, I present the mean change in diffuse support for the Court across respondents based on their preruling attitudes toward Obamacare. If attitude strength plays a role in evaluating institution vis-Г -vis policy, the most pronounced changes in diffuse support should take place among respondents with the strongest opinions on Obamacare.10 This is indeed the case. The Court suffers a significant decline in trust among only the strongest opponents of Obamacare. But as the first panel of figure 9.3 shows, the Court’s standing remains unchanged for respondents who indicated that they only somewhat disapproved of the law. Similar patterns can be seen in the other panels, where changes in supportPage 166 → for the Court are always the greatest among those with the strongest preruling opinions of the ACA. Changes in diffuse support have the relative magnitude and direction suggested by a theory of attitude change. While we cannot offer a full experimental test of the attitude strength account here, this evidence suggests that the potency of preferences plays an important role in relating institutional loyalty and policy support.

Discussion The Supreme Court altered the shape of attitudes toward the Patient Protection and Affordable Care Act in a subtle but significant way. Lukewarm feelings about the law and dim views of its constitutionality changed between June and July of 2012. Respondents expressed higher approval for the law and were more likely to view it as a constitutional exercise of power after the Court’s pronouncement, though most of these shifts took place among Democrats who had expressed some prior ambivalence about the ACA. Republican attitudes toward the law remained essentially unchanged. But the pattern of Supreme Court influence was quite complex. Diffuse support for the institution had no impact on attitudes toward its ruling; in fact, the opposite influence transpired. The findings offer some of the first evidence that in very specific instances, people may use their issue-specific attitudes to reshape their diffuse support for the judiciary. What can the Obamacare case add to the voluminous yet often divergent findings in the literature on the Court and public opinion? Consider how the findings fit with existing theories of Court influence: Policy agreement: There is no evidence that the ruling created consensus about Obamacare, despite the high-profile nature of the case; rather, positive opinion responses were concentrated among Democrats and liberals. These findings cast further doubt on the Court’s ability to create consensus, particularly when the public is polarized prior to a ruling and when the media highlight complexities and/or criticism of the decision. Structural responses: While Democrats and liberals became more supportive of Obamacare after the ruling, there is no evidence that Republicans became less supportive of it. In fact, Republican views

of the law did not change, even if we look at the most strong Republican identifiers or at those Republicans who expressed some Page 167 →support for the law prior to the ruling. These patterns provide incomplete evidence in favor of a structural response/polarization perspective. Preexisting opinion and interpretive context: The Court’s decision made salient a new piece of information—the ACA as a tax—that had a small but clear effect on public opinion. Similarly, experimental evidence suggests that exposure to specific information across different types of news coverage reshaped attitudes toward the law. Legitimacy theory and positivity theory: There is no evidence that the Court damaged or enhanced its reputation on average by involving itself in the health care controversy. But more remarkably, there is no evidence that Americans used their beliefs about the Court’s apolitical nature to guide their views about Obamacare. Rather, the opposite pattern occurred. The findings with respect to the conditional influence of the Court suggest the central role that the media played in shaping reactions to the high-profile case. The ACA ruling casts substantial doubt on theories of Supreme Court influence that posit a case-invariant relationship between the institution and public opinion. In this case, we observe not only strong partisan disagreements about the merits of the ruling and the ACA but also distinct patterns of partisan response that belie a polarization hypothesis. While the ruling is persuasive to some Democrats, it has essentially no effect—positive or negative—on Republicans. The case also provides further evidence that issue-specific considerations play a role in shaping reactions to judicial rulings (Brickman and Peterson 2006). The Court does not simply choose between two outcomes when it rules. Rather, the majority opinion upheld the ACA by redefining the basis of the law in the public sphere. By making salient the taxation consideration, the decision distinguished respondents’ attitudes based on their income levels. The Obamacare case also adds a new wrinkle to our knowledge about policy agreement and the Court, showing that attitudes toward the institution and the law can color one another under specific circumstances. This finding qualifies our understanding of institutional legitimacy: the Court can increase support for a policy by conferring its stamp of approval, but only to the extent that media coverage is deferential to the Court and that attitudes toward the institution are stronger than attitudes toward the policy. Consider how the attitude strength account may help make sense of two of the most controversial decisions in recent history. After the Court struck down bans on flag burning in Texas v. Johnson (1989), Americans expressed Page 168 →outrage at the decision (Grosskopf and Mondak 1998), as a movement took shape to amend the Constitution and punish the Court (Hanson 2008). The ruling in Kelo v. City of New London, which broadened the uses of eminent domain, had a similar effect. Such cases present a puzzle for a narrowly tailored view of Court influence that sees the institution as quite persuasive if Americans are aware of its decisions (see Persily 2008, 12–13). Why, then, would a Court ruling lead to public outcry and attacks on the judiciary? Alongside press coverage, attitude strength may play a central role. Indeed, in both of these controversies, American preferences for the protection of the flag and of private property might well have been stronger than their diffuse support for the Court itself. When the Court was depicted as having overturned these protections, Americans might have penalized the institution even as their policy-specific preferences remained intact.11 Because of its panel design, the Obamacare study makes an additional contribution to the literature by controlling for preruling attitudes toward the Court. These attitudes are a central component of legitimacy theory, implied by the tenet that the Court can increase policy approval when it rules. By what mechanism might this transpire if not a microlevel one? Would we not expect those holding the most favorable attitudes toward the Court to approve most heartily of the policies it endorses? A test of individual-level opinion change over time fails to provide support for this theory, however. Respondents essentially ignore their views about the Court in shaping their attitudes toward Obamacare postruling. Viewed alongside the results of chapter 8, this finding raises serious questions about whether the Court’s legitimacy affords it strong persuasive currency in the struggle over public policy or whether it is merely a bulwark against politicized interference with its responsibilities.

Generalizability Chapters 8 and 9 review experimental and panel evidence, demonstrating the Court’s relatively limited ability to draw on its institutional reputation to boast approval for the specific policies. However, it is important to consider fully the weight of this evidence. After all, convenience samples of populations from MTurk are clearly not representative of the general population. This fact should give us pause before we make broad generalizations about popular opinion on the issues studied here. Conversely, concerns about the quality of the data presented in these two chapters should be relatively limited. New research attests that when they utilize best-practice design principles,Page 169 → studies using MTurk produce high-quality data. Scholars have in fact recently replicated published experiments in political science and psychology using pools of MTurk subjects (Horton, Rand, and Zeckhauser 2011; Berinsky, Huber, and Lenz 2012). We are left, then, with high-quality data but concerns about drawing generalizable inferences from any single study. However, we have much more than a single study from which we can learn. The volume and diversity of the evidence presented in chapters 8 and 9 are among the best reasons to place confidence in our results. It would be one thing if an entire story of the Court’s inability to persuade rested on a sample of 165 subjects (e.g., chapter 8, study 1). Media effects, in this case, might be traceable to a unique or idiosyncratic subject population, to contextual factors in the real world (perhaps the experiment was administered during a time when trust in the Court had ebbed), or to experimental artifacts. It is quite another to have broadly consistent results across a range of studies, as we have here. All told, chapters 8 and 9 present evidence from nearly 1,200 individuals, six distinctive political controversies (religious freedom, religious expression, gun rights, private property rights, term limits, and Obamacare), three time periods (the winter of 2012, when popular attention to the Court was limited; the summer of 2012, both before and immediately after its high-profile health care decision; and the winter of 2012–13, when public attentiveness had again waned), and very distinctive treatments. Despite this diversity, the results across studies are broadly consistent, demonstrating that the Court’s influence is both limited and heavily dependent on context. But perhaps there is something unique about MTurk participants that introduces a similar bias into all of these studies. To be sure, the MTurk subjects surveyed here were distinct in a very clear way: their high levels of interest in and knowledge about the political process. (After all, these subjects elected to participate in a variety of surveys that were explicitly about political affairs.) This feature of a subject population, however, can be found in many experimental studies, yet it has not hampered scholars’ ability to draw meaningful insights about public opinion. In fact, high levels of political knowledge may make findings about the Court’s limited influence less likely, since knowledge is often associated with high levels of trust in the Court (Gibson and Caldeira 2009b). To the extent that participants in chapters 8 and 9 are more trusting of the Court than the general population, we should expect to see a higher degree of Court influence (the opposite of what we observe). We can place confidence in inferences to the extent that they have accrued a diverse array of evidence in their favor, both here and elsewhere. That Page 170 →being said, we can be most confident in the finding that judicial legitimacy does not afford the Court the ability to resolve any and all political controversies in the public sphere. Rather, there are significant constraints on the institution’s power to persuade. We can also be quite confident that one such constraint, across a broad range of issues, is the manner in which the press frames judicial decisions for public consumption. This fact is attested to by the conservative treatments that produce substantively meaningful effects in chapter 8, by the embedded experiment in chapter 9, and by a range of other evidence that suggests that the media shape public evaluations of Court decisions (Mondak 1994; Hoekstra 2000; Clawson and Waltenburg 2001; Baird and Gangl 2006). On a related note, the findings presented here suggest for one of the first times that that the Court’s inability to persuade stems from the fact that individuals’ trust in the institution becomes divorced from their specific policy views when the Court comes under attack. Evidence for this can be observed most clearly in study 1 of chapter 8 and throughout chapter 9. To a lesser degree, we can place confidence in the generalizability of findings presented in chapter 9: policy approval can exert a causal influence on views about the judiciary, and the strength of preferences toward both institution and policy can shape how they interact with one another. Evidence for these findings is clear, but comes from what may be a singular case in the history of the Court. Much more work needs to be done to explore the role of attitude strength

in generating changes in opinion about both institution and policy after the Court rules. Nonetheless, the Obamacare decision helps further our understanding of the critical role played by the media and of the processes by which institutional loyalty influences policy approval. And it begins to lay bare the conditions under which the causal arrow may run in the opposite direction.

Page 171 →

Chapter 10 Conclusion The U.S. Supreme Court does not release its decisions in a vacuum. The justices have neither the tools nor inclination to communicate simple judgments about the Constitution and public policy directly to the American public. The reasons for this are many: the law is often too complex and technical to evaluate in short written opinions, the justices themselves are accomplished jurists with lifetimes of experience writing for professional audiences, the Court as an institution places comparatively little value on making its work transparent and framing its decisions for public consumption. At the same time, the public often lacks both the expertise and motivation to wade into legal minutiae. To be sure, the Court maintains a relatively durable reserve of goodwill that both makes other political actors leery of criticizing it too sharply and engenders unease among many Americans about the potential erosion of its responsibilities. But it is one thing to recognize that the judiciary is trusted and quite another to assume that it can mold popular opinion and facilitate consensus with its rulings. Accounts of the Court as a consensus builder make this case, but the argument too often has been belied by a reality in which the American public responds with indifference or even displeasure toward Court decisions. Other accounts of Supreme Court influence have detailed a variety of factors that may constrain the institution, including the preexisting opinions held by Americans and the structural context that shapes their understanding of rulings. But missing from these accounts is the actor that plays the central role in defining popular understanding of the institution’s output: the press. For ordinary Americans, the law of the land is defined as much through the publications of journalists as it is through the pens of the justices. The media might play a less influential role if they communicated only Page 172 →the conclusions reached by the majority of justices who speak for the Court in a given controversy. If it focused only on the text of John Roberts’s opinion in the dispute over Obamacare, for example, or only on the tempered rhetoric employed by the majority in Kelo, the press might have given the Court greater currency with which to persuade the public. But it did not. Rather, in these instances and in many others, coverage of judicial decisions eschews deference for controversy. Indeed, variation in the coverage afforded high-profile Supreme Court rulings is quite pronounced. While the small number of unanimous holdings—which signal to reporters the need to provide simple and accurate accounts of widely accepted legal principles—attract uncontested depictions of the Court’s reasoning, many other nonunanimous decisions garner much more complex (and often critical) coverage. The arguments and framing perspectives employed by dissenting justices in their written opinions are sure to receive attention since they provide reporters with the opportunity to emphasize conflict and controversy over the law on the high court. But for the media, not all dissenting opinions are created equal. Dramatic, compelling, evocative language—“all private property is now vulnerable,” or “the Court today decides to save a statute Congress did not write,” as the dissenters charged in the Obamacare case—has a disproportionate influence on the contours of news coverage. Critical coverage is often further animated by arguments raised by issue activists, intemperate rhetoric employed by cable news pundits, and charges of judicial activism levied at the justices themselves. As scholars seek to understand the promise and peril associated with Supreme Court persuasion, these variations lay plain the importance of the press. Most existing work interrogates the influence of the Court alone on popular attitudes following a landmark ruling. Did the stamp of legitimacy cause Americans to defer to the Court’s judgment? Did its opinion exert a systematic persuasive influence? These questions raise important theoretical issues, yet their practical value is questionable. In the Kelo case, for example, it is entirely possible for one person to have watched a single news segment that characterized the decision as a limited instance of deference to local officials, while another individual who tuned in just minutes later to have heard a heartrending account of

Americans having their homes seized on the strength of the decision.1 Such differences in coverage may animate the diverse array of popular reactions that follow landmark rulings. But what of judicial legitimacy? Might the Court not still exert a persuasive influence, building popular consensus regarding the policies it endorses as a result of the trust most Americans place in it? Evidence suggests that is Page 173 →not the case, as Americans can distinguish their views about institution and policy in a rather sophisticated fashion. Across a range of issues, critical media coverage depresses support for Court-endorsed policies though not for the institution itself. These effects are most pronounced for individuals who place the highest levels of trust in the Supreme Court: they readily approve of policy when the press offers a sympathetic perspective on a decision but just as readily eschew their deference to the Court when its decisions come under fire. The pattern raises questions about the nature of judicial legitimacy and the individual-level processes that tie together institutional trust and policy-specific approval. It is unclear whether the Court’s stamp of legitimacy has any real currency in increasing support for controversial policies and fostering consensus, given the contested media environment through which its decisions reach most Americans. This finding casts doubt on many of the most optimistic theories of Court influence, which view the institution as capable of building consensus and acting as an unbiased mediator of political controversies in a polarized society. But a more limited view of judicial legitimacy remains appropriate. Though the perspective may well be tested anew in the wake of the Obamacare case, some evidence suggests that diffuse support for courts exerts a powerful influence on policy acceptance if not policy approval. This influence should not be taken lightly, since popular acceptance of the legitimacy of a Court decision ties closely to compliance with it, an important goal for an institution that lacks a formal enforcement mechanism. In a landmark study, Johnson (1967) found evidence that an Illinois school district ensured compliance with the Court’s school prayer decisions in spite of many residents’ private disagreements with the rulings. According to another study, “The degree to which people ascribe legitimacy to an institution does affect their willingness to acquiesce to its decisions” (Gibson, Caldeira, and Spence 2005, 196). Therefore, the most appropriate way in which we may conceive of judicial legitimacy is not as a persuasive currency but quite simply as a durable reservoir of goodwill that remains even when Americans are dissatisfied with specific outputs. To this perspective, we have added an important qualifier from chapter 9: belief that the Court exercises unbiased, apolitical decisionmaking procedures may erode, at least in the short term, when policy-specific attitudes are more potent than institutional support. The Court’s reservoir of goodwill affords it significant clout, if not the ability to readily persuade in the face of criticism. As Gibson, Caldeira, and Spence (2005) note, “The most important attitudes ordinary citizens hold toward institutions like the Supreme Court have to do with institutional Page 174 →loyalty. Institutions like courts need the leeway to be able to go against public opinion” (356). This is true both because of one of the traditional roles of courts—as a safeguard for the rights of minorities—and because courts themselves lack enforcement powers. It is thus imperative that citizens maintain their loyalty to the institution to ensure that its jurisdiction is safeguarded and that its policies (even the undesirable ones) are carried out. In these respects, the Supreme Court fares quite well. Even after the Court has released some of its most bitterly contested decisions in recent years, serious movements to resist their enforcement and to censure the justices remained almost nonexistent. The Republican nominee for president in 2012, Mitt Romney, for example, had made clear throughout the campaign his opposition to Obama’s national health care law. But he offered only a measured response to the Court’s ruling upholding the ACA, evincing an unwillingness to attack the institution in strident terms. Twelve years earlier, Al Gore ended a rancorous dispute over the rightful winner of the presidential election by quickly conceding the race after the Supreme Court ruled against him. The legitimacy of the Court in even these cases, it seems, proved a powerful deterrent for politicians as they weighed whether to launch aggressive attacks at the institution. And, as we have seen, judicial legitimacy has the ability to shape elite reactions to the specific policies the Court endorses as well. Evidence in Part 1 of the book demonstrates that political elites are reluctant and reporters are unwilling to denounce rulings in which the justices speak with a single voice. For this reason, we can say that the

