Coping With Divorce, Single Parenting, and Remarriage : a Risk and Resiliency Perspective 9781410602893, 1410602893

Table of contents :
Content: Cover
Half Title
Title Page
Copyright Page
Contents
Introduction and Overview
PART I: WHY MARRIAGES SUCCEED OR FAIL
1 Predicting the Future of Marriages
2 Black Couples, White Couples: The Early Years of Marriage
PART II: CHILD ADJUSTMENT IN DIFFERENT FAMILY FORMS
3 Multiple Risks and Adjustment in Young Children Growing up in Different Family Settings: A British Community Study of Stepparent, Single Mother, and Nondivorced Families
4 Family Structure, Parenting Practices, and Adolescent Adjustment: An Ecological Examination. PART III: FAMILY FUNCTIONING AND CHILD ADJUSTMENT IN DIVORCED AND SINGLE-PARENT FAMILIES5 Should We Stay Together for the Sake of the Children?
6 Father Absence and the Welfare of Children
7 Children of Divorced Parents as Young Adults
8 Young African American Multigenerational Families in Poverty: The Contexts, Exchanges, and Processes of Their Lives
9 Protective Factors in the Development of Preschool-Age Children of Young Mothers Receiving Welfare
PART IV: FAMILY FUNCTIONING AND CHILD ADJUSTMENT IN REPARTNERED RELATIONSHIPS AND IN STEPFAMILIES. 10 Contexts as Predictors of Changing Maternal Parenting Practices in Diverse Family Structures: A Social Interactional Perspective of Risk and Resilience11 From Marriage to Remarriage and Beyond: Findings From the Developmental Issues in StepFamilies Research Project
12 A Social Constructionist Multi-Method Approach to Understanding the Stepparent Role
13 The Dynamics of Parental Remarriage: Adolescent, Parent, and Sibling Influences
PART V: INTERVENTION
14 Psychological Interventions for Separated and Divorced Families
Author Index
Subject Index.

Citation preview

COPING WITH DIVORCE, SINGLE PARENTING, AND REMARRIAGE /4

a s u z t J^ e siC te ^ LC A / P e *-sf> e ^ t u se s

Edited by

E. Mavis Hetherington

Coping With Divorce, Single Parenting, and Remarriage A R isk a n d R esiliency P erspective

Coping With Divorce, Single Parenting, and Remarriage A R isk and Resiliency Perspective

Edited by

E. M avis H etherington University o f Virginia

Psychology Press Taylor & Francis Croup

H e a

-a ij

f.-a I ul

a

c

cl

3

-I

£ 9 £

hi

!J

I]! I if ill | I s

II s iii

X

li

S'

! i ;

L * w y (if Cw gltss Cilitopmt-irt-Puliliwrin* Dwu

■a Jn .

p tm

Coping ui & Swain, M. A. (1974). Comm unication, conflict, a nd marriage. San Francisco: Jossey-Bass. Shapiro, A. R, & Walker, K. (1997, April). M arital satisfaction: Predicting changes over the transition to parenthood. Paper presented at the the Biennial M eeting o f the Society for Research in Child Devel­ o pm ent, W ashington, DC. Weiss, R. L. (1980). Strategic behavioral m arital therapy: Toward a m odel for assessm ent and interven­ tion. In J. P. V incent (Ed.), Advances in fa m ily intervention, assessment a nd theory (Vol. 1, pp. 229271). Greenwich, CT: JAI Press. Weiss, R. I ., & Sum m ers, K. J. (1983). Marital interaction coding system III. In E. E. Filsinger (Ed.), M ar­ riage and fam ily assessment (pp. 35-115). Beverly Hills, CA: Sage Publications.

2

Black Couples, White Couples: The Early Years of Marriage Terri L. Orbuch Oakland University University o f Michigan

Joseph Veroff Andrea G. Hunter University o f Michigan

A s t h e d i v o r c e r a t e in A m erican society skyrocketed, m any concerned so ­ cial scientists u n d erto o k research to discover the factors affecting both the psycho­ logical well-being experienced by couples as their m arriages progress and the sta­ bility o f these m arriages over tim e. Som e o f these explanatory factors focus on changes in the interpersonal lives o f these couples— how couples try to m anage and regulate their interactions to m axim ize positive outcom es and reduce conflicts (C arstensen 1991;Carstensen, G ottm an, & Levenson, 1995; G ottm an, 1994; M arkm an, 1991; M arkm an, Renick, Floyd, Stanley, & Clem ents, 1993), how they express their feelings (Cowan & Cowan, 1994; G ottm an, 1994; Veroff, D o u v an ,& H atch ett, 1995), and how they do or do not reinforce each o th er (H uston & C horost, in press). O th er factors focus on structural issues in these couples’ lives— eco­ nom ic issues (V inokur & Van Ryn, 1993), the addition o f children to the family (A de-R idder & B rubaker, 1983; Belsky, Spanier, & Rovine, 1983; W hite & Edwards, 1990), the em beddedness o f couples to netw orks (T im m er, Veroff, & H atchett, 1996), and the stru ctu re o f w ork-fam ily connections (O rbuch, H ouse, Mero, 8c W ebster, 1996; O rbuch & Custer, 1995; O rb u ch & Eyster, 1997a). An im p o rta n t re­ cent analysis by A m ato and Rodgers (1997) o f a prospective longitudinal study re ­ veals ways in w hich interpersonal problem s experienced in m arriage m ediate som e o f the structural features found to be predictive o f divorce. T hus, various in te r­ active styles o r structural characteristics can be considered risk factors p ortending 23

