The Politics of Numbers 0871540150, 9780871540157

Table of contents :
Foreword
Contents
Contributors
Introduction
The Sociology of Official Statistics
Part I. The Politics of Economic Measurement
1 The Politics of Comparative Economic Statistics: Three Cultures and Three Cases
2 The Politics of Income Measurement
3 Political Purpose and the National Accounts
Part II. The Politics of Population Measurement
4 The 1980 Census in Historical Perspective
5 Politics and the Measurement of Ethnicity
6 The Social and Political Context of Population Forecasting
Part III. Statistics and Democratic Politics
7 Public Statistics and Democratic Politics
8 The Political Foundations of American Statistical Policy
9 Statistics and the Politics of Minority Representation: The Evolution of the Voting Rights Act Since 1965
Part IV. Statistics and American Federalism
10 The Politics of Printouts: The Use of Official Numbers to Allocate Federal Grants-in-Aid
11 Federal Statistics in Local Governments
12 The Managed Irrelevance of Federal Education Statistics
Part V. The New Political Economy of Statistics
13 Technology, Costs, and the New Economics of Statistics
14 Who Will Have the Numbers? The Rise of the Statistical Services Industry and the Politics of Public Data
Name Index
Subject Index

Citation preview

THE POLITICS OF NUMBERS

THE POLITICS OF NUMBERS W illiam Alonso and

Paul Starr Editors

for the National Committee for Research on the 1980 Census

RUSSELL SAGE FOUNDATION

/

NEW YORK

The Russell Sage Foundation The Russell Sage Foundation, one of the oldest of America's general purpose founda­ tions, was established in 1907 by Mrs. Margaret Olivia Sage for "the improvement of social and living conditions in the United States." The Foundation seeks to fulfill this mandate by fostering the development and dissemination of knowledge about the political, social, and economic problems of America. It conducts research in the social sciences and public policy, and publishes books and pamphlets that derive from this research. The Board of Trustees is responsible for oversight and the general policies of the Foundation, while administrative direction of the program and staff is vested in the Pres­ ident, assisted by the officers and staff. The President bears final responsibility for the decision to publish a manuscript as a Russell Sage Foundation book. In reaching a judg­ ment on the competence, accuracy, and objectivity of each study, the President is advised by the staff and selected expert readers. The conclusions and interpretations in Russell Sage Foundation publications are those of the authors and not of the Foundation, its Trus­ tees, or its staff. Publication by the Foundation, therefore, does not imply endorsement of the contents of the study. BOARD OF TRUSTEES John S. Reed, Chair Robert McCormick Adams Earl F. Cheit Philip E. Converse Reneé Fox

Henna Hill Kay Carl Kaysen Patricia King Gardner Lindzey

Gary MacDougal James G. March Madelon Talley Eric Wanner

library of Congress Cataloging-m-Publicadon Data The Politics of numbers. Chiefly papers prepared for a conference on "The political economy of national statistics," held in Washington, D.C., Oct. 13-15, 1983. Bibliography: p. Includes index. 1. United States—Statistical services—Political aspects. 2. United States—Census, 1980. I. Alonso, William. U. Starr, Paul, 1949HA37.U55P65 1986 320'.0723 86-10060 ISBN 0-87154-015-0 Copyright © 1987 by Russell Sage Foundation. All rights reserved. Printed in the United States of America. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the prior written permission of the publisher. The paper used in this publication meets the minimum requirements of American National Standard for Information Sciences—Permanence of Paper for Printed Library Materials. ANSI Z39 48-1984. Cover and text design: h u g u e tte f r a n c o 10 9 8 7 6 5 4 3

THE POPULATION OF THE UNITED STATES IN THE 1980s A Census Monograph Series

The National Committee for Research on the 1980 Census Charles F. Westoff Princeton University Chairman and Executive Director John S. Adams University of Minnesota Anthony Downs The Brookings Institution Leobardo Estrada University of California, Los Angeles Reynolds Farley University of Michigan Victor R. Fuchs Stanford University Bernard R. Gifford University of California, Berkeley Paul C. Glick Arizona State University

Tamara K. Hareven Clark University Nathan Keyfitz Harvard University Cora B. Marrett University of Wisconsin Robert K. Merton Columbia University Isabel V. Sawhill The Urban Institute William H. Sewell University of Wisconsin Michael S. Teitelbaum Alfred P. Sloan Foundation James R. Wetzel U.S. Bureau of the Census Raymond E. Wolfinger University of California, Berkeley

Sidney Goldstein Brown University

Staff

Charles V* Hamilton Columbia University

David L Sills Social Science Research Council

The committee is sponsored by the Social Science Research Council, the Russell Sage Foundation, and the Alfred P. Sloan Foundation, in collaboration with the U.S. Bureau of the Census. The opinions, findings, and conclusions or recommendations expressed the monographs supported by the committee are those of the author(s) and do not necessarily reflect the views of the committee or its sponsors.

Foreword

"The Population of the United States in the 1980s" is an ambitious series of volumes aimed at converting the vast statistical yield of the 1980 census into authoritative analyses of major changes and trends in American life. The Politics of Numbers resembles the other volumes in this series in that its point of departure is the United States census; it differs from the other volumes by going far beyond census data to an ex­ am ination of the compilation and analysis of other official data as well. It is the only volume in the series devoted to the governmental data sys­ tem itself, rather than to the inform ation contained in that data. More­ over, unlike the other series volumes, it is not a monograph but a collection of essays on official statistics, the outgrowth of a 1983 confer­ ence on the subject sponsored by the N ational Committee for Research on the 1980 Census. This series represents an im portant episode in social science re­ search and revives a long tradition of independent census analysis. First in 1930, and then again in 1950 and 1960, teams of social scientists worked w ith the U.S. Bureau of the Census to investigate significant so­ cial, economic, and demographic developments revealed by the decen­ nial census. 111686 census projects produced three landmark series of studies, providing a firm foundation and setting a high standard for our present undertaking. There is, in fact, more than a theoretical continuity between those earlier census projects and the present one. Like those previous efforts, this new census project has benefited from close cooperation between the Census Bureau and a distinguished, interdisciplinary group of schol­ ars. Like the 1950 and 1960 research projects, research on the 1980 census was initiated by the Social Science Research Council and the Russell Sage Foundation. In deciding once again to promote a coordi­ nated program of census analysis, Russell Sage and the Council were mindful not only of the severe budgetary restrictions imposed on the Census Bureau's own publishing and dissem ination activities in the

THE POLITICS OF NUMBERS

1980s, but also of the extraordinary changes that have occurred in so many dimensions of American life over the past two decades. The studies constituting "The Population of the United States in the 1980s" were planned, commissioned, and monitored by the Na­ tional Committee for Research on the 1980 Census, a special commit­ tee appointed by the Social Science Research Council and sponsored by the Council, the Russell Sage Foundation, and the Alfred P. Sloan Foun­ dation, w ith the collaboration of the U.S. Bureau of the Census. This com m ittee includes leading social scientists from a broad range of fields—demography, economics, education, geography, history, political science, sociology, and statistics. It has been the com m ittee's task to select the m ain topics for research, obtain highly qualified specialists to carry out that research, and provide the structure necessary to facilitate coordination among researchers and w ith the Census Bureau. The topics treated in this series span virtually all the major features of American society—ethnic groups (blacks, Hispanics, foreign-bom); spatial dimensions (migration, neighborhoods, housing, regional and m etropolitan growth and decline); and status groups (income levels, families and households, women). Authors were encouraged to draw not only on the 1980 Census but also on previous censuses and on subse­ quent national data. Each individual research project was assigned a special advisory panel made up of one com m ittee member, one member nominated by the Census Bureau, one nominated by the National Sci­ ence Foundation, and one or two other experts. These advisory panels were responsible for project liaison and review and for recommenda­ tions to the National Committee regarding the readiness of each m anu­ script for publication. W ith the final approval of the chairman of the National Committee, each report was released to the Russell Sage Foundation for publication and distribution. The debts of gratitude incurred by a project of such scope and or­ ganizational complexity are necessarily large and numerous. The com­ m ittee m ust thank, first, its sponsors—the Social Science Research Council, headed until recently by Kenneth Prewitt; the Russell Sage Foundation, under the direction of president M arshall Robinson; and the Alfred P. Sloan Foundation, led by Albert Rees. The long-range vi­ sion and day-to-day persistence of these organizations and individuals sustained this research program over many years. The active and w ill­ ing cooperation of the Bureau of the Census was clearly invaluable at all stages of this project, and the extra com m itm ent of tim e and effort made by Bureau economist James R. Wetzel m ust be singled out for spe­ cial recognition. A special tribute is also due to David L. Sills of the So­ cial Science Research Council, staff member of the committee, whose