single-largest threat to the Court’s ability to garner favorable media coverage and approving public responses for its decisions is internal. The choice facing each justice as he or she weighs the decision to join the majority in a given case implicates not only personal ideological preferences and the attendant strategic considerations but also the reception the ruling will receive from a conflict-attuned press and an often-agnostic public. This study’s focus on the Court-press-public dynamic has numerous implications for existing lines of research involving public opinion, political communication, and judicial decision making. The attitudinal model has played a key role in previous accounts of judicial behavior, suggesting that justices base their votes solely on personal policy preferences (Segal and Spaeth 2002). Scholars continue to refine this perspective, demonstrating that the justices also take into consideration legal principles (Knight and Epstein 1996; Clark and Carrubba 2012), the preferences of other political actors (Segal 1997; Sala and Spriggs 2004; Harvey and Friedman 2006; Clark Page 175 →2009) the state of public opinion (Marshall 1989; Mishler and Sheehan 1993), and the identities of claimants (Carrubba et al. 2012). Many of these approaches recognize a role for strategic behavior on the bench (see also Maltzman, Spriggs, and Wahlbeck 2000). So, for example, justices have at times preferred unanimous decisions to guard the legitimacy of the institution (Epstein, Segal, and Spaeth 2001). But the results presented here raise questions that may further refine the inquiry into voting behavior on the Court. Are justices aware of the effects their voting decisions have on popular opinion? And do they actively attempt to shape the type of press coverage a ruling receives with their votes? Both of these questions fit broadly within the realm of strategic behavior on the bench, which, recent evidence suggests, is much more prevalent than previously believed (Epstein and Knight 2013). Evidence indicates that Supreme Court justices meet the first necessary condition for such strategic behavior: an awareness of the important role the media play in shaping popular understanding of the Court and its members. Davis (2011) demonstrates that the justices are increasingly interacting with the press, public, and personal audiences for “individualistic external” goals. The choice to go public, which can be done in a variety of ways, can buttress a justice’s popularity and help him or her achieve career goals and triumph in personal conflicts on the Court. But does the motivation to go public influence the written opinions released by the justices? In a word, yes. According to Davis (2011), “Opinions are not just directed at the legal community. They are also intended to reach the public” (13). The justices’ concern with public opinion can be witnessed in a range of highprofile cases, beginning with Kelo, in which Sandra Day O’Connor spoke in easily interpretable, emotionally arresting terms with the apparent intent of reaching a popular audience. “The beneficiaries” of the decision, O’Connor warned, “are likely to be those citizens with disproportionate influence and power in the political process” (Kelo v. City of New London, O’Connor dissenting, 505). The strategy of undermining support for a Court decision through a dissent can be seen in a number of other cases. In Arizona et al. v. U.S. (2012), for example, the Court ruled that federal immigration law preempted an Arizona statute. In a stinging dissent, Justice Antonin Scalia took aim at the ruling: “Are the sovereign States at the mercy of the Federal Executive’s refusal to enforce the Nation’s immigration law?” (Arizona et al. v. U.S., Scalia dissenting, 392). In Ledbetter v. Goodyear Tire & Rubber Co. (2007), Justice Ruth Bader Ginsburg read an impassioned dissent from the bench, asserting, “The ball is in Congress’s court.В .В .В . [T]he Legislature may act to correct this” decision (Ledbetter v. Goodyear Tire Page 176 →& Rubber Co., Ginsburg dissenting, 661). These examples, and many others, attest that Supreme Court justices have one eye on the American public as they craft their opinions in high-profile cases. The strategic concern with a decision’s popular reception is not limited to dissenters. As we have seen, large majority coalitions on the Court lead to favorable portrayals of its rulings. And other evidence suggests that specific information presented in media coverage of rulings—for example, descriptions of procedural fairness in judicial decision making, references to unanimity on the Court, and depictions of legitimating symbols (Baird and Gangl 2006; Zink, Scott, and Spriggs 2009; Woodson, Gibson, and Lodge 2011)—can buttress support for both the Court and the policies it endorses. There are therefore reasons to expect that justices may seek to broaden majority coalitions to increase popular approval for rulings and the institution itself. And, in fact, examples of just

such behavior in the Court’s history include the drive for unanimity led by Chief Justice Earl Warren in Brown v. Board of Education and Chief Justice Roberts’s explicit statement of preference for “unanimity or near unanimity” on the bench (Davis 2011, 13). Clearly, then, justices in both majority and dissent are cognizant of and actively engaged with public interpretations of their written opinions. But the evidence is largely silent as to whether the public opinion concern can shift voting behavior in a case (or whether it merely shapes the types of opinions the justices craft). One useful framework worthy of exploration in future work would posit a balance between two potentially contradictory concerns that the justices face: buttressing the Court’s reputation and ensuring compliance with its decisions on the one hand, and ensuring a most-preferred-policy outcome even if it damages the institution on the other. These concerns are weighed in each case as a justice chooses whether to dissent (and what type of language to use in a dissenting opinion). If the strategic concern with public opinion were dispositive, one would expect distinctions in a justice’s voting behavior based on the popular profile of a case at hand. For example, to the extent that the chief justice is uniquely concerned with maintaining the Court’s reputation, he or she would be less likely to dissent in high-profile cases than others. To this point, there is some evidence that the justices evince a concern with popular opinion by behaving differently in high-profile cases (Unah and Hancock 2006; Bartels 2011), but much more work remains to be done on this topic. A significant portion of the voluminous literature concerning political communication explores the quality and content of the news (for an overview, see Graber and Smith 2005). This research regularly finds important links Page 177 →between mainstream government debate and media coverage (Bennett 1990). According to Althaus et al. (1996), “We should distinguish three ways of segmenting U.S. elites: governing elites as a whole; the executive branch or administration, which typically initiates foreign policy; and oppositional officials, whom the media generally identify among members of the opposition party in Congress. Together the administration and its critics in Congress compose the governing elite” (408). The vast majority of research involving media-government relations focuses on executive branch actors (Bennett 1990; Bennett and Manheim 1993; Hart 2002; Entman 2004; Campbell and Jamieson 2008; Owen and Davis 2008). The volume of research on communication strategies employed by the George W. Bush administration alone, for example, is impressive (Bennett, Lawrence, and Livingston 2006; John, Domke, and Coe 2007; Winkler 2007, 2008; Benoit and Henson 2009). A large subset of this work considers foreign policy crises in particular. Other areas of active research include new media institutions (Livingston and Bennett 2003; Xenos and Bennett 2007; Sobieraj and Berry 2011) and political communication in campaigns (Spillotes and Vavreck 2002; Falk, Grizard, and McDonald 2006; Benoit and Arne 2009; Gibson et al. 2010). Judges and judicial branch sources are conspicuously absent from these formulations. Such an imbalance is troubling. In fact, in studies of the relationship between elite influence and news coverage, researchers often pitch their findings too broadly since the elites in question are almost always executive branch actors. To what extent can we generalize these findings to other areas of the government? Does the reliance of studies on executive branch elites affect the conclusions they draw? This book suggests that models of elite influence capture an important portion of the Court-press dynamic. It demonstrates that the media’s reliance on elite expertise and willingness to index coverage to elite voices allow Supreme Court justices to use their votes to shape how their rulings will be framed. Consensus among these elites causes reporters to offer a unified and deferential perspective in coverage. But division among the justices opens the Court to framing challenges both internal and external. In this respect, elite disagreement is similarly influential whether it occurs on the bench or inside the White House. But the dissensus dynamics model also demonstrates how focusing on presidential administrations obscures other aspects of news production. Existing research has largely overlooked what we might call elite source credibility, which captures the extent to which various high-level sources have an incentive to offer counterframing perspectives. Given the durable nature Page 178 →of its legitimacy, the Court may have a higher level of inherent credibility than other branches of the federal government. Because Americans penalize aggressive attacks on the

Court (Shesol 2010), political actors have limited incentives to express disagreement with it.2 Wielding this credibility, the justices have more power to frame than other actors. Attacks on the Court are unlikely except when it is internally divided. Indeed, evidence shows that political disagreement with the Court does not affect media coverage once we control for the size of its majority voting coalition. The dissensus dynamics model also shows the importance of strategic communication. Framing challenges occur as a reaction to culturally incongruent frames. Presidents are unlikely to define issues successfully when they employ frames that are incongruent with dominant schemas in political culture. But “a good match between a news item and habitual schemas pulls a frame into people’s thoughts with virtually no cognitive cost” (Entman 2004, 14–15). The examples of Lingle and Kelo refine this perspective. In both cases, the Court offered a decision that challenged prevailing cultural norms: private property rights were not sacrosanct but rather subject to government interference. But the media rejected the Court’s frame only in the latter case, where the majority failed effectively to communicate its decision (Baron 2007). Rather, the dissenters offered successful strategic arguments that were designed to inflame public sentiment regarding the ruling. These dissenters drew on their credibility—after all, they were seen as some of the finest legal minds in the country—and offered emotionally evocative arguments that had a high level of personal relevance for the general public (e.g., “All private property is now vulnerable”). A central contribution of this study, then, is to demonstrate how research on political communication has drawn broad conclusions while focusing on a narrow range of elites. By paying more systematic attention to press coverage of other actors such as the justices of the Supreme Court, scholars may refine and improve existing models of elite influence. Finally, these findings have a number of implications for the study of public opinion. I explore a process that links elite action and public response more fully than other research on policy agreement. Scholars have long suggested that elites influence public opinion through messaging in the media (Zaller 1992, 1996), but studies of political communication and public opinion often exist separately from one another. With respect to the Supreme Court, this separation has generated disputes rooted in methodological differences, since experimental studies offer an incomplete look at the information conveyedPage 179 → to the public. Future research on public opinion would do well to explore media effects in a more realistic context. Recent studies have already taken steps in this direction. Chong and Druckman explore the impact of multiple competing frames on public opinion (Druckman 2004; Chong and Druckman 2007b, 2012; Druckman et al. 2010). Their work demonstrates some important limits on framing effects in competitive contexts. I have built on this work here, albeit in a context specific to the Court alone, but the study of contradictory cues and frames remains in its early stages. Much more remains to be explained about how public opinion takes shape in a complex media environment. At this point, evidence shows that the link between the Court and public is less direct than has commonly been assumed. Because the institution depends on the media to translate its decisions to ordinary Americans, its influence is circumscribed. As this book demonstrates, the press does not defer to the wisdom of the justices in all cases but rather uses voting signals to craft its coverage of the high bench. In their landmark 1989 study, Franklin and Kosaki captured the optimism that many observers felt about the ability of the judiciary to build consensus in a politically fractious society: “If, through their opinions, the courts can instruct the public, enforcement is gentle but effective.В .В .В . The conception of the Court as republican schoolmaster generally reflects the notion that the Court, through its explication of the law and its high moral standing, may give the populace an example of the way good republicans should behave” (751–52). Such a possibility seems all the more promising more than two decades hence, at a time when many Americans lament the country’s increasing political polarization. Might the Supreme Court be able to make careful legal judgments in a nonpartisan manner, throwing the institution’s credibility behind contentious policies to ensure popular consensus? This question, perhaps more than any other, has animated interest in the relationship between the Court and the public.

And to be sure, the Supreme Court maintains a measure of legitimacy, particularly when compared with other institutional actors in the American political system. Americans generally trust the justices to make decisions in the best interests of the country as a whole and recoil at overt attacks on the Court. But this legitimacy has largely failed to afford the Court the ability to generate popular consensus. For their part, Franklin and Kosaki (1989) expressed skepticism that the Court could do so: “Citizens do listen to the schoolmaster but they also talk in classВ .В .В . public response is not a simple Page 180 →matter of increased support for the positions taken by the Court majority” (768). Both contentions are true, and both speak to the tempered influence of judicial decisions on public opinion. But what do Americans say when they talk in class? How do they make sense of Court decisions, and why does the majority often fail to persuade? These questions cannot be answered by focusing on the Court alone. Rather, careful attention to the national news media—the channel through which most Americans learn about judicial decisions—can best help us understand the persuasive capabilities of the judiciary and the many points at which its popular influence is circumscribed.

Page 181 →

Appendix A Coding Standards for Media Coverage Variables Frame Dominance 0 (Frame parity): If frame parity is present, critics of the Court/decision will receive extensive voice in the article. Their arguments, either summarized or directly quoted, will contest the ruling on substantive (and possibly legal) grounds, often offering differing interpretations of the Court decision and its impact. While the majority and/or its supporters may receive voice in the article, they will clearly fail to dominate the discussion and establish the superiority of their preferred frame. Content of article focusing on Court majority’s frame at the expense of opposition frames: ~0–50%. .5: The article presents the viewpoints of both supporters and critics of the Court’s decision. The Court’s opinion might be quoted and/or explained in detail, but in this case, dissenters and/or opponents will also offer their viewpoints. Alternatively, some articles may fail to mention the positions of the Court and its critics in great detail. Articles that briefly mention a decision outcome but then discuss other matters related to it (that do not implicate support for or rejection of the majority’s frame) should be coded here. Content of article focusing on Court majority’s frame at the expense of opposition frames: ~50–75%. 1 (Frame dominance): If the Court’s issue frame is accepted, the article will describe extensively what the Court ruled and why, frequently choosing to quote from the majority opinion itself. Opposing viewpoints (such as those of dissenters on the Court, or activists upset with the decision) might also Page 182 →be present in the article, but they receive very little space relative to the majority’s opinion and the reasoning behind it. Content of article focusing on Court majority’s frame at the expense of opposition frames: ~75–100%.

Supportive Tone 0 (Opposition/unsupportive): Articles unsupportive of the Court’s decision may reject the Court’s framing of the issue (by giving extensive voice to critics of the ruling) and express either explicit or implicit disapproval of the ruling. Implicit approval is associated with the absence of challenges to the critics’ positions. Alternatively, the article may be neutral with respect to the Court’s issue frame but highlight explicitly a position that is unfavorable toward the ruling. .5 (Neutral/balanced): The article may be neutral about whether it accepts the Court’s issue frame and in doing so fail to indicate support or opposition to the ruling through its language or tone. It may also give equal voice to both critics and supporters of the decision—similar in their level of vehemence—in equal measure. Alternatively, the article may accept or reject the Court’s frame but explicitly convey ambivalence about the decision overall. 1 (Supportive): Articles supportive of the Court’s decision may accept the Court’s issue frame (by giving extensive voice to the majority and/or supporters of the ruling) and express either explicit or implicit approval of these supporters. Implicit approval is associated with the absence of challenges to the Court’s positions. Alternatively, the article may be neutral with respect to the Court’s issue frame but explicitly highlight a position that is favorable toward the ruling.