24

O R B U C H , VEROFF, H U N T E R

divorce or low m arital quality and a nu m b er can be considered sources o f re­ silience for a m arriage. T hese factors affect n o t only the couples b u t also the children o f these couples. Previous research indicates that m arital tension, conflict, and separation have a d ­ verse effects on children w ho are party to the disrupted m arital life their parents are perform ing; there are disabling long-term consequences o f m arital conflict and divorce on children’s developm ent, well-being, and social and interpersonal rela­ tionships (A m ato, 1993; A m ato & Keith, 1991; D em o & Acock, 1996; H etherington, 1993; H etherington & Clingem peel, 1992; M cLanahan & Sandefur, 1994; O rbuch, T h o rn to n , & Cancio, 1997; Seltzer & Bianchi, 1988; Webster, O rbuch, & House, 1995). It is no w onder th en that in the past two decades som e im p o rta n t longitudinal studies o f m arriage have em erged. M any o f them are still ongoing. M uch has been learned from them about the dynam ics and structural su p p o rts o f m arriage. H ow ­ ever, m ost o f this research has been constructed w ith the eye to establishing u n i­ versal principles about family dynam ics, and for the m ost part the studies ignore the ways in w hich the particular social context in which m arriages are em bedded shape the determ inants o f m arital quality and stability. In this chapter, we concen­ trate on a longitudinal study o f m arriage, the University o f M ichigan’s Early Years o f M arriage Project, because o f its unique focus on the social context o f m arriage. The prim ary objective o f this longitudinal study was to get a b etter understanding o f th e m arital processes and determ inants o f m arital quality and stability am ong African A m erican couples and W hite couples. W hich ones survive? W hich are re­ silient to stress? Given differences in m arital cultures and in the contexts o f m a r­ riage in the Black and W hite com m unities, we anticipated that the explanatory factors affecting m arital quality and stability w ould differ for Blacks and W hites. T hus, this chapter does n o t focus as m uch on the universal dynam ics o f m arriage as it does on ways in w hich m arital quality and stability depend on the social co n ­ text o f m arriage. O ne context is o f particular im portance: How does being African A m erican in A m erican society affect the way m arriage progresses, especially in contrast to being m em bers o f the d o m in an t W hite society?

RESEARCH ON BL A C K -W H IT E DIFFERENCES IN THE M EANING OF MARRIAGE The high rates o f divorce and separation are specifically pron o u n ced for African Am ericans. Yet little research attention has been given to Black Am erican m a r­ riages. According to Billingsley (1992), relatively little attention has been given to Black m arriages because o f researchers’ inability to go beyond the pathological approach to Black families, which relies heavily on single-parent families and teen pregnancies w ithin this cultural group: “O verlooked are the positive functions served by m arriage. O verlooked also are the patterns o f diversity w ithin and am ong

2.

BLACK C O U P L E S , W H I T E C O U P L E S

25

these families” (p. 205). F urtherm ore, Billingsley (1992) argued that given recent U.S. Census data, the thesis o f the vanishing Black family is n o t supported and em ­ pirically untrue. In 1990, “a m ajority o f African A m ericans lived in households; a m ajority o f these were fam ily households; a m ajority o f family households were m arried-couple households; and a m ajority o f m arried couples had young chil­ dren o f their ow n” (Billingsley, 1992, p. 207).

Structural Explanations In general, social psychologists and o th er social scientists w ho study m arriage have focused greater attention on th e structural rather than interactional factors that underlie the risks o f divorce or p rom ote psychological well-being and stability o f m arriages. This structural em phasis is especially evident w ithin the literature on racial differences in divorce rates. M arital quality an d interpersonal processes w ithin African A m erican m arriages have been scarcely studied and are n o t well u n ­ derstood (Tucker & M itchell-K ernan, 1995) Researchers have proposed various structural explanations o f the differential rates o f m arital quality and stability betw een African A m ericans an d W hites. O ne structural explanation focuses on the sex ratio o f Black males to females (Lichtcr, M cLaughlin, Kephart, & Landry, 1992; Staples & Johnson, 1993; Walker, 1988). T he m arriage m arket m ay be different for African A m ericans than for W hites b e ­ cause there is a sm aller num ber o f m arriageable Black m en than w om en. This sit­ uation increases the probability o f un m arried Black females and decreases the nu m b er o f possible b usband-w ife com binations (Staples & Johnson, 1993). T he unbalanced m arriage m arket also influences the degree to which Black m en have a larger pool (either physically o r psychologically) from which to choose a pa rtn e r o r m ate, which m ay affect the selection o f a m ate o r the likelihood that one stays w ithin a m arriage. The unbalanced m arriage m arket is viewed as also c o n trib u t­ ing to Blacks’ m arital dissolution (Elder 8c Caspi, 1988; H atchett, Veroff, 8c D ouvan, 1995). A nother structural explanation is that African A m erican couples face different structural obstacles and greater ongoing tensions and injustices th an W hite co u ­ ples. Such ongoing conflict w ith the external environm ent m ay carry over and affect the well-being and stability o f African A m erican m arriages. African A m eri­ can m arried couples m ay also have m ore difficulties and obstacles to overcom e given the unstable job m arket and lower earning pow er o f African Am erican males. In the 1980s Black w om en m ade significant gains in achieving stru ctu ral re­ sources, b u t Black m en did n o t (Staples, 1991). T here is also evidence to suggest th a t since 1954, unem ploym ent am ong African A m ericans in general has been twice as high as am ong W hite A m ericans (Billingsley, 1988). Therefore, African A m erican m arriages m ay suffer disproportionately from the long-term effects of unem ploym ent. In addition, although the results are inconclusive (A delm ann, Chadwick, 8c Baerger, 1996; B rom an, 1993), socioeconom ic factors such as lower

26

ORB UCH , VEROFF, HUNTE R

average household incom e am ong Black families may also explain the higher rates o f divorce and lower m arital quality o f Black marriages.