viii

Foreword organizational, administrative, and diplomatic skills kept this compli­ cated project running smoothly. The com m ittee also wishes to thank those organizations that con­ tributed additional funding to the 1980 Census project—the Ford Foun­ dation and its deputy vice president, Louis W innick, the National Sci­ ence Foundation, the National Institute on Aging, and the National Institute of Child H ealth and Human Development. Their support of the research program in general and of several particular studies is gratefully acknowledged. The ultim ate goal of the N ational Com m ittee and its sponsors has been to produce a definitive, accurate, and comprehensive picture of the U.S. population in the 1980s, a picture that would be prim arily descrip­ tive but also enriched by a historial perspective and a sense of the chal­ lenges for the future inherent in the trends of today. We hope our readers w ill agree that the present volume takes a significant step to­ ward achieving th at goal. C h a r l e s F. W e s t o f f

Chairman and Executive Director, National Committee for Research on the 1980 Census

Contents

Introduction William Alonso and Paul Starr

1

The Sociology of Official Statistics Paul Starr

7

PARTI

THE POLITICS OF ECONOMIC MEASUREMENT

1 The Politics of Comparative Economic Statistics: Three Cultures and Three Cases Raymond Vernon 2 The Politics of Income M easurement Christopher Jencks 3 Political Purpose and the National Accounts Mark Perlman PART R

THE POLITICS OF POPULATION MEASUREMENT

59

61

83

133

153

4 The 1980 Census in Historical Perspective Margo A. Conk

155

5 Politics and the M easurement of Ethnicity William Petersen

187

6 The Social and Political Context of Population Forecasting Nathan Keyfitz

235

xi

THE POUTICS OP NUMBERS

FART III

STATISTICS AND DEMOCRATIC POLITICS

259

7 Public Statistics and Democratic Politics Kenneth Prewitt

261

8 The Political Foundations of American Statistical Policy Steven Kelman

275

9 Statistics and the Politics of M inority Representation: The Evolution of the Voting Rights Act Since 1965 Abigail Themstrom

PART IV

STATISTICS AND AMERICAN FEDERALISM

10 The Politics of Printouts: The Use of Official Numbers to Allocate Federal Grants-in-Aid Richard P. Nathan

303

329

331

11 Federal Statistics in Local Governments Judith de Neufvßle

343

12 The Managed Irrelevance of Federal Education Statistics Janet A. Weiss and Judith E. Gruber

363

PART V

THE NEW POLITICAL ECONOMY OF STATISTICS

393

13 Technology, Costs, and the New Economics of Statistics Joseph W. Duncan

395

14 Who Will Have the Numbers? The Rise of the Statistical Services Industry and the Politics of Public Data Paul Starr and Ross Corson

xii

415

Name Index

449

Subject Index

461

Contributors

W illiam Alonso, Harvard University Paul Starr, Princeton University Margo Conk, University of Wisconsin Judith de Neufville, University of California, Berkeley Joseph Duncan, Dun é> Bradstreet Judith Gruber, University of California, Berkeley Christopher Jencks, Northwestern University Steven Kelman, Harvard University N athan Keyfitz, Harvard University Richard P. Nathan, Princeton University Mark Perlman, University of Pittsburgh W illiam Petersen, Carmel, California Kenneth Prewitt, Rockefeller Foundation Abigail Themstrom, Lexington, Massachusetts Raymond Vernon, Harvard University Janet Weiss, University of Michigan

THE POLITICS OF NUMBERS

INTRODUCTION WILLIAM ALONSO AND PAUL STARR

day, from the morning paper to the evening news, Ameri­ cans are served a steady diet of statistics. We are given the latest figures for consumer prices and the unemployment rate, lagging and leading economic indicators, reading scores, and life expectancies, not to m ention data on crime, divorce, and the money supply. Most of these numbers are official in the sense that they are produced by government in w hat are generally presumed to be impersonal and objec­ tive bureaucracies. Of course, in some countries, where the regimes are distrusted, official numbers are also routinely disbelieved. But where the statistical collecting and reporting agencies enjoy a reputation for professionalism (as they generally do in our society), their findings are commonly presented—and accepted—as neutral observations, like a w eatherm an's report on tem perature and atmospheric pressure. This view, we all know, is too simple. Official statistics do not merely hold a m irror to reality. They reflect presuppositions and theories about the nature of society. They are products of social, politi­ cal, and economic interests that are often in conflict w ith each other. And they are sensitive to methodological decisions made by complex organizations w ith lim ited resources. Moreover, official numbers, espe­ cially those that appear in series, often do not reflect all these factors in­ stantaneously: They echo their past as the surface of a landscape reflects its underlying geology. v e ry

E

1

THE POLITICS OF NUMBERS

Official statistics have always been subject to these influences, but more is now at stake. In the United States, an increased share of federal money is distributed to states and localities according to various statis­ tical formulae and criteria. The making of economic policy as well as private economic decisions hinges on fluctuations in key indicators. Standards for affirmative action in employment and school desegrega­ tion depend on official data on ethnic and racial composition. Several states now lim it their budgets to a fixed share of projected state income, and a proposed "balanced budget" constitutional amendment would do the same for the federal government, in effect incorporating the inexact science of economic measurement and forecasting into the Constitu­ tion. Official statistics directly affect the everyday lives of millions of Americans. They trigger cost-of-living adjustments of many wages and Social Security payments. They determine who qualifies as poor enough for food stamps, public housing programs, and welfare benefits. They are used to set the rates at which Medicare pays hospitals and to regu­ late businesses large and small. It is no wonder, then, that America has become a nation of statis­ tics watchers—from the congressmen concerned about redistricting to the elderly couples on Social Security worried about rising costs; from the bankers following changes in the money supply to the farmers watching the figures on cost-price "parity" for their crops. So well insti­ tutionalized are statistics such as the unemployment rate, the money supply, and various price indices that the date and even the hour of their release are regular events in the political and economic calendar, setting off debates on the performance of government policy and influencing both stock m arkets and elections. But official statistics also affect society in subtler ways. By the questions asked (and not asked), categories employed, statistical methods used, and tabulations published, the statistical systems change images, perceptions, aspirations. The Census Bureau's methods of clas­ sifying and measuring the size of population groups determine how many citizens w ill be counted as "Hispanic" or "N ative American." These decisions direct the flow of various federally mandated "prefer­ m ents," and they in turn spur various allegiances and antagonisms throughout the population. Such numbers shape society as they meas­ ure it. The absence of numbers may also be telling. For years after World War II Lebanon did not hold an official census, out of fear that a count of the tom country's Christians and M uslims might upset their fragile, negotiated sharing of power (which broke down anyway). Saudi Arabia's census has never been officially released, probably because of the Saudis' worry that publishing an exact count (showing their own popu2