Uncivil Debate (Cable News Sample Only) 0 (No): The program does not include two or more figures engaged in a back-and-forth discussion. Alternatively, a program may include the presence of multiple figures engaged in a discussion in a “civilized” manner. They may discuss the meaning and impact of a decision, with one of them acting as devil’s advocate for the

purposes of better understanding a ruling (not to generate a contentious substantive debate). Similarly, two speakers can debate the merits of a ruling, but if they do so in the most civilized manner Page 183 →(no charged language, no long/contentious exchanges, no interruptions of one another), the article should be coded here. 1 (Yes): The program includes two or more figures who debate the merits of an issue/decision in anything less than a cordial manner. Such debate is often characterized by charged language, lengthy back-and-forth discussions that take the form of an argument, and/or speakers interrupting one another.1

Activism 0 (No): No mentions of the term activism and its derivatives to characterize the decision. Brief or no mention of overturning a lower court and/or legal precedent. However, the article may use such phrases as groundbreaking, clarifies old law, answers new questions, and overturns lower court and still be coded as 0. 1 (Yes): The article clearly portrays the decision as activist. This is most typically done with an explicit use of the word activist or activism but might instead include a description of the Court’s overturning of a legal precedent. However, mentions of overruling a precedent should be explicit and obvious. Passing phrases about the establishment, clarification, and/or alteration of precedent do not indicate activism, but clear charges of an activist judiciary do. These charges might emphasize an unelected body making policy judgments or a deliberate refusal to honor precedent or law.

Page 208 → Page 209 →

Notes Chapter 1 1. SCOTUSblog, “Live Blog of the Health Care Decision,” accessed June 28, 2012, http://www.scotusblog.com/2012/06/live-blog-of-the-health-care-decision-sponsored-by-bloomberg-law/. 2. The idea of support stands in contrast to another manner in which the Supreme Court may influence public opinion: by increasing the acceptance/legitimacy of policies with which citizens disagree (see Mondak 1994; Woodson, Gibson, and Lodge 2011). 3. Many studies view the institutional legitimacy of the Court as a form of source credibility. 4. Josh Levs, “What the Supreme Court Ruled on Health Care вЂTax,’” accessed July 5, 2012, http://edition.cnn.com/2012/07/05/politics/scotus-health-care-tax/index.html; Bill Mears and Tom Cohen, “Emotions High after Supreme Court Upholds Health Care Law,” accessed June 29, 2012, http://edition.cnn.com/2012/06/28/politics/supreme-court-health-ruling/index.html; Kenneth T. Cuccinelli, “Victory in Defeat,” National Review, June 29, 2012. 5. Jeanne Sahadi, “Supreme Court: Mandate Penalty Is Tax,” accessed June 28, 2012, http://money.cnn.com/2012/06/28/pf/taxes/health_reform_new_taxes/index.htm; David G. Savage, “Supreme Court Upholds Healthcare Law as Tax Measure,” Los Angeles Times, June 29, 2012. 6. In the eminent domain case, the coverage paid even greater attention to the alarms sounded by four dissenting justices, led by Sandra Day O’Connor. The press recounted their criticisms that the decision abandoned a long-held basic limitation on government power, misapplying the Public Use Clause of the Fifth Amendment. Reporters also followed the emotional stories of ordinary homeowners like Susette Kelo in great detail. The Court’s majority opinion, however, received comparatively little attention. 7. New York Times, “Siding with the Liberal Wing,” June 29, 2012; Jan Crawford, “Roberts Switched Views to Uphold Health Care Law,” accessed July 1, 2012, http://www.cbsnews.com /8301-3460_162-57464549/roberts-switched-views-to-uphold-health-care-law.

Page 210 →Chapter 2 1. In some of these studies, evidence of backlash is incomplete because comparable pre- and postdecision opinion data do not exist. Nonetheless, the public viewed Court decisions in largely unfavorable terms in each case listed here. 2. Chris Cillizza, “John Roberts, Umpire,” accessed June 28, 2012, http://www.washingtonpost.com /blogs/the-fix/post/john-roberts-umpire/2012/06/28/gJQAx5ZM9V_blog.html. 3. In chapter 8, I develop a way to deal with questions about how to measure support for rulings. I create a decision approval scale with multiple items (support, approval, and so on).

Chapter 3 1. The justices themselves, however, pay careful attention to their public images (Davis 2011). 2. Throughout this book, I follow standard practice in using “the Court” to refer the majority coalition when discussing a particular ruling. 3. Nonetheless, the demand for timely coverage in this case was so great that CNN and Fox News initially reported incorrectly that the Court had struck down the Affordable Care Act. 4. USA Today, “Ruling Leaves Door Open to Abuse,” June 24, 2005; Don O’Brien and Kathleen Koch, CNN Newsroom, CNN, December 10, 2007, LexisNexis Academic. 5. A. J. Chavar, “Health-Care Ruling Reactions from the Supreme Court Steps,” accessed June 28, 2012, http://www.washingtonpost.com/politics/health-care-ruling-reactions-from-the-supreme-court-steps /2012/06/28/gJQA9kuJ9V_video.html. 6. Jeff Flock, News from CNN, CNN, June 23, 2003, LexisNexis Academic. 7. Liz Goodwin, “Conservatives Turn on Roberts, Supreme Court after Health Care Ruling,”

accessed July 16, 2012, http://news.yahoo.com/blogs/ticket/conservatives-turn-roberts-supreme-courthealth-care-ruling-184226194.html; S. A. Miller, “Supreme Court Decision on Obamacare Fuels Obama-Romney Campaign,” New York Post, June 29, 2012; Lewis Krauskopf, “Corporate Winners, Losers under Supreme Court Healthcare Ruling,” accessed June 29, 2012, http://www.insurancejournal.com/news/national/2012/06/29/253680.htm. 8. Matt Negrin, “Chief Justice John Roberts’ Switch on вЂObamacare’ Health Care Ruling Signals a Leaky Supreme Court,” accessed July 2, 2012, http://abcnews.go.com/Politics/OTUS/chiefjustice-john-robertss-switch-obamacare-health-care/story?id=16698557. 9. In chapter 4 and appendix B, I explore the relationships between issue-specific coverage and a broader measure of favorable content. 10. Furthermore, the attitudinal model of judicial behavior implies that a unanimous Court is unlikely to adopt a policy so far from the ideological mainstream that it would warrant strong criticism (Segal 1997; Segal and Spaeth 2002). 11. Michael D. Shear and Ashley Parker, “Romney Says He Will вЂRepeal Obamacare’ Page 211 →If Elected.” accessed June 28, 2012, http://thecaucus.blogs.nytimes.com/2012/06/28/romney-says-hewill-repeal-obamacare-if-elected/.

Chapter 4 1. Washington Post, “Judicial Takings and Givings,” May 29, 2005, LexisNexis Academic; Washington Post, “Eminent Latitude,” June 24, 2005, LexisNexis Academic. 2. A “regulatory taking” does not involve the physical taking of property but rather occurs when government restrictions become so onerous as to have similar consequences to a physical taking. 3. There is one possible exception. On June 20, the Supreme Court issued a decision in the case of San Remo Hotel v. San Francisco (2005), refusing to grant an exception to the full faith and credit statute for takings claims. The San Remo decision went largely unnoticed by the press and was soon overshadowed by Kelo. It appears to have had little effect on the popular discussions of property rights law from June 20 to 23 (before Kelo) or afterward. 4. One may question why the voting outcomes differed so dramatically in the cases if they were otherwise similar. Recent scholarship suggests that the justices have preferences not only about resolving the legal issues in a case but about resolving concrete disputes between parties (Carrubba et al. 2012). For this reason, we may speculate that the justices were more reluctant to rule against the New London homeowners than they were against Chevron. At the same time, the Court’s decision to remand in Lingle allowed it to avoid ruling directly against Chevron. Other property rights cases that have come before the Court feature more substantial distinctions than those between Lingle and Kelo. 5. This is the main drawback to the Lingle-Kelo comparison. But the alternatives are far less desirable for two reasons. First, Seawright and Gerring (2008) note that the most-similar design requires cases that differ on an explanatory variable and are similar along as many other potentially relevant dimensions as possible. No other two cases fit these criteria as well as Lingle and Kelo. Second, if we attempt to pair Supreme Court cases based on their perceived victims, we may have to go back as far as 1954, when the ruling in Berman v. Parker displaced thousands of low-income residents (Sax 2005; Nadler, Diamond, and Patton 2008). This introduces variation along a multitude of other dimensions that makes comparison with Kelo difficult. The best approach for making inferences about the factors that shape property rights coverage, then, involves a detailed exploration of the portrayals of the most similar cases (Lingle and Kelo). I take care to disentangle how any other differences between the rulings (beyond their voting outcomes) play out in the press. 6. All sources offer substantial coverage of property rights except Newsweek. By source, the number of articles/segments are New York Times, 71; Washington Post, 83; Washington Times, 51; USA Today, 19; CNN, 61; Fox News, 25; MSNBC, 30; Newsweek, 0; National Review, 17; and NPR, 37. 7. In Lingle, the Court makes no mention of the “victims” of its ruling. In Kelo, the majority and minority struggle over the impact of the decisions on blighted areas. More important, Thomas’s dissenting opinion expresses worries about the impact of Page 212 →the ruling. It mentions the poor, weak, blacks and other minorities and warns against the invasion of “traditional” property rights (Kelo v.

City of New London 2005, Thomas dissenting, 14). Based on an analysis of these opinions, I use the following search terms to denote victimization in the press: victim, invas, invad, intru, poor, weak, black, minor. 8. Peter Applebome, “Clear the Way, Fellows, the Yellow Brick Road Is Coming Through,” New York Times, May 8, 2005, LexisNexis Academic; Washington Post, “Defending the Neighborhood,” May 7, 2005, LexisNexis Academic; Kirstin Downey, “Revitalization Projects Hinge on Eminent Domain Lawsuit,” Washington Post, May 21, 2005, LexisNexis Academic; Richard W. Rahn, “A Run on the World Bank,” Washington Times, June 3, 2005, LexisNexis Academic; Martin Kasindorf, “Jackson Defense Strategy: вЂA Lot of Witnesses,’” USA Today, May 2, 2005, LexisNexis Academic. 9. Juliet Eilperin, “Legislators Working to Reshape Endangered Species Act,” Washington Post, May 20, 2005, LexisNexis Academic. 10. John Gibson, The Big Story with John Gibson, Fox News, May 6, 2005, LexisNexis Academic. 11. Bill O’Reilly, The O’Reilly Factor, Fox News, April 29, 2005, LexisNexis Academic. 12. Kirstin Downey, “Revitalization Projects Hinge on Eminent Domain Lawsuit,” Washington Post, May 21, 2005, LexisNexis Academic. 13. Diane Cardwell, “Ferrer and Mayor Clash on Brooklyn Rezoning,” New York Times, May 10, 2005, LexisNexis Academic; Deroy Murdock, “Scrap the Freedom Tower,” National Review, May 9, 2005, LexisNexis Academic; Peter Applebome, “Clear the Way, Fellows, the Yellow Brick Road Is Coming Through,” New York Times, May 8, 2005, LexisNexis Academic; David Nakamura, “Stadium Talks with Landowners Set to Begin,” Washington Post, May 5, 2005, LexisNexis Academic; Eric Fisher, “DCBSA Bid Gains Support on Council,” Washington Times, May 17, 2005, LexisNexis Academic. 14. Kirstin Downey, “Fighting the Power to Take Your Home,” Washington Post, May 7, 2005, LexisNexis Academic. 15. Linda Greenhouse, “In Free-Speech Ruling, Justices Say All Ranchers Must Help Pay for Federal Ads,” New York Times, May 24, 2005, LexisNexis Academic. 16. Washington Post, “Judicial Takings and Givings,” May 29, 2005, LexisNexis Academic. 17. Richard W. Garnett, “Unanimous!,” National Review, June 1, 2005, LexisNexis Academic. 18. Richard W. Rahn, “A Run on the World Bank,” Washington Times, June 3. 2005, LexisNexis Academic. 19. Steve Inskeep, Morning Edition, National Public Radio, June 7, 2005, LexisNexis Academic. 20. Soledad O’Brien, American Morning, CNN, June 24, 2005, LexisNexis Academic. 21. USA Today, “Ruling Leaves Door Open to Abuse,” June 24, 2005, LexisNexis Academic; Avi Salzman and Laura Mansnerus, “For Homeowners, Frustration and Anger at Court Ruling,” New York Times, June 24, 2005, LexisNexis Academic.Page 213 → 22. Soledad O’Brien, American Morning, CNN, June 24, 2005, LexisNexis Academic. 23. Tom Knott, “Half St. Citizens Continue to Fight Ballpark Bullies,” Washington Times, July 6, 2005, LexisNexis Academic. 24. Steve Inskeep, Morning Edition, National Public Radio, June 24, 2005, LexisNexis Academic; George F. Will, “Damaging вЂDeference,’” Washington Post, June 24, 2005, LexisNexis Academic; Scott Simon, Weekend Edition, National Public Radio, June 25, 2005, LexisNexis Academic. 25. Linda Greenhouse, “Justices Uphold Taking Property for Development,” New York Times, June 24, 2005, LexisNexis Academic. 26. New York Times, “Limits of Property Rights,” June 24, 2005, LexisNexis Academic; Edward Hudgins, “Your Castle No More,” Washington Times, June 27, 2005, LexisNexis Academic; Richard W. Rahn, “Uninformed Expropriation,” Washington Times, June 29, 2005, LexisNexis Academic; USA Today, “Homeowners Lose on Property Rights,” July 1, 2005, LexisNexis Academic; John Tierney, “Your Land Is My Land,” New York Times, July 5, 2005, LexisNexis Academic; Donald Lambro, “Home, Seized, Home,” Washington Times, July 18, 2005, LexisNexis Academic. 27. Tom Knott, “Land Grabbers’ Aim at Skyland Is Way Off Target,” Washington Times, July 21, 2005, LexisNexis Academic. 28. Wolf Blitzer, Wolf Blitzer Reports, CNN, June 23, 2005, LexisNexis Academic.

29. Greg Pierce, “Souter’s Property,” Washington Times, June 29, 2005, LexisNexis Academic. 30. Alex Chadwick, Day to Day, National Public Radio, June 23, 2005, LexisNexis Academic. 31. Washington Times, “A Win for Big Government,” June 24, 2005, LexisNexis Academic; Guy Taylor, “Supreme Court Backs Eminent Domain,” Washington Times, June 24, 2005, LexisNexis Academic; Edward Hudgins, “Your Castle No More,” Washington Times, June 27, 2005, LexisNexis Academic. 32. Charles Hurt, “Congress Assails Domain Ruling,” Washington Times, July 1, 2005, LexisNexis Academic. 33. Tom Knott, “Eminent Domain Ruling Slaps Down the Owner,” Washington Times, June 30, 2005, LexisNexis Academic; Todd Gaziano and Paul Rosenzweig, “A Fitting Property Rights Memorial,” Washington Times, July 7, 2005, LexisNexis Academic; Neil Conan, Talk of the Nation, National Public Radio, June 23, 2005, LexisNexis Academic; Tucker Carlson, The Situation, MSNBC, June 23, 2005, LexisNexis Academic. 34. I generate this data using the same Lexis-Nexis keyword searches described earlier in the chapter.