Cultural Explanations Past research has given little attention to the role of cultural factors beyond struc­ ture affecting differential rates o f quality and stability am ong Black and White marriages. We argue, however, that as social researchers we need to understand that m arital relationships are em bedded w ithin the context o f a racial and ethnic culture, consisting o f norm s and expectations that determ ine the m eanings of var­ ious interpersonal or m arital processes, such as conflict, parenthood, or decisions regarding the division o f household labor. These meanings are im portant to u n ­ derstanding racial differences in divorce rates; the various meanings are not nega­ tive or deviant b u t integral to the organization and functioning o f those marriages that endorse them (Walker, 1988). Thus, cultural norm s exist and are significant to explaining m arital quality and stability am ong both m ajority and m inority cul­ tures in American society. It is also im portant to note that there is considerable di­ versity in these cultural norm s, both between and w ithin cultures. If we focus on cultural factors beyond structure, we m ight expect Black-W hite differences in m arital processes for several reasons. First, Black and W hite m arriages may not be affected by the sa me factors and in the same ways. Specific factors may be im portant to one culture but not critical to m arriage in the other culture. For exam ­ ple, Orbuch and Eyster (1997a) found that husbands’ participation in household responsibilities and childcare was related to m arital well being for Black wives but n ot for W hite wives. Broman (1993) also dem onstrated that som e o f the large dif­ ferences found in Black-W hite com parisons of reported m arital quality (i.e., Blacks reporting lower m arital quality) can be attributed to the degree o f financial satisfac­ tion expressed by couples, particularly Black wives, and the degree o f spousal su p ­ port. A path m odel revealed im portant indirect effects of race on m arital well-being through these factors. The meanings o f interactions and styles o f behaviors within each ethnicity o r race can help clarify the cultural context w ithin which the well­ being and stability of marriages are experienced over time. Second, the underlying factors that determ ine m arital quality' and stability for Black and W hite couples may be similar, but how these factors get translated or ex­ pressed within m arriage differ for the two groups. For example, a supportive social netw ork is a significant predictor o f marital stability for Black and W hite couples, yet this support may be expressed or em bedded within different networks for Blacks and W hites: Stack (1974) found that neighborhood support is particularly im portant to the well-being of Black wom en raising their children. N eighborhood support m ight very well com pensate for o r perhaps underm ine the social support experienced w ithin the couple’s own relationship. Thus, the expression of support and where it stems from may differ within the two groups, although the basic m echanism or factor is similar.

2.

BLACK C O U P L E S , W H I T E C O U P L E S

27

We m ight also expect B lack-W hite differences in risk factors for m arital insta­ bility and for resiliency factors that p rom ote m arital well-being because Blacks and W hites m ay com e to m arriage w ith different family trajectories and histories (e.g., Black w om en m ay be m ore likely to be parents), which also have consequences for how factors are related to m arital processes. The norm ative transitions for the m a­ jority W hite culture (i.e., m arriage, then parenthood) m ay n o t be norm ative for oth er cultural groups. A delm ann et al. (1996), for example, argued that one p ro b ­ lem w ith perspectives th a t rely solely on stru ctu ral characteristics to explain m ari­ tal quality for Blacks over the life course is the assum ption o f negative m arital effects o f children. T he satisfaction and well-being o f African A m erican m arriages may be less integrally connected than W hite m arriages, to the presence o r absence o f children. This m ay be tru e because Black couples spend less than a year together before having a child and thus are m ore likely to enter m arriage as parents, com ­ pared w ith W hite couples w ho spend an average o f 2 years together before p aren t­ hood (L ittlejohn-B lake & D arling, 1993). T his em phasis on differential trajectories and histories based on race and ethnicity can lead to either u nique factors con­ tributing to m arital quality and stability in the two groups o r separate illustrations o f the sam e underlying factors. Finally, it is im p o rta n t to understand th at Black w om en and m en m ay organize their lives differently from W hite w om en and m en, specifically around issues o f gender and power. Studies find that Black couples are m ore egalitarian than W hite couples in their attitudes tow ard wom en and gendered roles (H unter & Davis, 1992; H u n ter & Sellers, 1998; Kane, 1992; O rbuch & Eyster, 1997a). H u n ter and Sellers (1997) claim ed that for African A m erican m en, one positive proactive adaptive strategy to their experiences w ith the lack o f stru ctu ral o p portunities has been to value the in p u t o f b o th w om en and m en inside an d outside the hom e and to ad o p t an egalitarian gender role ideology. F urtherm ore, m uch evidence suggests th a t Blacks are m ore likely than W hites to criticize gender inequality and trad i­ tional views o f m asculinity and fem ininity (Collins, 1990; H u n ter & Davis, 1992; H u n te r & Sellers, 1998; Kane, 1992). Given these differential views o f w om en and m en and th eir im p o rt to the econom ic and psychological survival o f the family, the factors th at influence m arriage or the underlying representations o f these factors m ay differ for Black and for W hite couples. In sum m ary, the cultural context o f early m arriage and the interpersonal p ro ­ cesses th at transpire w ithin these m arriages may be different for Blacks and W hites. F urtherm ore, the cultural context also filters ways that m en and w om en in m arriage experience the pow er they have w ithin the m arriage. As Fine (1993) stated, “diversity is som etim es narrow ly used to refer to differences in family stru c ­ ture, as opposed to differences in how families function and operate b o th w ithin and outside o f their hom es” (p. 237). E xplanations th at focus o n cultural factors beyond stru ctu re are im p o rta n t for a better understanding o f racial differences and sim ilarities in the quality and stability o f African A m erican and W hite m ar­ riages. O u r recent w ork stem s from a longitudinal study o f early m arriage and fo­

28

ORBUCH, VEROFF, HUNTER

cuses on the differential m eanings o f m arriage for Blacks and W hites w ith specific a tte n tio n to the cultural context o f those m arriages.