Introduction lation to be sm aller than many supposed) m ight encourage enemies to invade the country or promote subversion. In Britain a few years ago, Scotland Yard created a furor when, for the first time, it broke down its statistics on crime according to race. Some Britons objected that the mere publication of the data was inflammatory. Statistics are lenses through which we form images of our society. During the early decades of the Republic, Americans saw the rapid growth in population and industry that the census recorded as a confirmation, for all the world to see, of the success of the American ex­ perim ent. The historian Frederick Jackson Turner announced his famous conclusion about the closing of the American frontier on the basis of an observation in the report of the 1890 census. Today, our national self-perceptions are regularly confirmed or challenged by statistics on such fundamental m atters as the condition of the nuclear family (allegedly still eroding), reading and literacy, the (slight) reversal of rural-to-urban migration, and our industrial produc­ tivity and m ilitary strength relative to other countries. W hether the meanings that politicians or pundits read into the data are reasonable or fanciful, the numbers provide a basis for popular and specialized discus­ sion. Even when the numbers m isrepresent reality, they coordinate our misperceptions of it. The process is thus recursive. W inston Churchill observed that first we shape our buildings and then they shape us. The same may be said of statistics. Lest there be any confusion, we should emphasize that to say official statistics are entangled in politics and social life is not to say that they are "politicized" in the sense of being corrupt. In some cir­ cumstances, they may indeed be corrupt, but that is not our point. Far from it: In the United States, institutional safeguards for the m ost part shield the statistical agencies from meddling by politicians and interest groups. These safeguards are a political fact in their own right and a foundation essential for public trust in the numbers. Our point, rather, is that political judgments are im plicit in the choice of what to mea­ sure, how to measure it, how often to measure it, and how to present and interpret the results. These choices become embedded in die sta­ tistical systems of the modem state and the information they routinely produce. The forces that shape those systems and their consequences for politics and society are the subject of this book.

Overview of the Book Just as the collection of statistics is an act of selection, so is the production of a collection about statistics. We have not tried to cover 3

THE POLITICS OP NUMBERS

all the kinds of statistics that governments gather or even the full range of analytical problems raised by the interplay of statistics and politics. In designing this collaborative project, we have brought together au­ thors from different fields—economics, history, politics, sociology, and planning—to w rite on topics that we thought would be interesting in their own right and of broad intellectual reach. Our aim was not to con­ tribute to statistical policy or methodology but to open up a field that scarcely exists—the political economy and sociology of statistics. We hoped the collection would be suggestive, w ithout pretending that it might be definitive. In the second part of this introductory section, one of us (Starr) at­ tem pts to outline the analytical problems and approaches in this area of inquiry and to review not only w hat we publish here but also some of the relevant literature in history and the social sciences. This is an ef­ fort to sort out the analytical issues in the sociology of statistics and to put them in intellectual context and perspective. A central tenet of this book, as we have already indicated, is that statistics cannot be constructed on purely technical grounds alone but require choices that ultim ately turn on considerations of purpose and policy. The point is well illustrated by the three chapters in Part I, The Politics of Economic Measurement. In the first, Raymond Vernon first looks at the competing views of statistics held by professional govern­ m ent statisticians, political leaders and policymakers, and academic so­ cial scientists. He then considers three cases—comparative figures on economic growth, productivity, and m ilitary expenditures—that illus­ trate the policy choices that inevitably m ust be made in constructing statistical information. Christopher Jencks examines the choices in the measurement of income, and asks why official statistics of family in­ come in the United States in the 1970s continued to be reported in a highly misleading fashion when the deficiencies were well known. Mark Perlman examines the development of the national income ac­ counts and finds that the policy interests of Keynesian economists were critical in shaping the structure of the accounts in the United States. In the United States as well as W estern Europe, the census has been the subject of particularly open and strenuous political conflict in the last decade. Part n , The Pohtics of Population Measurement, begins w ith a chapter by Margo Conk on the historical roots of the controver­ sies that erupted over the 1980 U.S. census. W illiam Petersen takes a broad look at the history and nature of disputes over the definition and measurement of ethnicity, and N athan Keyfitz examines the political and social aspects of the inexact art of population forecasting. The constitutional mandate for a census grew prim arily out of the need to apportion seats in the House of Representatives and the Elec-

4

Introduction toral College. The functions of government statistics have since ex­ panded, but they continue to be tied closely to the demands of demo­ cratic government. Part in, Statistics and Democratic Politics, deals w ith those connections. Kenneth Prew itt puts the problems in the con­ text of democratic theory and argues that, despite its lim itations, the nation's number system has become vital to pursuing two essential goals of a democratic polity: accountability and the representation of diverse interests. Steven Kelman takes up sim ilar themes in providing an explanation for an apparent conundrum: why the federal government in the nineteenth century produced elaborate statistics at a tim e when theories of m inim al government prevailed. Why did Americans make an exception of statistical inquiries, some of which involved intrusive questions by government about private activities? Kelman rejects an ex­ planation offered by microeconomic theory, w hich emphasizes the use of government to overcome problems of m arket failure, and cites histor­ ical evidence to argue that statistics were sought for their use in inform­ ing political debate, confirming national identity, and securing group recognition. In the final chapter in Part III, Abigail Them strom offers a somewhat darker view of the use of statistics. She also focuses on an issue relevant to democratic practice—the assurance of m inority repre­ sentation—and suggests that statistical tests have served as a form of camouflage for changing political objectives in the enforcement of vot­ ing rights over the past two decades. Among their many functions, statistics also m ediate the resolution of conflict. In Part IV, Statistics and American Federalism, three chapters deal w ith the interplay of statistics and the various levels of government in the United States. Richard N athan analyzes the "politics of printouts": the use of statistical formulae to distribute federal aid and the resulting political burden imposed on the statistics and statistical agencies. Judith de Neufville looks at the effect on local statistical prac­ tice of changing federal policy, particularly the shift from categorical programs to revenue-sharing and block grants, which create different demands for data. She emphasizes the difficulties of local governments in coping w ith statistical needs, in part because of local "dependency in statistical production" and federal cutbacks, but also because of the dis­ tinctive problems of statistical politics and policymaking at the local level. Looking in the opposite direction—that is, from center to periphery—Judith Gruber and Janet Weiss identify problems in national statistics that derive, in part, from the fragmentation of power among the states and localities. They argue that the effort to create a Common Core of Data for national education statistics failed because of the difficulties of overcoming the dispersed authority for schooling in the federal system and because of a lack of political consensus about how to

5

THE POLITICS OF NUMBERS

measure education. The discouraging result is a public statistical sys­ tem w ith lim ited relevance to the vital problems of public policy. New technological and political developments are greatly altering the mode of statistical production and distribution. In Part V, The N ew Political Economy of Statistics, Joseph W. Duncan looks at the implica­ tions of new computer technology and the changing costs and methods of producing, analyzing, and disseminating data. He defends the increas­ ing role of private industry, which the Reagan adm inistration now en­ courages as a means of cutting back federal statistical commitments. In the final chapter of the volume, Paul Starr and Ross Corson provide an analysis of the rise of the private statistical services industry and its re­ lation to government. They take a critical view of the privatization of public data, suggesting that it threatens some democratic political values of fundamental importance. The chapters in this volume were initially prepared for a conference on "The Political Economy of National Statistics," held in Washington on October 13-15, 1983. The conference was sponsored by the Social Science Research Council's Committee for Research on the 1980 Census. We wish to thank David L. Sills, who coordinated the project for the Council; and Kenneth Prewitt, then president of the Council, who took a particular interest in this venture. Ross Corson assisted Paul Starr as a research assistant and Jane Skanderup served as David Sills's assistant throughout. We are of course grateful to the contribu­ tors to the volume for their work and their patience.