Chapter 5 1. In contrast, coverage at the local level can vary much more widely in the frames and tonal content employed (Haider-Markel 2004; Haider-Markel, Allen, and Johansen 2006). 2. I focus on cases that receive front-page attention to explore the content of press coverage, not to investigate why the press covers some decisions and ignores others.Page 214 → 3. The original version of this study oversampled decisions released before 1987 and after 2005. To correct for this imbalance, I report results that include a probability sampling weight correction (reported findings are robust to the exclusion of the weights as well). 4. Washington Times coverage available from 1989 to the present. I examined whether source-specific effects altered the major findings present here and found that they do not; consequently, I pool analyses across all sources. 5. The results are substantively similar for either set of measures. Because Bailey estimates allow for interinstitution comparisons, I report them here. 6. Ho and Quinn (2010) caution that labels such as liberal and conservative are arbitrarily assigned when describing Martin-Quinn scores. But because the press commonly assigns these designations to two distinct wings of the Court, I follow this practice here. 7. The justification for using the majority median to measure the substance of Supreme Court decisions is supplied by Carrubba et al. 2012. 8. Using the SCDB’s categorization of the major substantive issue on which the Court rules in a given case, I also create a measure of controversy for issues that arouse deeply held personal convictions in most individuals (Skitka, Bauman, and Sargis 2002; Wisneski, Lytle, and Skitka 2009). The issues coded as controversial were abortion, affirmative action, desegregation, First Amendment/obscenity, and sex discrimination. Noncontroversial issues, which frequently involve legal provisions, include libel, national supremacy, securities, and state regulation. Estimated effects are substantively similar for both measures of issue controversy, and I utilize the BH controversy measure in all reported models. 9. I generate this data by analyzing the number of amici mentioned in published case materials from LexisNexis. 10. Linda Ponce, “Justices’ Decisions in High Profile Suits Defy Easy Labeling,” Washington Times, July 5, 1989, LexisNexis Academic; Walter Dellinger, “The Abortion Decision: Momentum to Confusion; The Court Steers Us toward Social DisasterВ .В .В . and Threatens Other Personal Liberties as Well,” Washington Post, July 9, 1989, LexisNexis Academic; Dan Balz and David S. Broder, “Governors Assess Impact of Abortion,” Washington Post, August 2, 1989, LexisNexis Academic. 11. Jerry Knight and Joan Biskupic, “High Court Ruling May Add Billions to S&L Cleanup Cost,” Washington Post, July 2, 1996, LexisNexis Academic; Washington Post, “Costly but Correct,” July 5, 1996, LexisNexis Academic. 12. John Files, “Advocates Hope Supreme Court Ruling Can Renew Attention to вЂDon’t Ask,

Don’t Tell,’” New York Times, March 12, 2005, LexisNexis Academic; New York Times, “Prosecutorial Racial Bias in Texas,” June 15, 2005, LexisNexis Academic; Washington Post, “Strike against Bias,” June 14, 2005, LexisNexis Academic. 13. Nat Hentoff, “Scalia Outdoes the ACLU,” Washington Post, June 30, 1992, LexisNexis Academic; Bruce Fein, “Benchmarks of Absurdity,” Washington Times, March 12, 1992, LexisNexis Academic. 14. Ruth Marcus, “Supreme Court Overturns Law Barring Hate Crimes,” Washington Post, June 23, 1992, LexisNexis Academic; Frank J. Murray, “High Court Backs Asset Confiscation,” Washington Times, March 5, 1996, LexisNexis Academic; Joan Page 215 →Biskupic, “Court Upholds Criminal Forfeiture Law,” Washington Post, March 5, 1996, LexisNexis Academic. 15. Washington Post, “Handguns Supreme,” June 27, 2008, LexisNexis Academic; Washington Times, “Gun Ban Ends,” June 27, 2008, LexisNexis Academic; Paul Duggan, “Having Toppled D.C. Ban, Man Registers Revolver,” Washington Post, July 19, 2008, LexisNexis Academic. 16. Linda Greenhouse, “Justices, Ruling 5–4, Endorse Personal Right to Own Gun,” New York Times, June 27, 2008, LexisNexis Academic. 17. Activism charges, by definition, describe judges “legislating from the bench” (e.g., striking down laws, ignoring precedent); the formal concept has no relation to coalition size. 18. The effect is statistically significant for Washington Times coverage and for New York Times and Washington Post coverage. 19. Following existing applications of Wcopyfind in political science, I employed the program’s default parameters to identify the percentage of matching language (Corley 2008; Corley, Collins, and Calvin 2011; Collins, Corley, and Hammer 2013; Eshbaugh-Soha 2013). 20. For unanimously decided cases, the figure jumps to 7 percent, while it stands at 4 percent for five–four rulings. 21. Linda Greenhouse, “Justices, Ruling 5–4, Endorse Personal Right to Own Gun,” New York Times, June 27, 2008, LexisNexis Academic; Robert Barnes, “Justices Reject D.C. Ban on Handgun Ownership,” Washington Post, June 27, 2008, LexisNexis Academic. 22. Dan Balz and Keith B. Richburg, “Historic Decision Renews Old Debate,” Washington Post, June 27, 2008, LexisNexis Academic.

Chapter 6 1. Dan Abrams and Martin Savidge, The Abrams Report, MSNBC, June 23, 2005, LexisNexis Academic; Tucker Carlson, The Situation, MSNBC, June 23, 2005, LexisNexis Academic; Daryn Kagan, Jamie McIntyre, and Kimberly Osias, Live Today, CNN, June 23, 2005, LexisNexis Academic; Tony Harris, Ken Dolan, and Daria Dolan, Dolans Unscripted, CNN, June 25, 2005, LexisNexis Academic. 2. Wolf Blitzer, Jamie McIntyre, Jennifer Eccleston, Karl Penhaul, Christine Romans, and Brian Todd, Wolf Blitzer Reports, CNN, June 23, 2005, LexisNexis Academic. 3. Tucker Carlson, The Situation, MSNBC, June 23, 2005, LexisNexis Academic. 4. Kyra Phillips, Sean Callebs, Larry Smith, Kelli Arena, Jessica Yellin, Chad Myers, Miles O’Brien, Don Lemon, and Kathleen Koch, CNN Newsroom, CNN, December 10, 2007, LexisNexis Academic. 5. Brit Hume, Special Report with Brit Hume, Fox News, June 26, 1998, LexisNexis Academic; Lou Waters and Charles Bierbauer, CNN Today, CNN, June 19, 1998, LexisNexis Academic. 6. Joie Chen and Charles Bierbauer, The World Today, CNN, June 19, 2000, LexisNexisPage 216 → Academic; Lou Waters and Charles Bierbauer, CNN Today, CNN, June 19, 2000, LexisNexis Academic. 7. Sean Hannity and Alan Colmes, Hannity and Colmes, Fox News, May 23, 2000, LexisNexis Academic. 8. Greta Van Susteren, Roger Cossack, and Charles Bierbauer, Burden of Proof, CNN, May 18, 2000, LexisNexis Academic. 9. Judy Woodruff, Kate Snow, Kelli Arena, Candy Crowley, Ron Brownstein, Robert Novak, Eileen O’Connor, Rea Blakely, Major Garrett, William Schneider, and Charles Bierbauer, Inside Politics, CNN, June 28, 2001, LexisNexis Academic. 10. Sean Hannity and Alan Colmes, Hannity and Colmes, Fox News, May 23, 2000, LexisNexis Academic.

11. Recall again the erroneous reporting of CNN and Fox News about the Obamacare ruling. 12. Tucker Carlson, The Situation, MSNBC, June 23, 2005, LexisNexis Academic.

Chapter 7 1. Cues may also serve a persuasive function (Mondak 1990). 2. Nearly four out of five Americans trust the Court to make decisions that are “right for the country as a whole” (Gibson, Caldeira, and Spence 2003b, 540).

Chapter 8 1. However, the Court may well have succeeded in uniting Americans against a broad interpretation of the Public Use Clause. 2. The items used to construct the decision approval scale were: (1) Do you favor or oppose allowing public schools and universities to fund religious publications in equal measure as secular (nonreligious) publications? (2) In your personal opinion, do you believe that it is more important to ensure the separation of church and state by prohibiting government funding for all types of religious expression or to protect the freedom of speech by allowing the government to fund religious expression? (3) The Supreme Court recently ruled that the Constitution permits government funding of religious expression. In your personal opinion, do you agree with this Supreme Court decision? (4) Regardless of your personal opinion, do you agree that the Constitution permits government funding of religious expression? The scale reliability coefficient is 0.87 and the eigenvalue of its first factor is 2.65 (the eigenvalue of its second factor is в€’0.02). The factor loadings for variables 1–4 are 0.86, 0.74, 0.89, and 0.77, respectively. 3. In the gun rights condition, the Court was described as striking down the Gun Free School Zones Act (based on the decision in U.S. v. Lopez). In the religious expression case, the Court was portrayed as banning the display of the Ten Commandments in public places. The eminent domain case was based on the ruling in Kelo v. City of New London. Finally, the Court was described as striking down term limits for members of Congress based on its decision in U.S. Term Limits v. Thornton.Page 217 → 4. In contrast to Study 1, positive and negative frames raise distinct considerations. 5. For the gun rights issue, the items used to construct the decision approval index were: (1) In your personal opinion, should the federal government have the power to ban guns near schools, or should this be a matter left to state and local governments? (2) In your personal opinion, what is more important, ensuring safety from violence with strict gun control policies, or ensuring that the government does not infringe on the right to bear arms? (3) Do you support or oppose the law known as the Gun Free School Zones Act, which gives the national government the power to ban guns near schools? (4) Regardless of your personal opinion, what do you believe the Constitution permits? 6. For the various issues, scale properties were as follows. Gun rights—reliability coefficient 0.81; factor 1 eigenvalue 2.23 (factor 2 eigenvalue 0.05); factor loadings 0.79, 0.74, 0.89, 0.51. Takings—reliability coefficient 0.67; factor 1 eigenvalue 1.46 (factor 2 eigenvalue в€’0.01); factor loadings 0.65, 0.74, 0.57, 0.42. Term limits—reliability coefficient 0.69; factor 1 eigenvalue 1.49 (factor 2 eigenvalue 0.01); factor loadings 0.74, 0.75, 0.37, 0.50. Religious expression—reliability coefficient 0.71; factor 1 eigenvalue 1.71 (factor 2 eigenvalue 0.04); factor loadings 0.80, 0.51, 0.81, 0.39. 7. Because the cell sizes in Study 2 are small (the average size of a cell in a treatment group is 29), there is not enough evidence to suggest that diffuse support has a significant influence on decision approval across framing conditions (as we saw in Study 1).

Chapter 9 1. To verify the importance of the tax frame postdecision, I conducted a LexisNexis keyword search for mentions of the health care law alongside tax in six-month intervals in the New York Times and Washington Post. The number of published articles were as follows: 71 in the first half and 110 in the second half of 2009, 110 in the first half and 93 in the second half of 2010, 96 in the first half and 194 in the second half of 2011, and 274 in the first half (prior to the ruling) and 530 in the second half (after the ruling) of 2012. The

postruling tax coverage represented a 292 percent increase over the long-term average and was double the previous six-month period. 2. Once again, specifics matter. If one were to measure institutional support with a question about whether the Court should be abolished, this would likely tap into very strong, durable preferences that Americans hold about the institution. But other measures of support—including some loyalty indicators proposed by Gibson, Caldeira, and Spence (2005)—may prove less durable. 3. I take advantage of the widely anticipated release of the health care ruling by measuring opinion in June and July 2012 (at similar pre- and postdecision intervals). The goal is to learn about the effect of the decision in the respondent population, meaning that the appropriate comparison is pre- and postruling attitudes for all respondents. 4. Respondents answered 33 questions in the preruling phase on a range of political issues. 5. The concern is made more apparent by literature on demand effects (Crowne and Marlowe 1964; Fleming et al. 2007; Fleming and Zizzo 2011).Page 218 → 6. Out of an abundance of caution, I refrain from making strong statements about what results suggest regarding institutional “loyalty.” I instead employ the term institutional support throughout the chapter. 7. Among those who strongly or somewhat approved of the ACA before the ruling, Obamacare approval change is 0 on average. Among those who strongly or somewhat disapproved of the law, Obamacare approval change is .06 (p = 0.09). 8. We can explore evidence for positivity bias further by using the measure of news exposure. According to Woodson, Gibson, and Lodge (2011), exposure to legitimating symbols in the media increases diffuse support for the Supreme Court. But there is no evidence that news exposure influences changes in diffuse support; indeed, the measures are very weakly correlated (r = в€’0.008). 9. As such, it is essential to rule out all other accounts that could cast doubt on this finding. With the two time points of data, we have clear evidence of causality. And we have found no evidence for another explanation—that respondents’ views about the constitutionality of the ACA impacted their support for the Court. A final alternative explanation might argue that question-order effects contribute to changes in diffuse support, since respondents were first queried about their approval for Obamacare in the postdecision survey. This explanation suggests that diffuse support for the Court is a type of nonseparable preference, the only type for which question-order effects take hold (Lacy 2001). By definition, the explanation implies that diffuse support depends on the outcomes of other issues (such as policy matters before the Court). But in making such an argument, one would undermine a basic tenet of what we know about institutional legitimacy. Existing research suggests that diffuse support is far from a nonseparable preference, at least in the short term (e.g., Gibson, Caldeira, and Spence 2003b, 553). It is a troubling proposition for the Court, which depends on a reservoir of diffuse support to function effectively, if trust in it responds to its policy output (as appears to be the case here) or to something even more benign like question-order effects. 10. I measure strong attitudes here with self-reported preferences on the ACA—“strongly” approve or disapprove. This operationalization is closest to what Krosnick et al. define as the extremity and intensity of attitudes (1993). 11. The counterexample is the decision in Bush v. Gore, where support for the Court influenced support for the outcome (Gibson, Caldeira, and Spence 2003). One possible explanation for this pattern is that the strength of election-related attitudes diminished markedly over time. By early 2001, when the survey on which Gibson, Caldeira, and Spence rely was administered, respondents may have come to accept the resolution of the election more fully. In so doing, the strength of their election-related preferences would have declined, leading to the Court’s apparent persuasive power. If this were the case, we would expect a stronger influence of election-related preferences in the days that immediately followed the Court’s ruling. And indeed, one study, drawing on a tracking poll in the field at the time of the election, finds shifting public views about the Court; however, this study does not include a measure of diffuse support (Kritzer 2001).

Page 219 →Chapter 10

1. Soledad O’Brien, American Morning, CNN, June 24, 2005, LexisNexis Academic. 2. In this respect, we might see parallels to foreign policy conflicts where political actors become reluctant to criticize the president.

Appendixes 1. Follows the definition presented by Mutz and Reeves 2005.

Page 220 → Page 221 →

References Adarand Constructors, Inc. v. Pena. 1995. 515 U.S. 200. Allen, Mahalley D., and Donald P. Haider-Markel. 2006. “Connecting Supreme Court Decisions, Media Coverage, and Public Opinion: The Case of Lawrence v. Texas.” American Review of Politics 27 (Fall): 209–30. Althaus, Scott L., Jill A. Edy, Robert M. Entman, and Patricia Phalen. 1996. “Revising the Indexing Hypothesis: Officials, Media, and the Libya Crisis.” Political Communication 13 (4): 407–21. Arceneaux, Kevin. 2012. “Cognitive Biases and the Strength of Political Arguments.” American Journal of Political Science 56 (2): 271–85. Arizona et al. v. United States. 2012. 567 U.S. ___. Aronson, Elliot, Phoebe C. Ellsworth, J. Merrill Carlsmith, and Marty Hope Gonzales. 1990. Methods of Research in Social Psychology. 2nd ed. New York: McGraw-Hill. Baas, Larry R., and Dan Thomas. 1984. “The Supreme Court and Policy Agreement: Experimental Tests.” American Politics Research 12 (3): 335–60. Bailey, Michael A. 2007. “Comparable Preference Estimates across Time and Institutions for the Court, Congress, and Presidency.” American Journal of Political Science 51 (3): 433–48. Bailey, Michael A., and Forrest Maltzman. 2008. “Does Legal Doctrine Matter? Unpacking Law and Policy Preferences on the U.S. Supreme Court.” American Political Science Review 102 (3): 369–84. Bailey, Michael A., and Forrest Maltzman. 2010. The Constrained Court. Princeton: Princeton University Press. Baird, Vanessa A., and Amy Gangl. 2006. “Shattering the Myth of Legality: The Impact of the Media’s Framing of Supreme Court Procedures on Perceptions of Fairness.” Political Psychology 27 (4): 597–614. Baron, Jane B. 2007. “Winding toward the Heart of the Takings Muddle: Kelo, Lingle, and Public Discourse about Private Property.” Fordham Urban Law Journal 34 (2): 613–55. Baron, Reuben M., and David A. Kenny. 1986. “The Mediator-Moderator Distinction in Social Psychological Research: Conceptual, Strategic, and Statistical Considerations.” Journal of Personality and Social Psychology 51 (6): 1173–82. Page 222 →Barros, D. Benjamin. 2006. “At Last, Some Clarity: The Potential Long-Term Impact of Lingle v. Chevron and the Separation of Takings and Substantive Due Process.” Albany Law Review 69 (1): 343–56. Bartels, Brandon L. 2011. “Choices in Context: How Case-Level Factors Influence the Magnitude of Ideological Voting on the U.S. Supreme Court.” American Politics Research 39 (1): 142–75. Bartels, Brandon L., and Christopher D. Johnston. 2013. “On the Ideological Foundations of Supreme Court Legitimacy in the American Public.” American Journal of Political Science 57 (1): 184–99. Bartels, Brandon L., and Diana C. Mutz. 2009. “Explaining the Process of Institutional Opinion Leadership.” Journal of Politics 71 (1): 249–61. Baum, Lawrence. 2006. Judges and Their Audiences: A Perspective on Judicial Behavior. Princeton: Princeton University Press.