THE EARLY YEARS OF M ARRIAGE STUDY O u r ow n insights about the differential m eaning o f m arriage and m arital quality and stability am ong A frican A m ericans and W hites derives from a longitudinal Early Years o f M arriage study. Since 1986, a research group at the Institute for So­ cial Research has been collecting and analyzing data from an urban sam ple o f c o u ­ ples, m arried for the first tim e in 1986 w hen they were first interview ed and then followed up again yearly in 1987-1989 and again in 1992 d u ring their seventh year o f m arriage. T his project started w ith 199 African A m erican couples and 174 W hite couples w ho applied for m arriage licenses in Wayne County, M ichigan, d u ring 1986. O nly couples w ho were still m arried were reinterview ed. To be eligi­ ble for the study, both m em bers o f the couple had to be o f the sam e race and in their first m arriage. All eligible Black couples and a random sam ple of W hite co u ­ ples were contacted and asked to take p art in the study. M any different m ethods were used in both face-to-face and telephone interviews. Individuals were asked standard and open-ended questions in their interviews. C ouples were asked to do two unusual tasks together: O ne was a narrative procedure for them to tell the story o f their relationship (Veroff, Sutherland, C hadiha, & O rtega, 1993); the oth er was a behavioral observational procedure for them to resolve their differ­ ences ab o u t the im p o rta n t rules for a good m arriage (C oon & Veroff, 1996; Crohan, 1992). G ottm an (1994) em ployed sim ilar procedures in his prospective study o f m arriage. T he m ajor findings from this longitudinal study in m arriage contrasting A frican-A m erican and W hite couples’ m arital stability in the early years o f being a couple are found in M arital Instability (Veroff e ta l., 1995). In this book, m arital in ­ stability was exam ined by both an objective assessm ent o f w hether the couple d i­ vorced after Year 1 and a subjective assessm ent o f how often each spouse thought a b o u t leaving the m arriage. Veroff et al. (1995) explored the ways in which the de­ term inants o f m arital instability are identical in A frican A m erican and in W hite couples d uring the first 4 years o f their m arriages. These sim ilarities, however, are few and far betw een. T he authors have m ore to say about the ways in w hich the de­ term inants are different in these two groups. In som e instances the differences rep­ resent parallel phenom ena in the tw o groups, but these phenom ena are m anifested in different ways. In o ther instances, the differences truly represent ways in w hich the contexts for m arriage elicit different concerns in the two groups. These m ajor findings in M arital Instability relevant to African Am erican an d W hite com par­ isons in the m eaning o f m arriage are sum m arized next, after w hich we integrate other findings from the project that confirm o r add to the conclusions about Black-W hite differences that M arital Instability provided.

2.

BLACK C O U P L E S , W H I T E C O U P L E S

29

Results Emerging from Marital Instability First, we provide a su m m ary o f the differential m eanings o f m arriage discovered for Blacks and W hites in M arital Instability. We carefully list results that were high­ lighted in the book, and in this chapter we expand on the in terp retatio n o f the m eanings for m arriage that these results on Black-W hite differences m ight have for the two different cultural contexts. In particular, this discussion em phasizes the issues o f gender-pow er dynam ics th a t inform o u r in terpretations o f interactional processes am ong Black couples and W hite couples. T here are four instances where identical variables are associated w ith m arital instability in b o th African Am erican and W hite couples: frequency o f conflict, the wife having an affair, interferences experienced from wives’ friends, and the experi­ ence o f m arital unhappiness am ong wives. These are all interactional risks for d i­ vorce, rather than structural ones. The structural risk factors that em erged were specific to one group o r an o th er b u t n o t general to both. W hen we look closely at the four instances o f identical risk factors in the two groups, we note th a t beyond the expected finding that the m ore conflicted m arriages are, the m ore unstable they are, the oth er com m on results to both groups have to d o w ith wives specifi­ cally. This p attern o f results suggests that any m arriage, w hether involving African A m erican o r W hite partners, will likely becom e unhinged w hen a w om an takes an active stance about the difficulties in her m arriage (e.g., having an affair, co m plain­ ing to their friends) o r is free to acknowledge her desperate feeling (e.g., reporting her m arriage as unhappy). There are also certain them es th at are parallel in understan d in g m arital risk fac­ tors in both African A m erican and W hite m arriages b u t have different m anifesta­ tions. These them es im plicate im p o rta n t interactive phenom ena that are critical for stable m arriages o r gender-pow er phenom ena th a t can destabilize m arriages, o r both. We list each them e in Table 2.1. We also list in the table how th at them e is prototypically m anifested in the two groups. We discuss each in turn.