6

THE SOCIOLOGY OF OFFICIAL STATISTICS PAUL STARR

scientists routinely use statistics from official sources as a means of analysis, but less frequently do they take the statistics or their sources as an object of analysis. If they do, they generally are interested in knowing w hether—or showing how—social or political processes have distorted the results and, therefore, why they may be safely ignored or, if possible, what may be done to adjust the data or to improve official statistical practice. N ot surprisingly, social scientists in the academy, business, and government approach statistics prim arily w ith their own practical interests in m i n d —methodology, commerce, and policy. Yet statistical systems are worth understanding as social phenom­ ena in their own right. Although the sociology of statistics is not an es­ tablished field of inquiry, much has been w ritten in scattered places about statistical institutions, statistical policy and politics, the social and cultural history of statistics, the uses of statistics in organizations, and other related topics. From these diverse sources emerge general so­ ciological questions about statistics. In this introductory essay I want to set out some of those questions, the approaches taken to them, and the OCIAL

S

7

THE POLITICS OF NUMBERS

arguments that seem to me m ost persuasive and useful in understand­ ing statistics as a social phenomenon.1 The concept of "statistical system " lies at the center of this in­ quiry. By a statistical system I mean a system for the production, distri­ bution, and use of numerical information. A statistical system may be said to have two kinds of structure—social and cognitive. Its social or­ ganization consists of the social and economic relations of individual respondents, state agencies, private firms, professions, international organizations, and others involved in producing flows of data from original sources to points of analysis, distribution, and use. Cognitive organization refers to the structuring of the information itself, includ­ ing the boundaries of inquiry, presuppositions about social reality, systems of classification, methods of measurement, and official rules for interpreting and presenting data. About any information system—w hether for record-keeping, sur­ veillance, news, or statistics—it is possible to raise several kinds of so­ ciological questions, depending on w hether the object is to explain (or derive consequences from) the origin or pattem of development of the system, its social or cognitive structure, its uses and effects, or its direc­ tion of change.2 On this basis, it seems convenient to divide the sociol­ ogy of official statistics into five overlapping sets of questions: 1The reader should be alerted to three restrictions in the scope of this analysis. First, it is primarily concerned with descriptive statistics that measure social phenomena rather than with statistical data or techniques that measure the effects of different treatments or agents. In the latter category belong evaluation research and controlled social experimen­ tation. The politics of program evaluation, risk assessment, and cost-benefit analysis do not receive sustained attention here. Second, I do not deal with the development of statis­ tical methods. Although the term "statistics" refers both to numerical information and to a field of knowledge (die subject of statistics as a discipline), this analysis concerns statis­ tics only in the first sense. [For historical and sociological work on the development of sta­ tistical methods, see Donald A. MacKenzie, Statistics in Britain, 1865-1930: The Social Construction of Scientific Knowledge (Edinburgh: Edinburgh University Press, 1981); Paul F. Lazarsfeld, "Notes on the History of Quantification in Sociology—Trends, Sources and Problems," in Harry Woolf, ed., Quantification: A History of the Meaning of Measure­ ment in the Natural and Social Sciences (Indianapolis: Bobbs-Merrill, 1961), pp. 147-203; and Lorraine Dastin, Michael Heidelberger, and Lorenz Krüge, eds.. The Probabilistic Rev­ olution, 1800-1930 (Boston: Bradford Books, 1987).] Third, this analysis deals with official statistics, by which I mean statistics that governments produce, finance, or routinely in­ corporate into their decisions. Although the term "official" is sometimes used to describe any institutionally certified form (professional sports have their "official statistics"), the state is the main pathway to the distinctive cultural authority that official facts character­ istically assume. Part of my interest in this topic derives from my earlier work on cultural authority. See The Social Transformation of American Medicine (New York: Basic Books, 1982), esp. chap. 1. 2In developing these categories, I benefit from James B. Rule's work on surveillance systems. See his Private Lives and Pubhc Surveillance: Social Control in the Computer Age (New York: Schocken, 1974), esp. chap. 1 .1 owe a more general debt here to a classic essay that helped to organize the sociology of knowledge: Robert K. Merton, "Paradigm for the Sociology of Knowledge" [1945], reprinted in Merton, The Sociology of Science: Theoretical and Empirical Investigations (Chicago: University of Chicago Press, 1973), pp. 7-40.

8

The Sociology of Official Statistics 1. System origins and development: What causes statistical sys­ tem s to be established? Why do governments or other actors make a decision to count or to allow themselves to be counted? And w hat processes govern the system s' general historical evo­ lution? 2. Social organization of statistical systems: W hat are the sources and consequences of different designs or patterns in the social organization of statistical systems? 3. Cognitive organization of statistical systems: What are the sources and consequences of elem ents in the informational structure? 4. System uses and effects: What effects do the production and distribution of statistical information have on politics and so­ ciety? Do statistical systems shape understanding of social and economic reality so that effects are due, not to the phenomena measured, but to the system measuring it? 5. Contemporary system change: W hat processes, such as politi­ cal changes or technological innovations, are now shaping the current and future development of statistical systems? Like the first group, these questions are historical and developmental, but they concern the future rather than the past. The last of these sets of questions is partially dealt w ith in the final two chapters in this volume. Here I take up the first four problem areas and try to outline the sorts of questions raised and approaches taken for each of them.

I. Statistical Systems in the Making: Origins and Development Social and economic statistics, Otis Dudley Duncan has recently observed, share common features w ith other forms of m easurem ent.3 Like standard physical weights and measures, as well as measures of tim e and even common units of currency, they serve an interest in so­ cial coordination and control. In their origins and operation, they also have a specific relation to the state. O ther institutions or groups often do not command—and have no reason to invest—the financial resources, organization, and authority required to carry out censuses and large-scale surveys. However, precisely because of the state's power, its inquiries are often especially feared and resisted as invasions of privacy and because of their possible use for assessing taxes or catching suspected criminals and subversives. To the use of statistics for domes3Otis Dudley Duncan, Notes on Social Measurement: Historical and Critical (Nevy York: Russell Sage Foundation, 1984), pp. 12-38.

9

THE POLITICS OF NUMBERS

tic intelligence have been added other complex political functions in the allocation of money and power, the setting of norms, the making of policy, and the evaluation of government performance. All these func­ tions have increased the stakes in official statistics and underlined the importance of understanding their origin and development. Explaining why governments produce social and economic statis­ tics poses two separate tasks. The first is to explain the foundations of measurement: why governments establish the framework for a statisti­ cal system and w hat preconditions allow them to do so. The second is to explain specific decisions to introduce (or eliminate) a statistical series or undertake a particular survey once an intellectual and bureau­ cratic framework already exists. The first is a problem in institutional innovation; the second, in policy initiation (or termination). The inau­ guration of a population census or the advent of labor statistics is an example of the first; the introduction of a statistical series on job vacan­ cies is an instance of the second. I shall concentrate here on the founda­ tional problem.

Social Foundations of Official Statistics One difficulty in explaining the foundations of official statistics is that they have a somewhat confusing line of historical descent. Censustaking originated as a means, not of gathering quantitative data, but of surveillance, conscription, and tax assessment. N either did the term statistics originally have its present meaning. When first used in Ger­ man and English, “statistics" m eant facts about states and did not specifically refer to numerical data, m uch less to an abstract set of analytical principles. But although the earliest censuses and statistical studies do not correspond to modem social forms, the connections as well as the discontinuities between them are helpful in understanding the roots of modem statistical systems.

Censuses, Ancient and Modem Censuses of population can be traced to ancient societies, but from ancient Greece and Rome to early modem societies a census was a registration of people and property. The term census comes from Latin and in Rome referred to "a register of adult male citizens and their prop­ erty for purposes of taxation, the distribution of m ilitary obligations and the determ ination of political status."4 The Roman censor was also charged w ith the control of manners; hence the etymological associa4Waiter F. Willcox, "Census," Encyclopaedia of the Social Sciences (New York: Mac­ millan, 1930), vol. 2, pp. 295-300.