Baum, Matthew, and Tim Groeling. 2005. “What Gets Covered? How Media Coverage of Elite Debate Drives the Rally-’Round-the-Flag Phenomenon: 1979–1998.” In In the Public Domain: Presidents and the Challenges of Public Leadership, edited by Lori Cox Han and Diane J. Heith, 71–106. Albany: State University of New York Press. Bennett, W. Lance. 1990. “Toward a Theory of Press-State Relations.” Journal of Communication 40 (2): 103–27. Bennett, W. Lance. 2007. News: The Politics of Illusion. 7th ed. New York: Pearson. Bennett, W. Lance, and Shanto Iyengar. 2008. “A New Era of Minimal Effects? The Changing Foundations of Political Communication.” Journal of Communication 58 (4): 707–31. Bennett, W. Lance, Regina G. Lawrence, and Steven Livingston. 2006. “None Dare Call It Torture: Indexing and the Limits of Press Independence in the Abu Ghraib Scandal.” Journal of Communication 56 (3): 467–85. Bennett, W. Lance, and Steven Livingston. 2003. “A Semi-Independent Press: Government Control and Journalistic Autonomy in the Political Construction of News.” Political Communication 20 (4): 359–62. Bennett, W. Lance, and Jarol B. Manheim. 1996. “Taking the Public by Storm: Information, Cuing, and the Democratic Process in the Gulf Conflict.” Political Communication 10 (4): 331–51. Benoit, William, and David Arne. 2009. “Non-Presidential Political Advertising in Campaign 2004.” Human Communication 12 (1): 91–117. Benoit, William, and Jayne Henson. 2009. “President Bush’s Image Repair Discourse on Hurricane Katrina.” Public Relations Review 35 (1): 40–46. Berinsky, Adam J., Gregory A. Huber, and Gabriel S. Lenz. 2011. “Evaluating Online Labor Markets for Experimental Research: Amazon.com’s Mechanical Turk.” Political Analysis 20 (3): 351–68. Berkson, Larry. 1978. The Supreme Court and Its Publics. Lexington, MA: Lexington Books. Berman v. Parker. 1954. 348 U.S. 26. Blake, William D., and Hans J. Hacker. 2010. “вЂThe Brooding Spirit of the Law’: Supreme Court Justices Reading Dissents from the Bench.” Justice Systems Journal 31 (1): 1–25. Page 223 →Blumenthal, Mark. 2012. “Obamacare Polls Find Supreme Court Ruling Not Uniting Americans.” Accessed January 31, 2013. http://www.huffingtonpost.com/2012/07/03/obamacare-pollssupreme-court-health-caredecision_n_1646725.html?utm_hp_ref=@pollster. Brader, Ted. 2005. “Striking a Responsive Chord: How Political Ads Motivate and Persuade Voters by Appealing to Emotions.” American Journal of Political Science 49 (2): 388–405. Brady, Henry E., and Paul M. Sniderman. 1985. “Attitude Attribution: A Group Basis for Political Reasoning.” American Political Science Review 79 (4): 1061–78. Breau, David L. 2006. “A New Take on Public Use: Were Kelo and Lingle Nonjusticiable?” Duke Law Journal 55 (4): 835–63. Brickman, Danette, and Belinda Bragg. 2007. “Does the Media Matter? Understanding the Influence of Media Coverage on Public Response to Supreme Court Decisions.” Paper presented at the Western Political Science Association Conference, Las Vegas.

Brickman, Danette, and David A. M. Peterson. 2006. “Public Opinion Reaction to Repeated Events: Citizen Response to Multiple Supreme Court Abortion Decisions.” Political Behavior 28 (1): 87–112. Brigham, John. 1987. The Cult of the Court. Philadelphia: Temple University Press. Brooks, Deborah Jordan, and John G. Geer. 2007. “Beyond Negativity: The Effects of Incivility on the Electorate.” American Journal of Political Science 51 (1): 1–16. Brown v. Board of Education. 1954. 347 U.S. 483. Buhrmester, Michael, Tracy Kwang, and Samuel D. Gosling. 2011. “Amazon’s Mechanical Turk: A New Source of Inexpensive, Yet High-Quality, Data?” Perspectives on Psychological Science 6 (1): 3–5. Caldeira, Gregory A. 1986. “Neither the Purse nor the Sword: Dynamics of Public Confidence in the Supreme Court.” American Political Science Review 80 (4): 1209–26. Caldeira, Gregory A. 1987. “Public Opinion and the U.S. Supreme Court: FDR’s Court-Packing Plan.” American Political Science Review 81 (4): 1139–53. Caldeira, Gregory A., and James L. Gibson. 1992. “The Etiology of Public Support for the Supreme Court.” American Journal of Political Science 36 (3): 635–64. Caldeira, Gregory A., and John R. Wright. 1988. “Organized Interests and Agenda Setting in the U.S. Supreme Court.” American Political Science Review 82 (4): 1109–27. Campbell, Andrea, and Nathaniel Persily. 2012. “The Health Care Case in the Public Mind: How the Supreme Court Shapes Opinion about Itself and the Laws It Considers.” Working Paper. Accessed April 15, 2013. blog.oup.com. Campbell, Angus, Phillip E. Converse, Warren E. Miller, and Donald E. Stokes. 1960. The American Voter. New York: Wiley. Campbell, Donald T., and Julian C. Stanley. 1966. Experimental and Quasi-Experimental Designs for Research. Chicago: Rand McNally. Campbell, Karlyn Kohrs, and Kathleen Hall Jamieson. 2008. Presidents Creating the Presidency: Deeds Done in Words. Chicago: University of Chicago Press. Carpenter, Dale. 2012. Flagrant Conduct: The Story of Lawrence v. Texas. New York: Norton. Carrubba, Cliff, Barry Friedman, Andrew D. Martin, and Georg Vanberg. 2012. Page 224 →“Who Controls the Content of Supreme Court Opinions?” American Journal of Political Science 56 (2): 400–412. Casey, Gregory. 1974. “The Supreme Court and Myth: An Empirical Investigation.” Law and Society Review 8 (3): 385–420. Chaiken, Shelly. 1980. “Heuristic versus Systematic Information Processing and the Use of Source versus Message Cues in Persuasion.” Journal of Personality and Social Psychology 39 (5): 752–66. Chaiken, Shelly. 1987. “The Heuristic Model of Persuasion.” In Social Influence: The Online Symposium, vol. 5, edited by Mark P. Zanna, James M. Olson, and C. Peter Herman, 3–39. New York: Psychology Press. Chong, Dennis, and James N. Druckman. 2007a. “Framing Theory.” Annual Review of Political Science 10:103–26. Chong, Dennis, and James N. Druckman. 2007b. “A Theory of Framing and Opinion Formation in

Competitive Elite Environments.” Journal of Communication 57 (1): 99–118. Chong, Dennis, and James N. Druckman. 2012. “Counterframing Effects.” Journal of Politics 75 (1): 1–16. Claiborne v. U.S. 2007. 465 U.S. 1305. Clark, Tom S. 2009. “The Separation of Powers, Court-Curbing, and Judicial Legitimacy.” American Journal of Political Science 53 (4): 971–89. Clark, Tom S., and Clifford J. Carrubba. 2012. “A Theory of Opinion Writing in a Political Hierarchy.” Journal of Politics 74 (2): 584–603. Clawson, Rosalee, Elizabeth R. Kegler, and Eric N. Waltenburg. 2001. “The Legitimacy-Conferring Authority of the U.S. Supreme Court.” American Politics Research 29 (6): 566–91. Clawson, Rosalee, and Eric N. Waltenburg. 2003. “Support for a Supreme Court Affirmative Action Decision: A Story in Black and White.” American Politics Research 31 (3): 251–79. Collins, Paul M., Jr. 2004. “Friends of the Court: Examining the Influence of Amicus Curiae Participation in U.S. Supreme Court Litigation.” Law and Society Review 38 (4): 807–32. Collins, Paul M., Jr., Pamela C. Corley, and Jesse Hammer. 2013. “Me Too? An Investigation of Repetition in U.S. Supreme Court Amicus Curiae Briefs.” Paper presented at the Midwest Political Science Association National Conference, Chicago. Converse, Phillip E. 1964. “The Nature of Belief Systems in Mass Publics.” In Ideology and Discontent, edited by David E. Apter, 206–61. New York: Free Press. Corley, Pamela C. 2008. “The Supreme Court and Opinion Content: The Influence of Parties’ Briefs.” Political Research Quarterly 61 (3): 468–78. Corley, Pamela C., Paul M. Collins Jr., and Bryan Calvin. 2011. “Lower Court Influence on U.S. Supreme Court Opinion Content.” Journal of Politics 73 (1): 31–44. Crawford, Jarret T., and Jane M. Pilanski. 2012. “Political Intolerance, Right and Left.” Political Psychology 35 (6): 841–51. Crowne, Douglas P., and David Marlowe. 1964. The Approval Motive: Studies in Evaluative Dependence. New York: Wiley. Dana, David A. 2006. “The Law and Expressive Meaning of Condemning the Poor after Kelo.” Northwestern University Law Colloquy 101 (2): 5–22. Page 225 →Davis, Richard. 1987. “Lifting the Shroud: News Media Coverage of the U.S. Supreme Court.” Communications and the Law 9 (5): 43–59. Davis, Richard. 1994. Decisions and Images: The Supreme Court and the Press. Englewood Cliffs, NJ: Prentice Hall. Davis, Richard. 2011. Justices and Journalists: The U.S. Supreme Court and the Media. New York: Cambridge University Press. Doherty, David. 2013. “Who Do People Think Representatives Should Respond To: Their District or the Country?” Public Opinion Quarterly 77 (1): 237–55.

Dowling, Conor M., and Amber Wichowsky. 2013. “Does It Matter Who’s behind the Curtain? Anonymity in Political Advertising and the Effects of Campaign Finance Disclosure.” American Politics Research 41 (6): 956–96. Druckman, James N. 2001. “On the Limits of Framing Effects: Who Can Frame?” Journal of Politics 63 (4): 1041–66. Druckman, James N. 2004. “Political Preference Formation: Competition, Deliberation, and the (Ir)relevance of Framing Effects.” American Political Science Review 98 (4): 671–86. Druckman, James N., Cari Lynn Hennessy, Kristi St. Charles, and Johnathan Webber. 2010. “Competing Rhetoric over Time: Frames versus Cues.” Journal of Politics 72 (1): 136–48. Eagly, Alice H., and Shelly Chaiken. 1993. The Psychology of Attitudes. Fort Worth, TX: Harcourt, Brace, Jovanovich. Eagly, Alice H., Serena Chen, Shelly Chaiken, and Kelly Shaw-Barnes. 1999. “The Impact of Attitudes on Memory: An Affair to Remember.” Psychological Bulletin 125 (1): 64–89. Easton, David. 1965. A Systems Analysis of Political Life. Chicago: University of Chicago Press. Egan, Patrick J., and Jack Citrin. 2009. “Opinion Leadership, Backlash, and Delegitimation: Supreme Court Rulings and Public Opinion.” Accessed November 16, 2012. http://ssrn.com/abstract=1443631. Egan, Patrick J., Nathaniel Persily, and Kevin Wallsten. 2008. “Gay Rights.” In Public Opinion and Constitutional Controversy, edited by Nathaniel Persily, Jack Citrin, and Patrick J. Egan, 234–66. New York: Oxford University Press. Ely, James W., Jr. 2004. “вЂPoor Relation’ Once More: The Supreme Court and the Vanishing Rights of Property Owners.” In Cato Supreme Court Review: 2004–2005, edited by Mark K. Moller, 39–69. Washington, DC: Cato Institute. Entman, Robert M. 2004. Projections of Power: Framing News, Public Opinion, and U.S. Foreign Policy. Chicago: University of Chicago Press. Epstein, Lee. 1991. “Courts and Interest Groups.” In The American Courts: A Critical Assessment, edited by John B. Gates and Charles A. Johnson, 335–71. Washington, DC: Congressional Quarterly Press. Epstein, Lee, and Jack Knight. 2013. “Reconsidering Judicial Preferences.” Annual Review of Political Science 16:11–31. Epstein, Lee, and Jeffrey A. Segal. 2000. “Measuring Issue Salience.” American Journal of Political Science 44 (1): 66–83. Epstein, Lee, Jeffrey A. Segal, and Harold J. Spaeth. 2001. “The Norm of Consensus on the U.S. Supreme Court.” American Journal of Political Science 45 (2): 362–77. Page 226 →Epstein, Lee, and Thomas G. Walker. 2010. Constitutional Law for a Changing America: Rights, Liberties, and Justice. Washington, DC: CQ Press. Eshbaugh-Soha, Matthew. 2013. “Presidential Influence of the News Media: The Case of the Press Conference.” Political Communication 30 (4): 548–64. Eskridge, William N., Jr. 1991. “Overriding Supreme Court Statutory Interpretation Decisions.” Yale Law Journal 101 (2): 331–455.

Falk, Erika, Erin Grizard, and Gordon McDonald. 2006. “Legislative Issue Advertising in the 108th Congress: Pluralism or Peril?” Harvard International Journal of Press/Politics 11 (4): 148–64. Feather, N. T. 1964. “Acceptance and Rejection of Arguments in Relation to Attitude Strength, Critical Ability, and Intolerance of Inconsistency.” Journal of Abnormal and Social Psychology 69 (2): 127–36. Feldman, Stanley. 1995. “Answering Survey Questions: The Measurement and Meaning of Public Opinion.” In Political Judgment: Structure and Process, edited by Milton Lodge and Kathleen M. McGraw, 249–70. Ann Arbor: University of Michigan Press. Fiske, Susan T., and Shelley E. Taylor. 1984. Social Cognition. New York: Random House. Fleming, Piers, Ellen Townsend, Kenneth C. Lower, and Eamonn Ferguson. 2007. “Social Desirability Influences on Judgements of Biotechnology across the Dimensions of Risk, Ethicality, and Naturalness.” Journal of Risk Research 1 (7): 989–1003. Fleming, Piers, and Daniel John Zizzo. 2011. “Social Desirability, Approval, and Public Good Contribution.” Personality and Individual Differences 51 (3): 258–62. Fletcher, William A. 2006. “Kelo, Lingle, and San Remo Hotel: Takings Law Now Belongs to the States.” Santa Clara Law Review 46 (4): 767–79. Forgette, Richard, and Jonathan S. Morris. 2006. “High-Conflict Television News and Public Opinion.” Political Research Quarterly 59 (3): 447–56. Franklin, Charles H., and Liane C. Kosaki. 1989. “Republican Schoolmaster: The U.S. Supreme Court, Public Opinion, and Abortion.” American Political Science Review 83 (3): 751–71. Franklin, Charles H., and Liane C. Kosaki. 1995. “Media, Knowledge, and Public Evaluations of the Supreme Court.” In Contemplating Courts, edited by Lee Epstein, 353–75. Washington, DC: CQ Press. Franklin, Charles H., Liane C. Kosaki, and Herbert M. Kritzer. 1993. “The Salience of U.S. Supreme Court Decisions.” Paper presented at the annual meeting of the American Political Science Association, Washington, DC. Freelon, Dean G. 2010. “ReCal: Intercoder Reliability Calculation as a Web Service.” International Journal of Internet Science 5 (1): 20–33. Gans, Herbert J. 1979. Deciding What’s News. New York: Pantheon. Gash, Alison, and Angelo Gonzales. 2008. “School Prayer.” In Public Opinion and Constitutional Controversy, edited by Nathaniel Persily, Jack Citrin, and Patrick J. Egan, 62–79. New York: Oxford University Press. Gates, Melissa, and Jan P. Vermeer. 1992. “Reporting Supreme Court Decisions: Conflict, Dissents, and Other Cues.” Paper presented at the annual meeting of the Western Political Science Association, San Francisco. Page 227 →Gerring, John. 2007. Case Study Research: Principles and Practices. Cambridge: Cambridge University Press. Gibson, James L., and Gregory A. Caldeira. 2009a. Citizens, Courts, and Confirmations: Positivity Theory and the Judgments of the American People. Princeton: Princeton University Press. Gibson, James L., and Gregory A. Caldeira. 2009b. “Confirmation Politics and the Legitimacy of the U.S. Supreme Court: Institutional Loyalty, Positivity Bias, and the Alito Nomination.” American Journal of Political Science 53 (1): 139–55.