Supportive Wives T he first parallel them e is that in stable m arriages wives are supportive. T his is clearly a direct interactive ph en o m en o n b u t one th at can touch off m any concerns a b o u t what is appropriate behavior for m en and w om en in m arriage. However, this them e takes on slightly different channels in two subcontexts. For W hite stable m arriages, wifely n u rtu ran c e is found in w hether the wife is judged to be coopera­ tive by both the husband and the wife. This is n o t so for the African A m erican co u ­ ples, w hich suggests that being judged as cooperative for a W hite w om an implies an overt kind o f n u rtu ra n t com pliance to their husbands. Being cooperative in this way evidently can have other negative m eanings for som e African Am erican couplcs, for w hom a w om an’s collaborative style is im p o rta n t for understanding m a r­ ital stability.

30

ORBUCH, VEROFP, HUNTER

TABLE 2.1 Parallel Them es Found in Stable M arriages o f Both African American and W hite Couples W ith Different Prototypic M anifestations

Parallel Theme

Prototypic Manifestation in Stable African American Couples

Prototypic Manifestation in Stable White Couples

Stable m arriages depend o n n u rtu rin g wives.

Wives have collaborative style w ith their husbands.

Wives are rated high in being cooperative.

Stable m arriages require com patibility in interpersonal interactions.

Infrequent fighting. C om ple­ m entariness in openness to experience.

Infrequent fighting. T here is sexual tension, and spouses share leisure activities.

Stable m arriages depend o n husbands having d e a r evidence o f their achievem ent in society.

Husbands have highly educated m others, have low anxiety, and are low in alcohol usage.

H unsbands have high p e r­ sonal incom e and have jobs that are reported as n o t in te r­ fering w ith their marriages.

Stable m arriages depend on harm onious integrations w ith social networks.

Little interference from friends and from husbands' families.

Little interference from friends.

Stable marriages depend on husbands m aintaining som e independence and control o f their wives.

Wives have not had affairs.

Wives have not had affairs. H usbands don’t com plain about having no privacy. Wives are not high in alcohol usage.

We propose two explanations for the collaborative style o f interaction for Black couples. First, this measure can be a much more subtle kind o f nurturance. Collab­ orative style is a measure o f how much a person in a joint storytelling situation dovetails her remarks with those o f her partner. This collaborative style is not nec­ essarily agreement but a meshing o f ways o f interacting. It may very well be that som e Black men would be suspicious o f a more direct nurturant style from their wives, a style that would immediately suggest that the wife be rated as cooperative. African American men perhaps need to be subtly nurtured by their wives, to feel comforted by their wives’ way o f interacting, because they feel vulnerable about their power in the family and hence can easily read their wives’ cooperative efforts as conscious manipulations to make a go o f a marriage. Alternatively, this collaborative style o f interaction may represent the signifi­ cance and value that Black men place on w om en’s independence and strength. Black men desire an active, equal, and supportive partner in their spouse. Given differences in structural opportunities for Black men compared to White men, a reliance on an equal partner rather than a submissive or cooperative one is critical to Black couples’ marriages; Black families’ survival is often dependent on two ac­ tive partners.

2.

BLACK C O U P L E S , W H I T E C O U P LE S

31

Interpersonal Compatibility It seems like a m ere tru ism to state the second parallel them e as baldly as we did: Stable m arriages require interpersonal com patibility. In both sets o f couples the tendency to fight a lot, a m ark o f considerable incom patibility, is characteristic of unstable m arriages, b u t each set also has particular kinds o f com patibility that are im p o rta n t to m arital stability. For the Black couples, there has to be a balance o r a com plem entary orientation to experience; one spouse has to be cautious while the other is adventuresom e. For W hites, com patibility in their sexual and leisure lives is im p o rta n t for stability. T his set o f findings leads us to suggest that sexuality is n o t centrally connected to the risk factors to be found in the m arital lives o f African A m ericans b u t it becom es a m ore figural risk issue for W hite couples. We also suggest that Black husbands an d wives lead m ore independent lives as m arried couples than d o W hite husbands and wives. Such being the case, som e balance on each spouse’s orientation to experience seems critical for Black couples, whereas sharing o f leisure becom es m ore critical for W hite couples. All these findings deal w ith interpersonal com patibility, w hich takes different form s in the two subcultures. Som e separation and independence from each other m ight be m ore critical for com patibility to African Am erican than W hite couples. This may be the case because African Am erican couples have so m any ways that their m arried life m ight be d isrupted by external factors. F urtherm ore, as stated previously, there is greater gender parity, equality, and an appreciation o f w om en’s independence in Black than W hite m arriages (H unter & Davis, 1992). T hus, in ter­ personal com patibility takes on different form s in the two groups.

Male Achievement or Power The third parallel them e noted in M arital Instability speaks o f m en having to feel accom plished. W ith the W hite husbands, the them e em erges directly. T he higher W hite husbands’ incom es are and the less interference their jobs create in their m arital life, the m ore stable their m arriages. We can read the first result as a stabi­ lizer o f m en’s call o n pow er in the m arital relationship and the second result as show ing that the w ork dem an d s on m en should n o t disru p t sm ooth interactions in the m arriage. T his pattern is n o t the case for the African Am erican m en, w ho show ed m ore indirect evidence o f the them e: the education o f their m others an d w hether they themselves were anxious a b o u t their lives o r tu rn ed to alcohol. Black husbands w ith highly educated m others, w ho were free from anxiety and were not heavy drinkers, were m ore likely to be in m ore stable m arriages. In contrast to W hite co u ­ ples, husbands’ financial success d u rin g young adulth o o d did n o t have a significant im pact on the m arital stability o f Black couples. This m ay reflect the occupation trajectories o f Black m en, w ho often begin their adult work lives w ith lower paying jobs and experience a slower clim b to peak earnings than do W hite m en (Davis,