10

The Sociology of Official Statistics tion w ith our term s censor and censorship. As late as the sixteenth century, Jean Bodin's classic work on statecraft, The Six Bookes of a Commonweale, described a census as “nothing else but a valuation of every mans goods," useful not only for levying taxes but also for deter­ mining the number and age of m en available for war and public works, the needs of cities in the event of a siege, the resolution of civil disputes, and "the discovery of every mans estate and faculty, and whereby he gets his living, therby to expell all drones out of a common­ weale . . . and to banish vagabonds, idle persons, theeves, cooseners, & ruffians, which live and converse among good men, as woolves do among the sheepe. . . ."5 A census at this stage was unambiguously an instrum ent of state power and social control. A modem population census differs in several crucial ways.6 First, a modem census is an enumeration of an entire population in a nation or one of its geographical subdivisions. Premodem censuses, according to Willcox, were often lim ited to males, particular age groups and classes, or hearths and households, and thus were less inclusive.7 Second, a modem census records and counts individuals and provides data at the individual level. As late as the early nineteenth century, the household typically continued to be the unit of enumeration; m inim al informa­ tion, if any, was recorded about individuals other than the head. Third, a modem census is an enum eration at a fixed time, whereas the pre­ modem form was often a continuous register. Fourth, statistical data from modem censuses are typically expected to be published; premod­ em censuses were generally state secrets. Fifth, and perhaps most significant, nations today generally distinguish statistical agencies from those charged w ith tax assessment and law enforcement; no such dis­ tinction was made in the premodem form. The modem census has as its primary and manifest function the production of quantitative infor­ mation, whereas the premodem form was used explicitly for keeping people under surveillance and control. The primacy of statistical functions and their bureaucratic segrega­ tion from police and tax collection influence, or are m eant to influence, the social relation between census-takers and respondents. Although some respondents are suspicious of census-takers, many share the understanding that the purpose of an official census is to generate infor­ m ation about aggregates of people, not about individuals. Reports in the 5Jean Bodin, The Six Bookes of a Commonweale [1576], ed. Kenneth Douglas McRae (Cambridge, Mass.: Harvard University Press, 1962), pp. 637,641. The (Wilson), 209 deficit financing, 141 democratic government, 261-391; links to statistics of, 18-20. See also apportionment; democratic theory; federal government role in statistics, justifications of; Voting Rights Act of 1965 democratic pluralism, 268-69

464

democratic theory: accountability, 262r-67; representation of diverse interests, 267-74 demographers, 245-47, 253-54, 258 demographic aggregates, representation based on, 273 demographic changes, 184-85 demographics industry, 425-29 dependency in statistical production, 28, 351 Depression, the, 141, 168, 169, 172, 173, 184 desegregation: Civil Rights Act of 1964 and, 200; measurement of, 387-88 Detroit, Michigan, 182 developing countries, comparison of, 64—72 diglossia, 205, 206 Dillon's rule, 334-35 direct mail, 427 disclosure, 27, 32-33, 38-39. See also data reporting discrimination, 229, 230; against blacks, 195, 199-202; in education, 200, 201; electoral, 304,306-10; in employment, 199-200, 202-203; institutional, 270; principle of cumulation and, 195; reverse, 200 disfranchisement, 304-307,309,323-27 disinterested forecasts, 252-53 disposable income, 106, 107 dissemination of data. See data dissemination division of labor. See labor, division of divorce rates, 49 domain definition, 44 double-entry bookkeeping, 21, 146-48 dual formula for community development block grants, 340-41 Dual Integrated Map Encoding (DIME), 426, 434 Dutch heritage, 219

E Ebony, 212n economic conditions, voting behavior and, 263-64 economic forecasting, 28, 48-49, 422, 424 economic growth, 48-49, 53, 84, 136-37, 149, 287-88 economic indicators, 25,49-52, 54,264,266

Subject Index economic measurement, politics of. See comparative economic statistics; national income accounting; real family income (RFI) series economic output. See economic productivity economic policy, 2, 24-25, 42-43, 138 economic productivity: data manipulation and, 71-72; national income accounting and, 137-38; poor country identification and, 66-70; U.S., comparative, 72-77 economic stabilization, 42, 138 economic theory: Keynesian, 25, 42, 140, 142, 145-46, 150; macroeconomic, 150, 151; microeconomic, 276-78, 293-99 economic well-being, alternative measures of, 85-112; current-constant dollars conversion, 88-92; family definition, 92-105; income definition and measurement, 105-10; medians vs. means, 86-88 education: discrimination in, 200 , 201 ; population forecasting and, 257-58 education statistics, federal, 28, 363-91; federal level fragmentation of control and, 376-80; historical development of, 366-67; intergovernmental fragmentation of control and, 372-76; lack of consensus on measurement and, 382-89; pressures toward irrelevance of, 389-91; public-private sector fragmentation of control and, 380-82 electoral districts, 311-19 electronic data delivery, 403-405, 408-409 Elementary and Secondary Educational General Information Survey (ELSEGIS), 367,368 employment: affirmative action in, 200 , 202-203, 230, 272; discrimination in, 199-200 Employment Act of 1946, 125 energy forecasts, 248-50 energy policy, 410 Engel's Law, 99-101 English heritage, 219, 220 ENIAC (Electronic Numerical Integrator and Computer), 396 errors: accountability and, 266; in comparative economic statistics, 71, 72, 74, 78, 81; in data processing, 37-38; impact on local policies, 352; in

population forecasts, 242-44, 254-56; in surveys, 31; tolerated, 38 Eskimos, 216, 231 Espanol para los Hispanos (Baker), 227 estimates, 241, 242. See also population forecasting ethnic categories vs. ethnic groups, 205-208 ethnic classification, 44, 187-233; American ethnicity, special characteristics of, 191-204; American Indians, 189, 193, 201, 212-18; blacks, 190, 193-95, 201, 208-12, 231; conflicts with democratic government, 271-73; electoral distribution and, 314-18; ethnic group vs. ethnic category, 205-208; Europeans, 190, 218-23; factors in, 188-91; vs. hard data classification, 187-88; Hispanic, 45, 207, 223-29, 231-32 ethnic competition, 199-200 ethnic groups, 45; vs. ethnic categories, 205-208. See also ethnic classification ethnicity: shift from race to, 201-204; vs. race, 190 ethnic polarization, 318-19 European immigrants, 164; assimilation of, 194-95; bilingualism of, 206; census classification of, 190, 218-23; cultural pluralism and, 196-97 exchange-rate comparisons, 69, 70, 75

F FAIR v. Klutznick (1980), 180n, 18 In family definition, 92-105 family equivalence scales, 96-104 family income, real. See real family income (RFI) series federal government: democratic politics [see apportionment; democratic theory; federal government role in statistics, justifications of; Voting Rights Act of 1965); federalism (see education statistics, federal; grants-in-aid; local government, federal statistics in); history of the census and, 158-74; information services of (see federal government information products and services); "inkind" services of, 107, 108, 138 federal government information products

465

THE POLITICS OF NUMBERS

and services: data comparability, 407- 408; data quality, 406-407; interaction with private sector, 406, 409-13, 416,434-47; local access to census and, 357-59; policy issues, 408- 409; pricing of, 440-44; user fees, 358, 405-406, 439 federal government role in statistics, justifications of, 275-302; current, 299-302; legislation-aiding function, 19, 280-86, 295-96,302; microeconomic theory and, 276-78, 293-99; patrioticpride function, 19-20, 286-88, 297, 301; private decision-aiding function, 292-93, 295,301-302; social-recognition function, 19, 271, 288-90, 297; specialvalue-of-knowledge function, 290-92, 297 federalism. See education statistics, federal; grants-in-aid; local government, federal statistics in Federalist Papers, The (Madison), 18,33, 263, 268 Fifteenth Amendment, 199, 303,304, 305, 307,323,324 Filipino-Americans, 325 Filipinos, 225 Finnish immigrants, 219 "first statistical era," 23-24 fixed weight price index, 90n Florida, 324 folded-in programs, 338 food budgets, USDA, 97, 98, 119 food consumption, 98, 101, 120, 123-24 food equivalence scale, 97, 98 food expenditures, 97-102, 119, 120, 121-24 Food Stamps, 109 Ford administration, 130,340 forecasting: economic, 28, 48-49, 422, 424; population (see population forecasting) foreign heritage, 190, 191 FOSDIC (Film Optical Sensing Device for Input to Computer), 173, 397 Fourteenth Amendment, 174,199, 228,324 Framework for Planning U.S. Statistics for the 1980's (Duncan and Shelton), 134n France, 12; classification in, 53; emergence of statistical inquiry in, 17, 22 ; output per worker in, 74 Freedom of Information Act, 222 French language, 224