Gibson, James L., and Gregory A. Caldeira. 2011. “Has Legal Realism Damaged the Legitimacy of the U.S. Supreme Court?” Law and Society Review 45 (1): 195–219. Gibson, James L., Gregory A. Caldeira, and Vanessa Baird. 1998. “On the Legitimacy of National High Courts.” American Political Science Review 92 (2): 343–58. Gibson, James L., Gregory A. Caldeira, and Lester Kenyatta Spence. 2003a. “Measuring Attitudes Toward the United States Supreme Court.” American Journal of Political Science 47 (2): 354–67. Gibson, James L., Gregory A. Caldeira, and Lester Kenyatta Spence. 2003b. “The Supreme Court and the U.S. Presidential Election of 2000: Wounds, Self-Inflicted or Otherwise?” British Journal of Political Science 33 (4): 535–56. Gibson, James L., Gregory A. Caldeira, and Lester Kenyatta Spence. 2005. “Why Do People Accept Public Policies They Oppose? Testing Legitimacy Theory with a Survey-Based Experiment.” Political Research Quarterly 58 (2): 187–201. Gibson, James L., Jeffrey A. Gottfried, Michael X. Delli Carpini, and Kathleen Hall Jamieson. 2010. “The Effects of Judicial Campaign Activity on the Legitimacy of Courts: A Survey-Based Experiment.” Political Research Quarterly 64 (3): 545–58. Ginsburg, Ruth Bader. 1995. “Communicating and Commenting on the Court’s Work.” Georgetown Law Journal 83 (6): 2119–30. Goldstein, Tom. 2012. “We’re Getting Wildly Differing Assessments.” Accessed February 8, 2013. http://www.scotusblog.com/2012/07/were-getting-wildly-differing-assessments/. Goux, Darshan, Patrick J. Egan, and Jack Citrin. 2008. “The War on Terror and Civil Liberties.” In Public Opinion and Constitutional Controversy, edited by Nathaniel Persily, Jack Citrin, and Patrick J. Egan, 310–32. New York: Oxford University Press. Graber, Doris A. 2001. Processing Politics: Learning from Television in the Internet Age. Chicago: University of Chicago Press. Graber, Doris A., and James M. Smith. 2005. “Political Communication Faces the 21st Century.” Journal of Communication 55 (3): 479–507. Green, Joshua A., and Matthew G. Jarvis. 2008. “The Right to Die.” In Public Opinion and Constitutional Controversy, edited by Nathaniel Persily, Jack Citrin, and Patrick J. Egan, 267–85. New York: Oxford University Press. Greenhouse, Linda. 1996. “Telling the Court’s Story: Justice and Journalism at the Supreme Court.” Yale Law Journal 105 (6): 1537–61. Griswold v. Connecticut. 1965. 381 U.S. 479. Groseclose, Tim, and Jeffrey Milyo. 2005. “A Measure of Media Bias.” Quarterly Journal of Economics 120 (4): 1191–1237. Grosskopf, Anke, and Jeffrey J. Mondak. 1998. “Do Attitudes toward Specific Supreme Page 228 →Court Decisions Matter? The Impact of Webster and Texas v. Johnson on Public Confidence in the Supreme Court.” Political Research Quarterly 51 (3): 633–54. Haider-Markel, Donald P. 2004. Media Coverage of Lawrence v. Texas: An Analysis of Content, Tone, and Frames in National and Local News Reporting. New York: GLAAD Center for the Study of Media and Society.

Haider-Markel, Donald P., Mahalley D. Allen, and Morgen Johansen. 2006. “Understanding Variations in Media Coverage of U.S. Supreme Court Decisions: Comparing Media Outlets in their Coverage of Lawrence v. Texas.” Harvard International Journal of Press/Politics 11 (2): 64–85. Hall, Matthew. 2009. “High Court Headlines: An Analysis of Case Characteristics Associated with Media Attention to Supreme Court Rulings.” Accessed January 14, 2011. http://ssrn.com/abstract=1428548. Hanley, John. 2008. “The Death Penalty.” In Public Opinion and Constitutional Controversy, edited by Nathaniel Persily, Jack Citrin, and Patrick J. Egan, 108–38. New York: Oxford University Press. Hanson, Peter. 2008. “Flag Burning.” In Public Opinion and Constitutional Controversy, edited by Nathaniel Persily, Jack Citrin, and Patrick J. Egan, 184–208. New York: Oxford University Press. Hart, Roderick. 2002. “Why Do They Talk That Way? A Research Agenda for the Presidency.” Presidential Studies Quarterly 32 (4): 693–709. Harvey, Anna, and Barry Friedman. 2006. “вЂPulling Punches’: Congressional Constraints on the Supreme Court’s Constitutional Rulings, 1987–2000.” Legislative Studies Quarterly 31 (4): 533–62. Hayes, Andrew F. 2009. “Beyond Baron and Kenny: Statistical Mediation Analysis in the New Millennium.” Communication Monographs 76 (4): 408–20. Hawaii Housing Authority v. Midkiff. 1984. 467 U.S. 229. Hazelwood School District et al. v. Kuhlmeier et al. 1988. 484 U.S. 260. Hess, Stephen. 1981. The Washington Reporters. Washington, DC: Brookings. Ho, Daniel E., and Kevin M. Quinn. 2008. “Measuring Explicit Political Positions of Media.” Quarterly Journal of Political Science 3 (4): 353–77. Ho, Daniel E., and Kevin M. Quinn. 2010. “How Not to Lie with Judicial Votes: Misconceptions, Measurement, and Models.” California Law Review 98 (3): 813–76. Hoekstra, Valerie J. 1995. “The Supreme Court and Opinion Change: An Experimental Study of the Court’s Ability to Change Opinion.” American Politics Quarterly 23 (1): 109–29. Hoekstra, Valerie J. 2000. “The Supreme Court and Local Public Opinion.” American Political Science Review 94 (1): 89–100. Hoekstra, Valerie J. 2003. Public Reaction to Supreme Court Decisions. Cambridge: Cambridge University Press. Horton, John J., David G. Rand, and Richard J. Zeckhauser. 2011. “The Online Laboratory: Conducting Experiments in a Real Labor Market.” Experimental Economics 14 (3): 399–425. Hoyle, Rick H., Monica J. Harris, and Charles M. Judd. 2002. Research Methods in Social Relations. Belmont, CA: Wadsworth. Huber, Gregory A., Seth J. Hill, and Gabriel S. Lenz. 2012. “Sources of Bias in Retrospective DecisionMaking: Experimental Evidence on Voters’ Limitations in Controlling Incumbents.” American Political Science Review 106 (4): 720–41. Page 229 →Huber, Gregory A., and Celia Paris. 2013. “Assessing the Programmatic Equivalence Assumption in Question Wording Experiments: Understanding Why Americans Like Assistance to the Poor More Than Welfare.” Public Opinion Quarterly 77 (1): 385–97.

Iyengar, Shanto. 1987. “Television News and Citizens’ Explanations of National Issues.” American Political Science Review 81 (3): 815–32. Iyengar, Shanto. 1991. Is Anyone Responsible? How Television Frames Political Issues. Chicago: University of Chicago Press. Iyengar, Shanto, and Kyu S. Hahn. 2009. “Red Media, Blue Media: Evidence of Ideological Selectivity in Media Use.” Journal of Communication 59 (1): 19–39. John, Sue Lockett, David Scott Domke, and Kevin Coe. 2007. “Going Public, Crisis after Crisis: The Bush Administration and the Press, from September 11 to Saddam.” Rhetoric and Public Affairs 10 (2): 195–220. Johnson, Richard M. 1967. “Compliance and Supreme Court Decision-Making.” Wisconsin Law Review 1967 (1): 170–85. Johnson, Timothy R., and Andrew D. Martin. 1998. “The Public’s Conditional Response to Supreme Court Decisions.” American Political Science Review 92 (2): 299–309. Joinson, Adam N. 1999. “Anonymity, Disinhibition, and Social Desirability on the Internet.” Behaviour Research Methods, Instruments, and Computers 31 (3): 433–38. Kanner, Gideon. 2006. “Kelo v. City of New London: Bad Law, Bad Policy, and Bad Judgment.” Urban Lawyer 38 (2): 201–35. Katsch, Ethan. 1983. “Supreme Court Beat: How Television Covers the U.S. Supreme Court.” Judicature 67 (1): 6–11. Kearney, Joseph D., and Thomas W. Merrill. 2000. “The Influence of Amicus Curiae Participation Briefs on the Supreme Court.” University of Pennsylvania Law Review 148 (3): 743–855. Kelo v. City of New London. 2005. 545 U.S. 469. Kent, Michael B., Jr. 2011. “More Questions Than Answers: Situating Judicial Takings with Existing Regulatory Takings Doctrine.” Virginia Environmental Law Journal 29 (2): 143–69. King, Gary, Robert O. Keohane, and Sidney Verba. 1994. Designing Social Inquiry: Scientific Inference in Qualitative Research. Princeton: Princeton University Press. Kmiec, Keenan D. 2004. “The Origin and Current Meanings of вЂJudicial Activism.’” California Law Review 92 (5): 1441–77. Knight, Jack, and Lee Epstein. 1996. “The Norm of Stare Decisis.” American Journal of Political Science 40 (4): 1018–35. Knoche, Eldon. 1987. “A Reporter’s View of Relations with Judges.” Judicature 70 (5): 268–69. Krantz, John H., and Resshad Dalal. 2000. “Validity of Web-Based Psychological Research.” In Psychological Experiments on the Internet, edited by Michael H. Birnbaum, 35–60. San Diego: Academic Press. Krippendorff, Kraus. 1987. “Association, Agreement, and Equity.” Quality and Quantity 21 (2): 109–23. Kritzer, Herbert M. 2001. “The Impact of Bush v. Gore on Public Perceptions and Knowledge of the Supreme Court.” Judicature 85 (1): 32–38.

Krosnick, Jon A., David S. Boninger, Yao C. Chuang, Matthew K. Berent, and CatherinaPage 230 → G. Carnot. 1993. “Attitude Strength: One Construct or Related Constructs?” Journal of Personality and Social Psychology 65 (6): 1132–51. Krosnick, Jon A., and Richard E. Petty. 2014. “Attitude Strength: An Overview.” In Attitude Strength: Antecedents and Consequences, edited by Jon A. Krosnick and Richard E. Petty, 1–24. New York: Psychology Press. Lacy, Dean. 2001. “A Theory of Nonseparable Preferences in Survey Responses.” American Journal of Political Science 45 (2): 239–58. Larson, Stephanie. 1985. “How the New York Times Covered Discrimination Cases.” Journalism and Mass Communication Quarterly 62 (4): 894–96. Le, Loan, and Jack Citrin. 2008. “Affirmative Action.” In Public Opinion and Constitutional Controversy, edited Nathaniel Persily, Jack Citrin, and Patrick J. Egan, 162–83. New York: Oxford University Press. Ledbetter v. Goodyear Tire & Rubber Company. 2007. 550 U.S. 618. Lee, Patricia H. 2006. “Eminent Domain: In the Aftermath of Kelo v. City of New London, a Resurrection in Norwood: One Public Interest Attorney’s View.” Western New England Law Review 29 (1): 121–40. Lerman, Amy E. 2008. “The Rights of the Accused.” In Public Opinion and Constitutional Controversy, edited by Nathaniel Persily, Jack Citrin, and Patrick J. Egan, 41–61. New York: Oxford University Press. Lijphart, Arend. 1975. “The Comparable-Case Strategy in Comparative Research.” Comparative Political Studies 8 (2): 158–77. Lingle v. Chevron U.S.A. Inc. 2005. 544 U.S. 528. Livingston, Steven, and W. Lance Bennett. 2003. “Gatekeeping, Indexing, and Live-Event News: Is Technology Altering the Construction of News?” Political Communication 20 (4): 363–80. Lombard, Matthew, Jennifer Snyder-Duch, and Cheryl Campanella Bracken. 2002. “Content Analysis in Mass Communication and Reporting of Intercoder Reliability.” Human Communication Research 28 (4): 587–604. Lombard, Matthew, Jennifer Snyder-Duch, and Cheryl Campanella Bracken. 2004. “Practical Resources for Assessing and Reporting Intercoder Reliability in Content Analysis Research Projects.” Accessed January 31, 2013. http://www.temple.edu/sct/mmc/reliability. Lucas v. South Carolina Coastal Council. 1992. 505 U.S. 1003. Luks, Samantha, and Michael Salamone. 2008. “Abortion.” In Public Opinion and Constitutional Controversy, edited by Nathaniel Persily, Jack Citrin, and Patrick J. Egan, 80–107. New York: Oxford University Press. Lupia, Arthur. 1994. “Shortcuts versus Encyclopedias: Information and Voting Behavior in California Insurance Reform Elections.” American Political Science Review 88 (1): 63–76. Lupia, Arthur, and Mathew D. McCubbins. 1998. The Democratic Dilemma: Can Citizens Learn What They Need to Know? Cambridge: Cambridge University Press. MacKinnon, David P., Amanda J. Fairchild, and Matthew S. Fritz. 2007. “Mediation Analysis.” Annual Review of Psychology 58:593–614.