32

O R B U C H , VEROFF, H U N T E R

1995; Lichter, 1988). However, because m others’ education is positively related to educational achievem ent in Black m en, it may have an effect on m en’s achievem ent potential (Epps, 1985; Sm ith, 1991). Because educational achievem ent is related to occupational status, Black husbands w ith highly educated m others m ay have b e t­ ter long-term prospects for occupational achievem ent and in tu rn m ay have less provider-role anxiety th an the sons o f less educated m others. Indeed, despite the gains in occupational o p portunity, Black m en’s w orries ab o u t job security and financial well-being are well founded (Davis, 1995; Lichter, 1988). Being anxious and seeking illusory pow er th rough drinking, o th er significant risk factors for African A m erican m en, can be sym ptom atic o f som e m en’s concerns about how well they are providing for their families. In addition, unlike W hite husbands, black m en m ust m anage race-related stres­ sors (e.g., discrim ination) that may underm ine m arital relationships (Johnson, 1990; Taylor & Z hang, 1990). Race-related stressors, particularly as they are m an i­ fested in em ploym ent, m ay be reflected in increased anxiety and alcohol use in Black males (Bow m an, 1988; W illiams, Lavizzo-Mourey, & W arren, 1994). Accord­ ing to Johnson (1990), findings su p p o rt a direct relationship betw een Black m en’s perceived race-related job stress and strained couple interactions as well as the p o ­ tential for separation and divorce.

Integration With Social Networks T he fourth them e, stable m arriages being d ependent on h arm o n io u s integrations w ith social netw orks, also has som e com m on m anifestations in both groups. Both sets o f m arriages find it hard to w ithstand interferences from netw orks o f friends, b u t Black couples in addition find it difficult to rem ain stable when the h u sb a n d ’s family is interfering. Thus, netw ork interactions are im plicated by these results. P rior research suggests th at integration w ith families and intergenerational re­ lationships are m ore im p o rta n t to the well-being and stability o f African A m erican families than they are to W hite families (Stack, 1974; Taylor, C hatters & Jackson, 1993). It is n o t hard to suppose th at the su p p o rt th at African A m ericans derive from their families seems especially critical for them to w ithstand the disruptive forces on m arriages that occur in a racist society. Families can provide the in te r­ active reassurance that helps one o r the o th er spouses, b u t m aybe particularly the husbands, to feel com fortable about their threatened position in the society that, in tu rn , has repercussions ab o u t a threatened sense o f pow er in the m arriage.

Husband’s Independence and Control T he last them e reiterates the patriarchal ideal, that husbands in successful m a r­ riages m aintain th eir independence and som e control o f their wives. In both groups a wife’s affair was disruptive to the m arriage, b u t a h u sb a n d ’s affair was not consistently disruptive. A m an whose wife has com m itted adultery has had a blow to the m ale ego, m uch stronger than the blow to the fem ale ego caused by a ph ilan ­

2.

BLACK C O U P L E S , W H I T E C O U P L E S

33

dering husband. In this way African A m erican and W hite m arriages are alike. However, W hite males, in contrast to Black m ales, seem to need even m ore reassur­ ance a b o u t being in control o f their ow n lives w ith o u t overt need o f their wives. H aving no privacy is a com plaint W hite b u t n o t Black husbands voice in unstable m arriages. The im age o f the solitary m ale w ithstanding obstacles and also having adventures w ith o u t the restriction o f a family is a highly rom anticized them e in o u r society. W itness the popularity o f the western m ovie. This ideal perhaps has a stronger foothold in the lives o f the m ajority W hite group. It is w ith that ideal in m ind th at we note that M arital Instability finds that stable W hite m arriages depend on wives n o t being alcoholic, unlike stable Black m arriages. O ne could suppose th at w hen their wives are alcoholic, W hite males need to be especially caretaking. Being tied dow n like th at m ight u n d o som e o f the hu sb an d s’ autonom y. These inroads into autonom y m ay n o t shake the African A m erican m arriages as m uch because these m arriages m ay hold m ore cultu ral n o rm s for independence am ong Black w om en as well as m en. Black husbands already have the skills and ideology to deal with problem s in the household should they com e up as a result of a wife’s d rinking problem . We have clear evidence th a t only in African Am erican m arriages does stability have anything to do w ith the sense o f pow er th at wives have in controlling their lives (Veroff et al., 1995). F urtherm ore, we find that only in Black m arriages is an egalitarian role orientation a positive force for m arital sta ­ bility. As we noted in the first and th ird them es, Black husbands are not threatened as m uch as W hite husbands are about th eir wives being m ore independent o f them on a day-to-day basis, although having an affair is still a threat to their m asculinity. From o u r presentation o f parallel them es noted in M arital Instability, we have suggested a n u m b er o f ways in which the African A m erican context for a stable m arriage m ight be different from the d o m in a n t w hite context. We sum m arize these here: 1. Black wives are m ore likely to be collaborative in their interactions with husbands. In white m arriages, direct cooperativeness is desired. T his collabo­ rative style m ay represent a subtle form o f n u rtu ran c e tow ard Black husbands o r m ore likely the validation and significance o f Black wives’ independence and strength. 2. Given the differential stru ctu ral positions betw een Black m en and W hite m en, there are o th er issues beyond direct financial take-hom e pay that may represent the accom plishm ents o f the Black m ale in Black couples (e.g., antici­ patory achievem ent given m o th e r’s education, fewer job-related stressors lead­ ing to anxiety). A m ong W hite husbands, financial success is p a ram o u n t to a m ale’s achievem ent. 3. For Black m arriages, com patibility about sex and spending tim e together in leisure are not critical, b u t these issues are im p o rta n t in white m arriages. 4. Integration w ith families o f origin is a m ajor su p p o rt o f a m arriage for Blacks, but it is less critical for Whites.