466

6 General Agreement on Tariffs and Trade (GATT), 65 Generalized System of Preferences, 65 General Theory of Employment, Interest, and Money, The (Keynes), 140 geodemographics industry, 426-29 geographical groupings, 45 German heritage, 219-20 German immigrants, 164, 194, 218, 219 German language, 224 Germany : classification in, 53-54; emergence of statistical inquiry in, 22 ; university statistics in, 13. See also West Germany Ghost Dance movement, 215 government: manipulation of statistics by, 38-39, 71-72, 265-66. See also federal government; local government, federal statistics in Gramm-Rudman plan, 55 grants-in-aid, 185,331-42; automatic formulas in, 55-57; categorical, 343-48; history of, 158, 168-71; shift to broadening of, 331-32; targeting and, 334; technical challenges to administration of, 332—33; undercounts and, 232. See also block grants; revenue­ sharing Great Britain: emergence of statistical inquiry in, 22 ; political arithmetic in, 14—15; U.S. immigration quotas for, 164 Green Party, 33 gross national product (GNP), 49, 51; in comparative economic statistics, 62-63, 67-70, 72-75, 78; in national income accounting, 143, 145, 147 group-based representation, shift toward, 273 group classification, 272. See also ethnic classification group identity, 19, 271,288-90, 297. See also ethnic classification grouping, 44-45 Guarani, 205

H hard disks, 403 Harlem, New York City, 209, 317n

Subject Index Havana Charter for an International Trade Organization, 65 Hawaiians, 189-90, 231 Health and Nutrition Examination Survey (HANES), 124 health statistics, 30-31, 112-15 Hebrew heritage, 219 Hebrew immigrants, 219 Hidalgo County, Texas, 312 Hispanics: census classification of, 45, 207, 223-29, 231-32; in census of 1970, 158; population growth of, 184; selfidentification of, 207,314—15, 321; undercounts of, 176-77, 179, 180, 188, 223; use of national statistical system by, 271; Voting Rights Act of 1965 and, 310-17, 323-27 Hispanos, 223 hold-harmless provision, 338 holography, 403 homeless population, 272 home ownership, 88-89 Housing Assistance Plan (HAP), 346-47 housing expenditure trends, 115-19 How to Pay for the War (Keynes), 142 Hyde County, North Carolina, 304-^305

I identity cards, 401 ideology, 25; population forecasting and, 250-52; real family income (RFI) series and, 129-30 illegal aliens, 180-81, 225-26, 272 Illinois, 157, 173, 203 illiteracy, 305, 324, 325, 370 immigrants, 104, 162-64, 167. See also European immigrants immigration, 272, 284; policy, 220; restriction of, 162, 164, 194-95, 203-204, 220-21, 229 Immigration Act of 1924, 220 impartiality, 258 implicit price deflator, 90 income: national [see national income accounting); per capita, 93-96, 251; per family [see real family income (RFI) series) income accounting. See national income accounting income distribution: automatic formulas

in, 55-57; national income accounting and, 135-36. See also grants-in-aid income measurement, politics of. See real family income (RFI) series independent agencies, 26 indexing, 55-56 India, 69 Indians. See American Indians individual noncompliance, 32 infant mortality, 50, 112-14 inflation, 53, 56, 84, 142; consumer price index exaggeration of, 128; cost of living adjustments and, 55; current-constant dollar conversion and, 88-92; mean vs. median measures of, 87 information-gathering, justifications of. See federal government role in statistics, justifications of information products and services. See federal government information products and services; statistical services industry information technology. See computers; technological change "in-kind" income, 105, 107-12 innovation. See technological change inquiry: framing of, 41-43; periodicity of, 49-50 "instant Indians," 214, 216-17, 230 institutional racism, 270 integrated statistical services firms, 429-33 intelligence function, 10, 12, 13 interagency data sharing, 407-408 intergovernmental fragmentation of education control, 372-76 International Comparison Project (ICP), 69-71, 74 Ireland, 164 Irish heritage, 219 Irish immigrants, 202, 222 Italian language, 224

J Japan, 67, 74-77 Japanese-Americans, 231, 325 Jewish immigrants, 221 Jews, 42, 44, 202, 207 job classifications, 46 Johnson administration, 99 Jones v. City of Lubbock (1984), 319n journalistic use of forecasts, 245

467

THE POLITICS OF NUMBERS

K

M

Kentucky, 161 Keynesian economics, 25, 42, 140, 142, 145-46, 150 knowledge-as-special-value function of statistics, 290-92, 297 Korea, 65 Koreans, 316, 325

macroeconomic theory, 150, 151 Madisonian pluralism, 268 mail census, 34, 172, 174-76 management information systems, 421, 422, 424-25 manufacturing census of 1810, 281 manufacturing census of 1850, 289 marketing, demographic data in, 427, 429 Maryland, 161 Massachusetts, 161 mass data storage, 403, 418 mean vs. median family income, 86-88 measurement, 46-49 median vs. mean family income, 86-88 medical expenditures trends, 112-15 melting-pot era, 194—96, 198 metropolitan statistical areas (MSAs), 45, 339, 340 Mexican-Americans: census classification of, 45, 207, 223, 225-29; Voting Rights Act of 1965 and, 312-13, 323-26 Mexican heritage, 228 microcomputers, 395, 402, 403 microeconomic theory, 276-78, 293-99 microfiche, 358 migration, 170, 184, 272 military expenditures: as economic output or economic input, 67-70; U.S. vs. U.S.S.R., 77-81 minicomputers, 395, 403 minimum competency tests, 384 minority groups: assimilation of, 194-97; ethnic classification of [see ethnic classification); language, 324, 325; representation of [see apportionment; Voting Rights Act of 1965); undercounts of, 155, 158, 174-79, 182, 188, 216, 223, 232-33,349. See also names of minority groups minority officeholding, 304,308, 309,312, 313,320, 321 misreporting: by government, 38-39, 71-72, 265-66; by respondents, 34-35 Mississippi, 323 Missouri, 428 "modem statistical era," 24 money income, reported, 105-107 "Money Income of Households, Families, and Persons in the United States"

L labeling, 45 labor, division of: in population forecasting, 29, 238, 253-54; in statistical systems, 25-29 labor productivity, U.S., 73, 76 labor statistics, 10, 24, 28 land grants, 168 Language and Nationality in the Compilation of Demographic Data (Strassoldo), 469 language minorities, 324,325 languages: foreign, 205-206, 224; technical vs. everyday, 45 laser cards, 402 laser video disks, 403 legislation-aiding function of statistics, 19, 280-86, 295-96, 302 legitimacy of statistics, 353-57 Letters from an American Farmer (de Crèvecœur), 191 liberal democratic government, 19 libraries, 358, 359, 412, 438 life expectancy, 112, 113 literacy tests, 303, 306-10, 323-26 Lithuanian immigrants, 219 local government, federal statistics in, 27-28, 343-62; access to census, 357-60; demand for accuracy, 351-53; education and: [see education statistics, federal); future, 360; grants-in-aid [see grants-inaid); legitimization of, 353-57; local dependencies on, 351 local policymaking and federal statistics, 350-60 local vs. national division of statistical labor, 27-28 Louisiana, 316n Loving v. Virginia (1966), 199 lumpiness of grants, 338

468

Subject Index (Census Bureau), 86 , 91, 93, 94, 105, 106, 109, 111 mortality statistics, 165, 286, 293 mortgage payments, 88-89 mulattoes, 208-209