Mahoney, James. 2000. “Strategies of Causal Inference in Small-N Analysis.” Sociological Methods and Research 28 (4): 387–424. Page 231 →Maltzman, Forrest, James F. Spriggs, and Paul Wahlbeck. 2000. Crafting Law on the Supreme Court: The Collegial Game. Cambridge: Cambridge University Press. Marshall, Thomas R. 1989. Public Opinion and the Supreme Court. Boston: Unwin Hyman. Martin, Andrew D., and Kevin M. Quinn. 2002. “Dynamic Ideal Point Estimation via Markov Chain Monte Carlo for the U.S. Supreme Court, 1953–1999.” Political Analysis 10 (2): 134–53. Mate, Manoj, and Matthew Wright. 2008. “The 2000 Presidential Election Controversy.” In Public Opinion and Constitutional Controversy, edited by Nathaniel Persily, Jack Citrin, and Patrick J. Egan, 333–52. New York: Oxford University Press. Mayeri, Serena, Ryan Brown, Nathaniel Persily, and Sun-Ho Kim. 2008. “Gender Equality.” In Public Opinion and Constitutional Controversy, edited by Nathaniel Persily, Jack Citrin, and Patrick J. Egan, 139–61. New York: Oxford University Press. McDermott, Rose. 2002. “Experimental Methods in Political Science.” Annual Review of Political Science 5:31–61. Merrill, Thomas W. 2010. “Why Lingle Is Half Right.” Vermont Journal of Environmental Law 11 (3): 421–35. Mishler, William, and Reginald S. Sheehan. 1993. “The Supreme Court as a Countermajoritarian Institution? The Impact of Public Opinion on Supreme Court Decisions.” American Political Science Review 87 (1): 87–101. Mondak, Jeffrey J. 1990. “Perceived Legitimacy of Supreme Court Decisions: Three Functions of Source Credibility.” Political Behavior 12 (4): 363–84. Mondak, Jeffrey J. 1992. “Institutional Legitimacy, Policy Agreement, and the Supreme Court.” American Politics Research 20 (4): 457–77. Mondak, Jeffrey J. 1994. “Policy Agreement and the Supreme Court: The Sources and Contexts of Legitimation.” Political Research Quarterly 47 (3): 675–92. Mondak, Jeffrey J., and Shannon Ishiyama Smithey. 1997. “The Dynamics of Public Support for the Supreme Court.” Journal of Politics 59 (4): 1114–42. Mullin, Megan. 2008. “Federalism.” In Public Opinion and Constitutional Controversy, edited by Nathaniel Persily, Jack Citrin, and Patrick J. Egan, 209–33. New York: Oxford University Press. Murakami, Michael. 2008. “Desegregation.” In Public Opinion and Constitutional Controversy, edited by Nathaniel Persily, Jack Citrin, and Patrick J. Egan, 18–40. New York: Oxford University Press. Murphy, Walter F., and Joseph Tanenhaus. 1968. “Public Opinion and the United States Supreme Court: Mapping of Some Prerequisites for Court Legitimation of Regime Change.” Law and Society Review 2 (3): 357–84. Mutz, Diana C. 2007. “Effects of вЂIn-Your-Face’ Television Discourse on Perceptions of a Legitimate Opposition.” American Political Science Review 101 (4): 621–35. Mutz, Diana C., and Byron Reeves. 2005. “The New Videomalaise: Effects of Televised Incivility on Political Trust.” American Political Science Review 99 (1): 1–15.

Nadler, Janice, Shari Seidman Diamond, and Matthew M. Patton. 2008. “Government Takings of Private Property.” In Public Opinion and Constitutional Controversy, edited by Nathaniel Persily, Jack Citrin, and Patrick J. Egan, 286–309. New York: Oxford University Press. Page 232 →National Federation of Independent Business v. Sebelius. 2012. 567 U.S. ____. Nelson, Thomas, Rosalee Clawson, and Zoe Oxley. 1997. “Media Framing of a Civil Liberties Conflict and Its Effect on Tolerance.” American Political Science Review 91 (3): 567–83. Newland, Chester A. 1964. “Press Coverage of the United States Supreme Court.” Western Political Quarterly 17 (1): 15–36. O’Callaghan, Jerome, and James O. Dukes. 1992. “Media Coverage of the Supreme Court’s Caseload.” Journalism Quarterly 69 (1): 195–203. Osgood, Charles E. 1960. “Cognitive Dynamics in the Conduct of Human Affairs.” Public Opinion Quarterly 24 (2): 341–65. Osgood, Charles E., and Percy H. Tannenbaum. 1955. “The Principle of Congruity in the Prediction of Attitude Change.” Psychological Review 62 (1): 42–55. Owen, Diana, and Richard Davis. 2008. “Presidential Communication in the Internet Era.” Presidential Studies Quarterly 38 (4): 658–73. Peake, Jeffrey S. 2007. “Presidents and Front-Page News: How America’s Newspapers Cover the Bush Administration.” International Journal of Press/Politics 12 (4): 52–70. Persily, Nathaniel. 2008. Introduction to Public Opinion and Constitutional Controversy, edited by Nathaniel Persily, Jack Citrin, and Patrick J. Egan, 3–17. New York: Oxford University Press. Persily, Nathaniel, Gillian E. Metzgar, and Trevor W. Morrison. 2013. Introduction to The Health Care Case: The Supreme Court Decision and Its Implications, edited by Nathaniel Persily, Gillian E. Metzgar, and Trevor W. Morrison, 1–8. Oxford: Oxford University Press. Petty, Richard E., and John T. Cacioppo. 1981. “Issue Involvement as a Moderator of the Effects on Attitude of Advertising Content and Context.” In Advances in Consumer Research, vol. 8, edited by Kent B. Monroe, 20–24. Ann Arbor, MI: Association for Consumer Research. Petty, Richard E., John T. Cacioppo, and David Schumann. 1983. “Central and Peripheral Routes to Advertising Effectiveness: The Moderating Role of Involvement.” Journal of Consumer Research 10 (2): 135–46. Postman, Neil. 1985. Amusing Ourselves to Death: Public Discourse in the Age of Show Business. New York: Penguin. Plessy v. Ferguson. 1896. 163 U.S. 537. Reips, Ulf-Dietrich. 2000. “The Web Experiment Method: Advantages, Disadvantages, and Solutions.” In Psychological Experiments on the Internet, edited by Michael H. Birnbaum, 89–117. San Diego: Academic Press. Riker, William H. 1995. “The Political Psychology of Rational Choice Theory.” Political Psychology 16 (1): 23–44. Roberts, John G., Jr. 1993. “Article III Limits on Statutory Standing.” Duke Law Journal 42 (6): 1219–32.

Roe v. Wade. 1973. 410 U.S. 113. Rosenberg, Gerald N. 1991. The Hollow Hope: Can Courts Bring about Social Change? Chicago: University of Chicago Press. Rosenberg, Gerald N. 1995. “The Real World of Constitutional Rights: The Supreme Court and the Implementation of Abortion Decisions.” In Contemplating Courts, edited by Lee Epstein, 390–419. Washington, DC: Congressional Quarterly. Page 233 →Rucker, Derek D., and Richard E. Petty. 2006. “Increasing the Effectiveness of Communications to Consumers: Recommendations Based on Elaboration Likelihood and Attitude Certainty Perspectives.” Journal of Public Policy & Marketing 25 (1): 39–52. Sala, Brian R., and James F. Spriggs II. 2004. “Designing Tests of the Supreme Court and the Separation of Powers.” Political Research Quarterly 57 (2): 197–208. Sax, Joseph L. 2005. “Kelo: A Case Rightly Decided.” University of HawaiвЂi Law Review 28 (2): 365–71. Schlesinger, Philip, and Howard Tumber. 1994. Reporting Crime: The Media Politics of Criminal Justice. Oxford: Clarendon. Schudson, Michael. 2002. “The News Media as Political Institutions.” Annual Review of Political Science 5:249–69. Seawright, Jason, and John Gerring. 2008. “Case Selection Techniques in Case Study Research: A Menu of Qualitative and Quantitative Options.” Political Research Quarterly 61 (2): 294–308. Segal, Jeffrey A. 1997. “Separation-of-Powers Games in the Positive Theory of Congress and Courts.” American Political Science Review 91 (1): 28–44. Segal, Jeffrey A., and Harold J. Spaeth. 2002. The Supreme Court and the Attitudinal Model Revisited. Cambridge: Cambridge University Press. Shear, Michael D., and Ashley Parker. 2012. “Romney Says He Will вЂRepeal Obamacare’ If Elected.” Accessed January 12, 2013. http://thecaucus.blogs.nytimes.com/2012/06/28/romney-says-he-willrepeal-obamacare-if-elected/. Sherman, Rorie. 1988. “The Media and the Law.” National Law Journal, June 6, 32–36. Shesol, Jeff. 2010. Supreme Power: Franklin Roosevelt vs. the Supreme Court. New York: Norton. Sigal, Leon V. 1986. “Sources Make the News.” In Reading the News, edited by R.K. Manoff and Michael Schudson, 9–37. New York: Pantheon. Sirico, Louis J., Jr. 2010. “Failed Constitutional Metaphors: The Wall of Separation and the Penumbra.” University of Richmond Law Review 45 (2): 459–89. Skitka, Linda J., Christopher W. Bauman, and Edward G. Sargis. 2002. “Moral Conviction: Another Contributor to Attitude Strength or Something More?” Journal of Personality and Social Psychology 88 (6): 895–917. Slotnick, Elliot E. 1991. “Media Coverage of Supreme Court Decision Making: Problems and Prospects.” Judicature 75 (3): 128–42. Slotnick, Elliot E. 1993. “Television News and the Supreme Court: A Case Study.” Judicature 77 (1):

21–33. Slotnick, Elliot E., and Jennifer Segal. 1998. Television News and the Supreme Court: All the News That’s Fit to Air? New York: Cambridge University Press. Sniderman, Paul M., Richard A. Brody, and Philip E. Tetlock.1991. Reasoning and Choice: Explorations in Social Psychology. Cambridge: Cambridge University Press. Sobel, Joel. 1985. “A Theory of Credibility.” Review of Economic Studies 52 (4): 557–73. Sobieraj, Sarah, and Jeffrey M. Berry. 2011. “From Incivility to Outrage: Political Discourse in Blogs, Talk Radio, and Cable News.” Political Communication 28 (1): 19–41. Spaeth, Harold J. 2010. “The Supreme Court Database.” Accessed February 8, 2013. Scdb.org. Page 234 →Spill, Rorie L., and Zoe M. Oxley. 2003. “Philosopher Kings or Political Actors: How the Media Portray the Supreme Court.” Judicature 87 (1): 22–29. Spillotes, Constantine, and Lynn Vavreck. 2002. “Campaign Advertising: Partisan Convergence of Divergence?” Journal of Politics 64 (1): 249–61. Spriggs, James F., and Paul J. Wahlbeck. 1997. “Amicus Curiae and the Role of Information at the Supreme Court.” Political Research Quarterly 50 (2): 365–86. Stanton, Jeffrey M. 1998. “An Empirical Assessment of Data Collection Using the Internet.” Personnel Psychology 51 (3): 709–25. Sterk, Stewart E. 2006. “The Demise of Federal Takings Legislation.” William and Mary Law Review 48 (1): 251–301. Stoutenborough, James W., and Donald R. Haider-Markel. 2008. “Public Confidence in the U.S. Supreme Court: A New Look at the Impact of Court Decisions.” Social Science Journal 45 (1): 28–47. Stoutenborough, James W., Donald P. Haider-Markel, and Mahalley D. Allen. 2006. “Reassessing the Impact of Supreme Court Decisions on Public Opinion: Gay Civil Rights Cases.” Political Research Quarterly 59 (3): 419–33. Tannenbaum, Percy H. 1967. “The Congruity Principle Revisited.” In Advances in Experimental Social Psychology, vol. 41, edited by Leonard Berkowitz, 271–320. New York: Academic Press. Tannenbaum, Percy H., Jacqueline R. Macauley, and Eleanor L. Norris. 1966. “Principle of Congruity and Reduction of Persuasion.” Journal of Personality and Social Psychology 3 (2): 233–38. Terkildsen, Nayda, Frauke I. Schnell, and Christina Ling. 1998. “Interest Groups, the Media, and Policy Debate Formation: An Analysis of Message Structure, Rhetoric, and Source Cues.” Political Communication 15 (1): 45–61. Texas v. Johnson. 1989. 491 U.S. 397. Toobin, Jeffrey. 2002. Too Close to Call: The Thirty-Six Day Battle to Decide the 2000 Election. New York: Random House. Toobin, Jeffrey. 2012. The Oath: The Obama White House and the Supreme Court. New York: Doubleday. Tyler, Tom R. 1990. Why People Obey the Law: Procedural Justice, Legitimacy, and Compliance. New Haven: Yale University Press.

Tyler, Tom R. 2006. “Does the American Public Accept the Rule of Law? The Findings of Psychological Research on Deference to Authority.” DePaul Law Review 56 (2): 661–94. Unah, Isaac, and Ange-Marie Hancock. 2006. “U.S. Supreme Court Decision Making, Case Salience, and the Attitudinal Model.” Law and Policy 28 (3): 295–320. Valentino, Nicholas, Vincent Hutchings, and Ismail White. 2002. “Cues That Matter: How Political Ads Prime Racial Attitudes during Campaigns.” American Political Science Review 96 (1): 75–90. Winkler, Carol. 2007. “Parallels in Preemptive War Rhetoric: Reagan on Libya, Bush 43 on Iraq.” Rhetoric and Public Affairs 10 (2): 303–34. Winkler, Carol. 2008. “Encroachments on State Sovereignty: The Argumentation Strategies of the George W. Bush Administration.” Argumentation 22 (4): 473–88. Wisneski, Daniel C., Brad L. Lytle, and Linda J. Skitka. 2009. “Gut Reactions: Moral Convictions, Religiosity, and Trust in Authority.” Psychological Science 20 (9): 1059–63. Page 235 →Witt, Elder. 1990. Guide to the Supreme Court. Washington, DC: CQ Press. Woodson, Benjamin, James L. Gibson, and Milton Lodge. 2011. “Judicial Symbols and the Link between Institutional Legitimacy and Acquiescence.” Paper presented at the annual meeting of the American Political Science Association, Seattle. Wroth, L. Kinvin. 2005. “Lingle and Kelo: The Accidental Tourist in Canada and NAFTA-Land.” Vermont Journal of Environmental Law 7 (1): 62–83. Xenos, Michael, and W. Lance Bennett. 2007. “The Disconnection in Online Politics.” Information, Communication, and Society 10 (4): 443–64. Young, Ernest A. 2002. “Judicial Activism and Conservative Politics.” Colorado Law Review 73 (4): 1139–1216. Zaller, John R. 1992. The Nature and Origins of Mass Opinion. Cambridge: Cambridge University Press. Zaller, John R. 1996. “The Myth of Massive Media Impact Revived: New Support for a Discredited Idea.” In Political Persuasion and Attitude Change, edited by Diana C. Mutz, Paul M. Sniderman, and Richard A. Brody, 17–78. Ann Arbor: University of Michigan Press. Zaller, John R., and Stanley Feldman. 1992. “A Simple Theory of the Survey Response: Answering Questions versus Revealing Preferences.” American Journal of Political Science 36 (3): 579–616. Zink, James R., James F. Spriggs II, and John T. Scott. 2009. “Courting the Public: The Influence of Decision Attributes on Individuals’ Views of Court Opinions.” Journal of Politics 71 (3): 909–25.