34

O R B U C H , VEROFF, H U N T E R

5. There is an acceptance o f parallel but relatively independent lives in the way Black husbands and wives conduct a m arriage, unlike the n o rm for inter­ dependence that is m ore im p o rta n t in W hite m arriages but that m akes W hite husbands feel defensive ab o u t their autonom y. T his acceptance illustrates that unlike W hites, Black husbands and wives organize their m arried lives with greater attention to w om en’s independence and equality.

Other Results from the Study M any o th er results have been reported about Black-W hite com parisons from the Early Years o f M arriage study. M ost dovetail well w ith the in terpretations we have discussed; som e offer new insights about the different m eanings o f m arriage in the two contexts. In all instances these findings confirm the general orien tatio n we have taken in this chapter, that different cultural contexts as we find in W hite and Black subgroups represented in o u r urb an A m erican sam ple can differentially shape the gender-pow er dynam ics that inform o u r interpretations o f the interac­ tive processes found in m arriages at risk. S u p p o rtiv e W ives. F urther evidence for how im p o rta n t it is for Black wives to subtly m anage their relationships in n u rtu ra n t ways com es from findings that Acitelli, D ouvan, and Veroff (1997) have docum ented. These researchers report that w hen African A m erican wives indicate that they u nderstand w hat their hus­ bands are doing w hen faced with a conflict, their m arriages are happier. They evi­ dently translate their understan d in g into behaviors that p e rm it their husbands’ attem pts to deal with their fights to be truly constructive. T he results were just the opposite for the W hite wives. T heir u n derstanding correlated w ith m arital u n h a p ­ piness. These findings suggest that the sam e type o f understan d in g in the W hite wives gets translated into som ething that fu rth er exacerbates couples’ conflicts. We conclude that Black wives subtly take their understanding in to account in their interactions, whereas W hite wives perhaps feed th at understanding into further confrontations about their conflicts. A sim ilar type o f subtle su p p o rt by African A m erican wives was reported by Veroff, D ouvan, O rbuch, a n d Acitelli (1998). T hey found that the especially happy African Am erican m arriages in the fourth year o f m arriage (i.e., those w ho have never thought about leaving the m arriage and profess extrem e happiness and sat­ isfaction with the m arriage) were ones w ith p a rtn e rs who had considerable over­ lap in reported interests and hobbies. It is not so m uch th at they share leisure tim e together but that they share interests. This finding was n o t apparent for h a p p i­ ness in the W hite m arriages, although, as we n oted previously, sharing leisure tim e was im p o rta n t for m aintaining the stability o f W hite m arriages. We interpret this set o f findings to m ean that w hen a Black w om an has an interest in w hat her hus­ band is interested in, this interest m ay be totally sincere and th u s positively con­ veyed as caring for him . W hy does this re iu lt n o t apply to W hite couples? We can

2.

BLACK C O U P L E S , W H I T E C O U P L E S

35

only suggest th at a W hite w om an m ay be less sincere o r less subtle in the way she conveys a m u tu al interest, so that the husband m ay th in k o f her as being p o ten ­ tially m anipulative. It has becom e p a rt o f the d o m in an t W hite folklore that m en and w om en d o not share the sam e interests. W hen they do, o r appear to, som e m en m ay be uncom fortable o r suspicious. In te rp e rso n a l C o m p a tib ility . The different m eanings o f interpersonal com ­ patibility in the two contexts o f A frican A m erican and W hite societies have also show n up in o th er investigations w ith the Early Years o f M arriage Project data. Veroff et al. (1998) have reported that, for couples w ho rem ain especially happy after four years o f m arriage, the presence o f positive sex and the absence o f nega­ tive sex seems m ore critical for W hite couples, and the absence o f conflict seems m ore critical for Black couples. We have already noted that un tro u b led sexuality is m ore im p o rta n t to the stability o f W hite m arriages than it is for Black m arriages. W hat these new results tell us is th at enjoym ent o f sexuality is im p o rta n t for the es­ pecially happy side o f m arriage for the W hite couples as well. It is difficult for us to state that sexuality is u n im p o rta n t for African A m erican couples, b u t we can c o n ­ clude that evaluating a m arriage on the basis o f sexual feelings is less p a ram o u n t in Black m arriages th an in W hite m arriages. H enderson-K ing and Veroff (1994) have also found m any com plicated differ­ ential pattern s for the effects o f sexual feelings on various facets o f m arital well­ being. In their analyses, sexuality does n o t drop out o f the picture for African A m erican w ell-being, n o r is it always pro m in en t in W hite well-being. N everthe­ less, it does vary by race and also by gender, and they advocate that fu rth er analyses o f sexuality need to be pursued differentially for African A m ericans and W hites. This finding alerts us in the future to consider in m ore d ep th the specific contex­ tual m eaning o f sexuality in m arriage for the two groups. The fact that the absence o f conflict seem s particularly critical for Black wives to be especially happy in their m arriages (Veroff et al., 1998) led us to consider the possibility th at the m eaning o f m arital conflict m ay be different for African A m eri­ can couples and W hite couples. Results from a study by O rb u ch , Veroff, and H olm berg (1993) corroborate this possibility. O rbuch et al. (1993) found that Black co u ­ ples were significantly m ore likely to present stories w ith couple conflict than were W hite couples. However, findings from this study also indicate th at for all but the Black husbands, som e conflict in telling th eir courtship story in the first year of m arriage was predictive o f less m arital happiness in the third year. These findings confirm K ochm an’s (1981) work, w hich suggests that Black couples are m ore likely to conflict w ith each o th er in interaction as a general style o f discourse, b u t it is n o t necessarily abusive o r destructive and can be seen as playful and engaging. Thus, Black couples m ay display m ore conflict w ith each other, b u t the m eaning behind th at interactional style m ay differ for Blacks and W hites. C rohan (1992), however, has also show n that for bo th Blacks and W hites, h o ld ­ ing the follow ing attitudes tow ard m arital conflict have negative effects on m arital