N national economic statistics, comparative. See comparative economic statistics national income accounting, 24, 133-51; changes in objectives of, 138-51; conceptual ambiguities in reporting, 51-52; data-theory linkage and, 42-43; measurement rules and, 48; origins of, 135-38 National Income and Its Composition (Kuznets), 150 National Income and Products Account (NIPA) series, 172 national income comparisons, 69-70 National Income in the United States 1929-35 (1936), 139 National Income in the United States 1929-36 (1937), 139 national income redistribution, 169-70 National Origins Act, 162, 164 national vs. local division of statistical labor, 27-28 négritude, 198 "negro" designation, 211 Negro Digest, 211, 212n Neshoba County, Mississippi, 310 neutrality, 50-51 New Deal, 141, 169, 170, 172, 199, 212-13 New England, 162 New Federalism, 332,337 New Hampshire, 161 New Immigration, 202 New Jersey, 195-96, 325 New Mexico, 223, 324 New York City, 176, 179,333, 349; Voting Rights Act of 1965 and, 305, 309-11, 313-22 New York State, 176,182, 184, 325 NICs (newly industrialized countries), 66 Nigeria, 239 Nixon administration, 130,307 noncompliance, 32-33 nonparticipation, 269 "non-whites" category, 231

norms, 54 North Carolina, 220, 304-305,316n, 319 numeracy, 2 0-22 Nuremberg, 12

o observations vs. data, 42 occupational classification, 46 officeholding, minority, 304,308,309,312, 313, 320, 321 one man-one vote principle, 157 one-seventh rule, 104 online systems, 418-20 On the Accuracy of Economic Observations (Morgenstern), 34 operating agency statistical branches, 26 optical memory media, 403 optical scanning techniques, 402 ordering, 46 organizational structure, politics of, 25-29 Other America, The (Harrington), 270-71 output. See economic productivity overcounts, 163 overproduction, 141

P Pakistan, 69 Paraguay, 205 patriotic-pride function of statistics, 19-20, 286-88, 297, 301 Pennsylvania, 219, 220, 325 People's Republic of China, 13,38, 254-55 per capita GNP, 67-70, 74 per capita income, 93-96, 251 periodicity of inquiry, 49-50 perpendicular magnetic recording, 402-403 Personal Consumption Expenditures (PCE) deflator, 87, 90, 91, 128 phenomenological approach, 35-37 Phoenix, Arizona, 226n Plains Indians, 215 planners, 245-46 planning, 24-25, 243-45 pluralism, 28; cultural, 197-98, 203; democratic, 268-69; Madisonian, 268 policymakers, comparative economic statistics and, 62-63, 81 Polish-Americans, 203

469

THE POLITICS OF NUMBERS

Polish immigrants, 219 Polish language, 224 political arithmetic, 14-16, 20 Political Arithmetick (Petty), 14 political campaigning, 429 political organizations, use of demographic industry by, 427-28 politicians, comparative economic statistics and, 62-63, 81 politics: of economic forecasting, 48-49; of economic measurement (see comparative economic statistics; income measurement; national income accounting); federalism and (see educational statistics, federal; grants-inaid; local government, federal statistics in); of organizational structure, 25-29; political economy of statistics (see statistical services industry; technological change); of population measurement (see census of 1980; ethnic classification; population forecasting); statistical rules and, 56-57; statistics and democratic government (see democratic government) poor countries: census-taking by, 239; identification of, 64-72 population: demographic changes in, 93-94; politics of measurement of (see census of 1980; ethnic classification; population forecasting) population censuses. See censuses population density maps, 162 population expansion, 166-67 population forecasting (estimates, projections), 235-58; conflict with policy, 254—58; disinterested, 252-53; division of labor in, 29, 238, 253-54; ideological element in data presentation, 250-52; impossibility of vs. need for, 236-^37; journalistic use of, 245; long­ term nature of, 240-41; as persuasion, 247-50; in planning, 243-45; professional caution with, 245-47; supporting data, 238-40; terminology, 241-43 population growth, 161-63,166-67, 184-85, 240-41, 250-51. See also population forecasting population maps, 162 population movements, 167 population policy, 254—58

470

population projections. See population forecasting poverty ratios, 96-97, 99, 102 poverty thresholds, 96-99,102 preclearance provision, Voting Rights Act of 1965,306,308-14,322, 326 prejudice, 195 President's Private Sector Survey on Cost Control in the Federal Government, 399-400 price ceilings, 71, 72 price deflators, 75 price indexes, 90; in comparative economic statistics, 75; consumer (see consumer price index [CPI]); food, 121-23; government manipulation of, 71, 72 prices: in comparative economic statistics, 68-69, 79-80; of federal government products and services, 440-44 price surveys, 401 Principles of Economics (Marshall), 146 printing, private sector, 409-10 printouts, politics of, 331-42 privacy, 13, 32m33, 175 private decision-aiding function of statistics, 292-93, 295, 301-302 private sector: vs. public sector division of statistical labor, 28-29; vs. public sector fragmentation of education control, 380-82; statistical services (see statistical services industry) Problems of a Changing Population, The (Nutritional Resources Committee), 170 procurement process, federal, 398-400 productivity. See economic productivity professionals, 266, 345, 346; in comparative economic statistics, 61-62, 81-82; in population forecasting, 245-47, 253, 258 projection pursuit, 251-52 projections, 48-49; vs. forecasting, 241-43. See also population forecasting proportionality, statistical, 272 Protestantism, 22, 222 protostatistical era, 23 Prussia, 16 public acceptance of statistics, 353-57 publications, federal government, 435,437, 440 public data: accesrf to, 357-59,441-42, 444-47,- social distribution of, 38-39 public education. See education; education statistics, federal

Subject Index public goods theory, 277-80, 294-95 public policy norms, 54 public sector: vs. private sector division of labor, 28-29; vs. private sector fragmentation of education control, 380-82 publishing, 430 Puerto Ricans, 45, 189, 207, 226, 227, 310, 321, 325 pump-priming, 141 punch card tabulation, 396 purchasing power, 69-70, 141

Q quantification, 20-23 Quon v. Stans (1970), 176n quotas, 164, 203-204, 230, 273

R race: vs. ethnicity, 190; immigration and, 164; shift to ethnicity from, 201-204; undercounts and, 174-79, 182 racial classification, 271-72. See also ethnic classification racial discrimination. See discrimination racial groupings, 45 racial polarization, 318-22 racial separation, 320 racism, 201,270,311, 319,349 railroads, 284 reading achievement tests, 388 Reagan administration, 80, 299, 390, 409, 417, 435, 437 real family income (RFI) series, 83-131; Census Bureau publicizing of, reasons for, 126-31; current-constant dollars conversion and, 88-92; family definition and, 92-105; food expenditure trends and, 119-24; housing expenditure trends and, 115-19; improvement of, 111-12; income definition and measurement and, 105-10; medians vs. means and, 86-88; medical expenditure trends and, 112-15; problems with, 110-11, 125-26; welfare and expenditures and, 124-25 real per capita income (RPCI) series, 95 reapportionment, 157, 160, 163-65, 173-74, 184 reference data bases, 418 Reflections of America: Commemorating

the Statistical Abstract Centennial (Commerce Department), 134n religious affiliation, 41-42, 44, 221-23 religious institutions, 428 removable hard disks, 403 repackaging industry, 423 "Report on Indians Taxed and Indians Not Taxed in the United States," 214 representation, 18, 267-74. See also apportionment; Voting Rights Act of 1965 representative government. See democratic government "Requiem for USAC, A" (Kraemer), 359n resource allocation, 272 respondent compliance, 32-34 retrospective evaluations by voters, 263 revenue-sharing, 343; allocation formula for, 334-37; legislation, 157-58; undercounts and, 177 reverse discrimination, 200 Revolution in United States Government Statistics, 1926-1976, 134n Roman Catholic Church, 180 Roman Catholics, 221, 222 Rome, ancient, 10-11, 17 Roosevelt administration, 169, 170, 172, 175 rural areas, 163, 167, 170 Russian immigrants, 218-19