Page 236 → Page 237 →

Index acceptance (of Supreme Court rulings), 10, 15–19, 173, 209n2 acquiescence (to Supreme Court rulings). See acceptance (of Supreme Court rulings) activism, judicial definition of, 47, 215n17 dissensus dynamics model and, 86–88, 104, 108, 112, 172 measurement of, 183 Adarand v. Pena, 6, 123 Affordable Care Act (ACA). See National Federation of Independent Business v. Sebelius Althaus, Scott L., 177 amicus curiae, 82, 90, 214n9 Arizona et al. v. U.S., 175 attitudes changes in, 143–48, 162–68, 170 (see also receive-accept-sample model [RAS]) crystallization of, 20, 145–46 elaboration likelihood model of, 121–22 properties of, 147 strength of, 146–48, 165–66, 217n2, 218n10 Atkins v. Virginia, 105 Atwater v. Lago Vista, 33 backlash examples of, 10–11, 13, 57–59, 70, 216n1 theories of, 4–5, 10–11, 16, 21–22 unanswered questions about, 141–42, 210n1 (chap. 2) Bailey, Michael A., 80, 81, 82, 188, 214n5 See also Bailey ideal point estimates Bailey ideal point estimates, 80, 81, 82, 188, 214n5 Baron, Reuben M., 111

Bennis v. Michigan, 85 Berman v. Parker, 51, 56–57 Blake, William D., 81 Breyer, Stephen, 7, 11 Brown v. Board of Education, 16, 176 Burlington Industries v. Ellerth, 104 Bush, George W., 17, 99, 177 See also Bush v. Gore Bush v. Gore, 13, 14, 17, 218n11 Bush v. Vera, 109 Caldeira, Gregory A., 13, 146, 150, 162, 173 See also legitimacy Carrubba, Clifford J., 214n11 cascading activation model, 41, 88–89 case salience definition of, 16–17 dissensus dynamics model and, 35, 78, 83–84, 102–5, 109, 185 influence of, 16–17, 20, 122, 124, 136, 140, 144 Kelo v. City of New London and, 49, 59, 61–62, 129 Page 238 →case salience (continued) measurement of, 26, 88–89 National Federation of Independent Business v. Sebelius and, 145 case study method, 50–51, 211n5 Chong, Dennis, 179 See also cues; framing effects Citrin, Jack, 4, 11 Claiborne v. U.S., 37 Clawson, Rosalee A., 6, 123 Clinton, Bill, 104

compliance (with Supreme Court rulings). See acceptance (of Supreme Court rulings) conditional response hypothesis, 17–18, 167 receive-accept-sample model (RAS) and, 18 controversial rulings (of the Supreme Court). See backlash See also Bush v. Gore; District of Columbia v. Heller; Furman v. Georgia, Kelo v. City of New London; Miranda v. Arizona; National Federation of Independent Business v. Heller; Roe v. Wade; Texas v. Johnson; Webster v. Reproductive Health Services Court-packing plan, 15 CNN Crossfire, 111 reporting of, 6, 32–33, 67, 69, 70, 97, 99, 105, 109, 111–12 cues definition of, 120–21 dissensus dynamics model and, 77–78, 91, 98, 107 influence of, 121–28, 129–42, 216n1 (chap. 7) theories of, 121–28 Davis, Richard, 175 decision approval. See policy agreement, measurement of decision support, 79–80, 83, 85, 86, 94, 102–4, 108–14, 185–90 diffuse support (for the Supreme Court). See legitimacy District of Columbia v. Heller, 85, 89, 95–96 dissensus dynamics model alternatives to, 41–47 applications of, 8–9, 74–76, 98–101, 172–73, 177–78 measurement of, 80–82, 93–94, 102–3 predictions of: elite influence, 40–41; dissent hypothesis, 39; ideological diversity hypothesis, 40; televised incivility hypothesis, 101 robustness of, 93–95, 115 scope of, 34–36 summary of evidence concerning, 115–16 tests of, 77–116

dissent examples of (see O’Connor, Sandra Day; Thomas, Clarence) explanations for (see Supreme Court, strategic behavior on) influence of (see dissensus dynamics model) Druckman, James N., 179 See also cues; framing effects Easton, David, 13, 124 See also legitimacy Egan, Patrick J., 4, 11 elite influence on news coverage, 40–41 Entman, Robert M., 41, 42 See also cascading activation model Epstein, Lee, 78 Eskridge, William N., Jr., 46 Fox News Hannity and Colmes, 111–12 O’Reilly Factor, 63 reporting of, 32, 63, 97–99, 105 framing definition of, 122 dissensus dynamics model and, 48, 59, 64, 68, 73–74, 172 Page 239 →National Federation of Independent Business v. Sebelius and: case salience, 143–45; discussion of taxes, 145; front page coverage, 119; television coverage, 159–62 influence of, 6, 122–27 measurement of, 102, 181–82 models of: cascading activation, 41, 88–89; indexing, 41, 75 operationalizations of, 130–31, 137, 159, 197–207 tests of: discussion of, 140–42, 168–70; participant pool in, 130–31, 136–38, 148–49, 151–53; results of, 132–36, 138–40, 159–62; treatments in, 131, 136–38, 197–207 theories of, 41, 122–25

framing effects. See framing Franklin, Charles H., 11, 15, 17, 19, 157, 179 See also structural response hypothesis Furman v. Georgia, 11 Gerring, John, 211n5 See also case study method Gibson, James L., 13, 14, 146, 150, 162, 173, 218n8 See also legitimacy Ginsburg, Ruth Bader, 7, 38, 70, 175–76 Gore, Al, 174 See also Bush v. Gore Greenhouse, Linda (New York Times), 26, 39, 69 Griswold v. Connecticut, 29 See also Roe v. Wade Grosskopf, Anke, 13, 146 Grutter v. Bollinger, 105 gun control. See District of Columbia v. Heller; U.S. v. Lopez Hacker, Hans J., 81 Hawaii Housing Authority v. Midkiff, 56–57, 71 heuristics, 121–22 See also attitudes; cues Ho, Daniel E., 214n6 indexing model, 41, 75 institutional loyalty hypothesis, 124–25, 133–34, 162–66 Johnson, Richard M., 173 Johnson, Timothy R., 17 See also conditional response hypothesis Kagan, Elena, 7, 38 Kelo v. City of New London

backlash to, 4, 11, 123, 127, 129, 142, 216n1 dissensus dynamics model and, 8, 48–50, 53–76, 91–93, 125, 168 long term influence of, 75–76, 213n34 news framing of: dissent-laden, 61–62, 72–76, 105, 110, 139, 172, 209n6; government power, 71–72, 98–99, 189; human interest stories, 33–34, 68–69, 211n5; ideological motivations, 100–101, 113; mentions of constitutional principles, 69–70; privileging of wealthy interests, 72, 98–99; public property in danger, 70–71 role of ideology in, 71–72, 113 salience of, 59–60, 75, 83–84, 105 strategic behavior in, 175, 178 summary of, 49, 52–53 Kennedy, Anthony, 1, 34, 39, 111 Kenny, David A., 111 Kimbrough v. U.S., 104, 109 Kosaki, Liane C., 11, 15, 17, 19, 157, 179 See also structural response hypothesis Lane, Charles (Washington Post), 69 Ledbetter v. Goodyear Tire & Rubber Co., 70, 175 Page 240 →legitimacy changes in: 143–48, 162–68, 170; hypothesis about, 125 definition of, 13 formation of, 11–14 influence of, 2–3 limitations on, 124–27, 132–36, 138–42, 144, 147–48, 162–68, 170, 173–74, 179–80 measurement of, 14, 132, 138, 143, 150, 216n2 (chap. 8), 217n2 strategic behavior with respect to, 176 See also institutional loyalty hypothesis; relevance hypothesis Lerman, Amy E., 10 Lingle v. Chevron dissensus dynamics model and, 8, 48–50, 53–76 news coverage of, 64–66, 72–76, 211n5

salience of, 55, 60–62 summary of, 49, 51–52 Lodge, Milton, 218n8 Lorrilard Tobacco Company v. Reilly, 109 loyalty (to the Supreme Court). See legitimacy Lupia, Arthur, 121 Marshall, Thomas, 4 Martin, Andrew D., 17, 80, 214n6 See also conditional response hypothesis; Martin-Quinn scores Martin-Quinn scores, 80, 214n6 McCain, John, 95 media effects. See media effects hypothesis media effects hypothesis, 124–28, 132–33, 138–42, 159–62, 168–70 See also cues; framing mediation analysis, 111–12 Miller-El v. Dretke, 85 Miranda v. Arizona, 11 Mondak, Jeffrey J., 13, 122, 123, 146 MSNBC Hardball, 111 reporting of, 100, 113 The Situation, 100–101 Mutz, Diana C., 100, 103, 219n1 (appendix) National Federation of Independent Business v. Sebelius attitudes about, 2–4, 9, 128, 142, 143–70, 172–74 Commerce Clause and, 7, 32 complexity of, 37–38, 97–98 dissensus dynamics model and, 28–34, 37–38, 43–44, 172 news coverage of, 2, 6–7, 28–34, 37–38, 97–98, 119, 145, 159–62, 172, 205–7

reactions to, 33, 34, 43, 95, 174 strategic behavior in, 12 summary of, 1–2 National Public Radio (NPR), 66, 69, 71 negativity bias, 13, 146, 162 see also legitimacy; positivity bias New York Times dissensus dynamics model and, 59, 79, 185–89 examples of coverage in, 65, 69, 85, 119 reporting principles of, 7, 16, 26–27, 43, 57, 65, 78, 83, 106 newspaper coverage (of the Supreme Court) constraints in: accuracy, 30–32; conflict, 33–34; drama, 32–33; simplicity, 29–30; timeliness, 32 content of, 79–80, 84–86 (see also decision support) dissensus dynamics model and, 8, 35–41, 49, 74–76, 86–96 elite influence on, 40–41 influence of ideology in, 71–72, 79, 188–89, 214n6 influence on attitudes from (see framing) reporting practices in, 27–29 See also New York Times; Washington Post; Washington Times Page 241 →O’Connor, Sandra Day, 52–54, 62–63, 65, 68–70, 75, 83, 175, 209n6 Obama, Barack, 33, 95 Obamacare. See National Federation of Independent Business v. Sebelius Patient Protection and Affordable Care Act. See National Federation of Independent Business v. Sebelius Persily, Nathaniel, 4 Plessy v. Ferguson, 12 Poletown Neighborhood Council v. City of Detroit, 56–57 policy agreement examples of, 15–16 evidence for, 15–16

evidence against (see backlash) limitations on, 16–19, 122–25, 153–68, 173, 179–80 literature on: 11–22; shortcomings of, 19–22 long term trends in, 16 measurement of, 19–20, 132, 138, 149–50, 210n3 (chap. 2), 216n2 (chap. 8), 217nn5–6 moderating factors for: case salience, 16–17; decision characteristics, 18; news coverage, 120–28, 132–36, 138–42, 144–46, 153–68, 173; preexisting opinion, 17; sequence of decisions, 17–18 tests of, 132–36, 138–42, 153–68 theory of, 3–5, 11–16, 116, 144 policy legitimation. See policy agreement positivity bias, 13, 127, 133, 146, 167, 218n8 See also legitimacy; negativity bias procedural fairness, 11–15, 22, 144–45, 176 Public Use Clause. See Kelo v. City of New London; Lingle v. Chevron Quinn, Kevin M., 80, 188, 214n6 See also Martin-Quinn scores R.A.V. v. City of St. Paul, 85 Reeves, Byron, 100, 103, 219n1 receive-accept-sample model (RAS), 18 Rehnquist, William, 52, 53, 69, 99 relevance hypothesis, 126–27, 134–35, 140, 162–66 See also cues; framing; media effects reporting (on the Supreme Court) content of, 84–86, 104–6 negativity in (see televised incivility) objectives of, 26–34 partisanship and ideology in, 107–8, 113, 115 See also newspaper coverage (of the Supreme Court) response (to Supreme Court decisions), theories of acceptance, 10, 15–19, 173

acquiescence, 10, 15–19, 173 backlash, 4–5, 10–11, 16, 21–22 conditional response hypothesis, 17–18, 167 negativity bias, 13, 146, 162 policy agreement, 3–5, 11–16, 116, 144 policy legitimation, 3–5, 11–16, 116, 144 positivity bias, 13, 127, 133, 146, 167 procedural fairness, 11–15, 22, 144–45, 176 receive-accept-sample model (RAS), 18 structural response hypothesis, 17, 157–58, 166, 179–80 Roberts, John as Chief Justice, 151 legal philosophy of, 12, 176 role in National Federation of Independent Business v. Sebelius: ally with liberal justices, 39, 119, 144; author of majority opinion, 2, 4, 6–7, 37–38, 128, 172; subject of criticism, 34, 43, 205–6 Roe v. Wade attitudes about, 10, 15–19, 157 news coverage of, 29–32, 188 Romney, Mitt, 34, 43, 174 Page 242 →Rosenberger v. Virginia, 131 Rumsfeld v. Forum for Academic and Institutional Rights, 84–85 San Remo Hotel v. San Francisco, 211n3 Scalia, Antonin, 39, 52, 69, 175 Schudson, Michael, 40 Seawright, Jason, 211n5 See also case study method Segal, Jeffrey A., 78 Smiley v. Citibank, 109 Sotomayor, Sonia, 7 source credibility, 11, 14, 121, 122, 127–28, 135, 139, 142, 178, 179

Souter, David, 71, 127 Spence, Lester Kenyatta, 13, 150, 162, 173 See also legitimacy Stevens, John Paul, 39, 52, 67, 69 structural response hypothesis, 17, 157–58, 166, 179–80 Supreme Court controversial rulings of (see backlash) credibility of, 11, 14, 122, 127–28, 135, 139, 142, 178, 179, 209n3 critics of: justices, 6–7, 172, 175–76, 178; media, 6–7, 85, 104–6, 112, 119, 128; political elite, 42–43, 95–96, 174; public, 11, 13, 17 (see also backlash; Kelo v. City of New London, backlash to) docket of, 26, 28 history of, 26–27 information released by, 36–39 justices of (see Breyer, Stephen; Ginsburg, Ruth Bader; Kagan, Elena; Kennedy, Anthony; O’Connor, Sandra Day; Rehnquist, William; Roberts, John; Scalia, Antonin; Sotomayor, Sonia; Souter, David; Stevens, John Paul; Thomas, Clarence) opinions of, 27, 37 (see also dissent) oral arguments on, 26–27 packing of, 15 persuasive influence of, 15 (see also policy agreement) popular indifference to, 20 Public Information Office of, 27 republican schoolmaster role of, 10, 120, 179 (see also Franklin, Charles H.; Kosaki, Liane C.; policy agreement) strategic behavior on, 12, 174–76, 210n1 (chap. 3) support for (see legitimacy) symbols of authority of, 12–13 unanimity on: approval for, 18; desire for, 12, 45, 113, 175, 176; examples of, 51, 56, 176; news coverage of, 38, 42, 44, 49, 58, 65, 86, 89–91, 110, 172 voting outcomes on, 38–40 (see also dissent) Supreme Court Database, 80, 214n8 Takings Clause. See Kelo v. City of New London; Lingle v. Chevron

Taylor, Guy (Washington Times), 69 televised incivility definition of, 100–101, 219n1 dissensus dynamics model and, 101, 108–13 examples of, 98–101, 105, 110, 113; on Crossfire (CNN), 111; on Hannity and Colmes (Fox News), 111–12; on Hardball (MSNBC), 111 measurement of, 103 statistical test regarding, 111–12 television news coverage (of the Supreme Court) constraints on: accuracy, 30–32; conflict, 33–34; drama, 32–33; simplicity, 29–30; timeliness, 32 content of, 102, 104–6 differences from newspaper coverage in (see televised incivility) dissensus dynamics model and, 8, 35–41, 49, 74–76, 100–101, 108–9, 113–15 Page 243 →influence of ideology in, 71–72, 79, 106–7, 188–89 influence on attitudes from (see framing) reporting principles of, 27–34, 77–78, 210n3 (chap. 3) See also CNN; Fox News; MSNBC Texas v. Johnson, 11, 13 Thomas, Clarence behavior during oral arguments of, 26 dissenting opinions of: in Establishment Clause cases, 105; in Kelo v. City of New London, 52, 54, 63, 68, 70, 72, 75; in National Federation of Independent Business v. Sebelius, 38 Toobin, Jeffrey (CNN), 67, 69 Totenberg, Nina (NPR), 69 U.S. Term Limits v. Thornton, 216n3 U.S. v. Lopez, 216n3, 217n5 (chap. 8) U.S. v. Playboy, 105–6, 111–12 U.S. v. Winstar Corporation, 84 Waltenburg, Eric N., 6, 123 Warren, Earl, 176

Washington Post dissensus dynamics model and, 59, 79, 185–89 examples of coverage in, 49, 65, 84–85, 95–96, 119 reporting principles of, 27, 33, 49, 63–66, 69, 78, 106, 130 Washington Times, 59, 66, 71, 70, 85, 119, 185–89, 214n4 Webster v. Reproductive Health Services, 83–84, 185 Woodson, Benjamin, 218n8