S sampling, 31, 63, 173 San Antonio, Texas, 323 sanskritization, 189 Scandinavian immigrants, 219 school districts, 367,374, 381 schools: performance of, 375,383-87. See also education; education statistics, federal Scottish heritage, 219 Seasonal Variations in Industry and Trade (Kuznets), 136 secondary publishing, 423 "second statistical era," 24 segmented marketing, 427 self-identification, ethnic, 204, 206,208, 233; American Indian, 216-18; Hispanic, 207,314-15,321 self-interest, 297-98 "701" planning grants, 345

471

THE POLITICS OF NUMBERS

site location, 426 Six Books of a Commonweale (Bodin), 11 Sixteenth Amendment, 33 slavery, 161-62, 192-93, 199, 201, 208, 209, 267 Slavic immigrants, 203, 218-19 social accounts, systematizing of, 146-48 social class, classification and, 190 social classification systems, 43—46 social distribution of public data, 38-39 Social Forces, 201 social foundations of official statistics, 10-23 Social Framework, The (Hicks), 146 Social Framework of the American Economy, The (Hicks), 146 social indicators, 25, 49-50, 54, 264, 266 social-recognition function of statistics, 19, 271, 288-90, 297 social reform, 269-71. See also civil rights social relations: of data collection, 34-37; of data processing, 37-38 Social Security Act of 1935, 140, 169, 257 Social Security benefit indexing, 55, 56 Social Security taxes, 107 sociology of official statistics, 7-57. See also statistical systems source data bases, 419 South, the, 156, 161, 174, 183, 184, 189, 192, 306-307,309, 310, 323,327, 340 South Carolina, 220, 319 South Dakota, 324 Southwest, 223, 225, 228, 325,326 Soviet Union. See Union of Soviet Socialist Republics Spain, 16, 17 Spanish heritage, 224-27, 325 Spanish language, 224, 227-28 Spanish surnames, 224—25,325 Sri Lanka, 69 Staatenkunde, 13 standardization of data, 373-74 standard metropolitan statistical areas, 339, 340 standard of living, 84, 87; food expenditure trends, 119-24; housing expenditure trends, 115-19; impact of children on, 96-104; inflation and, 91; medical expenditure trends, 112-15 state: early concepts of statistics and, 13-15; intelligence function of census

472

and, 13; role in collecting statistics, 9 10, 15-17; statistics and intervention by, 2425. See also federal government State and Local Finance in the National Economy (Hansen and Perloff), 170-71 state building, 15-17 state governments. See education statistics, federal; grants-in-aid Statistical Abstract of the United States, 84, 248-49, 423 Statistical Account of Scotland (Sinclair), 15 statistical agencies, 26, 28,37, 81, 242, 265, 266, 372, 440 statistical blackouts, 38-39 statistical eras, 23-24 statistical policy, political foundations of. See federal government role in statistics, justifications of statistical process, sociology of, 25, 29-39 statistical services industry, 404, 415—47; adding value to public data function, 423-25; data bases, 417-21, 430-31; demographic analysis, 425-29; integration of services in, 429-33; interaction with federal government products and services, 406, 409-13,416, 434-47; local access to census data, 357-60; sources of growth of, 421-23; technological change and, 416-17, 421, 429 statistical systems, 7-57; cognitive organization of, 8, 40-52; defined, 8; development of, 9-25; effects of, 52-57; politics of organizational structure of, 25- 29; social foundations of official statistics, 10-23; sociology of statistical process, 25, 29-39 statisticians, 61-62, 81-82, 245-46, 253, 258, 265-66 statistics cutbacks, 299-302 “statistics strikes,“ 32 strategic data reporting, 35-35 student achievement, 385-89 subjective orientation of interviewers, 35-37 suicide statistics, 35-36 Sumter County, South Carolina, 316 surveillance, 10, 12, 13 Survey of Current Business, 139n, 145, 146

Subject Index Survey of Income and Program Participation, 92 surveys, 30-^32 Sweden, 12 Swiss immigrants, 218 Syria, 67

T Tabellenstatistik, 13 tabulation software, 357, 358 T-accounts, 147-48 tapes, computer, 357,358, 401, 404 tax assessment, 10-11, 16, 18,33 tax-effect factor, 335 taxes: income measurement and, 107-108; revenue-sharing formula and, 335-36 teachers, 257, 369-71, 380-81, 385 technological change, 395-413; context of, 400; in data collection, 401-402; in data dissemination, 402-405; growth of, 396-97; lag in, 397-99; policy issues and, 405-13; recent evaluations of, 399-400; statistical services industry and, 416-17, 421, 429. See also computers telecommunications, 395, 401, 404, 417, 431, 433 telephone interviewing, 401 Tennessee, 157, 161, 174 Texas, 157, 323, 324 text retrieval, 418-19 theory, linkage between data and, 42-43 Thirteenth Amendment, 199 ''Three Black Histories" (Sowell), 208 time-sharing, computer, 404, 418, 421, 431, 433 TPL (Table Producing Language), 397 two-thirds rule, 101-104

u undercounts, 33; of American Indians, 216; apportionment and, 174-84; of blacks, 158, 175, 176, 188; block grants and, 349; census of 1920,163; census of 1970,158; census of 1980,31, 155, 156, 176-84; court decisions and, 182-84; grants-inaid and, 232-33; of Hispanics, 176-77, 179, 180, 188, 223; revenue-sharing and, 177; of rural areas, 163

underreporting of income, 105-106, 108-11 undocumented workers, 272 unemployment, 128-30, 141, 258; black, 209-10; grants for, 168-69; short-run changes vs. long-run trends in, 125-26; statistics, 30-31, 53 Union of Soviet Socialist Republics: central planning and statistics in, 24-25; intelligence function of censuses in, 13; military expenditures in, comparative, 77-81; statistical blackouts in, 38-39 United Kingdom: output per worker in, 74. See also Great Britain United States: dissemination of Soviet statistics by, 39; first use of census in, 12; military expenditures in, comparative, 77-81; organizational structure of statistics in, 27-28; productivity of, comparative, 72-77; protests against undercounts in, 33. See also federal government UNIVAC I, 396-97 universities, 29 university statistics, 13 unobserved economy, 266 unreported income, 107 urban areas, 33, 167, 170, 173, 184. See also local government, federal statistics in urbanization, 163, 164, 167 urban renewal programs, 344—45 user fees, 358, 405-406, 439

V "Valuation of the Social Income, The" (Hicks), 142-43 Value and Capital (Hicks), 245 Vermont, 161 Virginia, 161 voice communications, digital, 404, 409 vote dilution, 311-13 voting behavior, 263-64, 296 Voting Rights Act of 1965, 199, 273, 303-27; amendments of 1970,307-11; amendments of 1975,322-24; amendments of 1982, 231,326-27; central provision of, 306-307; criticisms of, 304-306; electoral districting and, 311-19; groundwork for, 305-306; New York City example, 305,309-11,313-22; preclearance provision of, 306, 308-14,

473

THE POLITICS OF NUMBERS

322,326; racial polarization and, 318-22; subterfuge and, 325-27

W wages, reporting of, 105 wage statistics, 283, 285 Wall Street Journal, The, 39n, 186n, 304 "War Expenditure and National Production" (Gilbert), 145 War on Poverty, 199, 270-71 wartime production allocation problem, 143-45 welfare, 124, 137,140, 168-70 West Germany: organizational structure of statistics in, 27; output per worker in, 74; "statistics strike" threat in, 32r-33

474

West Indian blacks, 208, 209 whites: ethnic heritage of, 189; reverse discrimination and, 200; undercounts of, 158, 175 women, 34; Hispanic, 224; unemployed, 125-26; working wives, 110 World War I, 140, 163, 178, 208, 302 World War H, 43, 84, 93, 143-46, 173, 197, 213, 333, 343, 396

Y Yiddish, 224 Young v. Klutznick (1980), 182-84