36

O R B U C H , VEROFF, H U N T E R

happiness: beliefs th at conflicts should be avoided, beliefs th at conflicts are un healthy, or beliefs that conflicts cannot be resolved. T herefore, a general shying away from conflict resolution puts couples from both groups in jeopardy. We had started o u r research into B lack-W hite differences in the m eaning o f m arriage w ith an expectation th a t Black couples w ould be m ore im m u n e from conflict troubles in m arriage than W hites, given the research by Kochm an (1981), which suggests that conflict m ay be taken for granted m o re in hum an interactions in A frican A m erican society. O u r finding n oted in Table 2.1 th at infrequent fighting was in fact a stronger predictor o f stability in Black couples than in W hite couples suggested som ething to the contrary. C rohan’s (1992) analysis o f conflict attitudes also gives us few new insights ab o u t why we found this difference and why Black wives are particularly happy w hen they experience infrequent m arital conflict. O u r best speculation is th a t there are fewer barriers tow ard divorce in the African A m erican c om m unity in general, so that frequent fighting if seen as prog­ nostic o f a problem atic future can m ore quickly lead to divorce in an African A m erican couple. We indeed found th at African A m erican couples rated the fol­ lowing as a good reason for divorce significantly higher than the W hite couples: fighting and arguing and financial problem s. The W hite couples, com pared to the African A m erican couples, rated the following statem ent higher as a good reason for n o t divorcing: th eir families were opposed to divorce. A m ong A frican A m eri­ cans, there m ay be a lower threshold for tolerance o f fighting to preserve a m a r­ riage, especially if the fighting is ab o u t financial m atters, which it often is. O u r cou­ ples rep o rt fighting ab o u t financial m atters as a very frequent source o f difficulty. T here are no differences betw een Blacks and W hites in reported frequency o f fight­ ing ab o u t m oney m atters, b u t Black husbands were m ore dissatisfied w ith the way things were left after fighting ab o u t m oney m atters than were W hite husbands. T his result suggests th at fighting a b o u t m oney m atters is hard to resolve for Black m ales w ho perhaps feel especially vulnerable to their own anxieties about financial m atters. It is n o t hard to guess that the African A m erican m ale’s anxiety about be­ ing an adequate provider induces considerable m arital conflict, which m ay be the trigger for m arital unhappiness and instability. M a le A c h ie v e m e n t a n d Power. There are also oth er pieces o f evidence for how the them e o f m asculine achievem ent as a force for m arital stability and h a p p i­ ness m ay be differentially experienced by African A m erican and W hite husbands, and this evidence m ay cast a differential light on g e nder-pow er dynam ics in each group. Veroff et al. (1993) show ed that couples’ discussions o f financial issues as p a rt o f their story o f their relationships correlates significantly w ith the m arital unhappiness o f Black husbands b u t w ith no other group. So does m entioning is­ sues dealing w ith children. It is not hard to interpret these findings to m ean th at financial insecurity w ith regard to raising children in a family is a m ore p ro m in e n t issue for how a Black h u sb an d feels about his m arriage than for the other groups. So long stereotyped as an unreliable family person and provider, the Black h u s­

2.

BLACK CO UP LE S, W H I T E CO UP LE S

37

band has a handicap in ever thinking o f him self as an adequate achiever. M ost W hite couples have inherited a legacy o f financial su p p o rt from their families if they are doing well. For exam ple, m any young W hite couples can co u n t on their families for help in puttin g together dow n paym ents for a first hom e. N ot so for m ost African Am erican couples whose rise in the m iddle class has com e w ith little financial backing from parents. For this reason we th in k financial insecurity re­ m ains a d o m in an t anxiety for Black m ales in particular, even w hen and perhaps es­ pecially when their incom es rise. N um erous studies (B rom an, 1988; Hossain & R oopnarine, 1993; Taylor, C h a t­ ters, Tucker, & I^w is, 1991; W illie, 1985), including findings from the Early Years of M arriage Project, indicate that Black husbands are m ore egalitarian and m ore likely to participate in household and childcare tasks than W hite husbands (O r­ buch 8c Custer, 1995; O rbuch 8c Eyster, 1997a; S utherland, W ent, 8c D ouvan, 1990). African A m erican husbands do not appear to be threatened o r challenged w hen they participate in fem inine tasks in the household. F urtherm ore, O rbuch and Custer (1995) fo u n d th at although career wives m ay threaten o r challenge both W hite and Black husbands, this threat is especially salient for W hite husbands w ho participate in traditionally fem inine household and childcare tasks. Participation in housew ork seems to threaten their positions as d o m in a n t males in the family and is related to increased anxiety. For Black husbands, however, wives’ w ork has adverse effects on well-being, b u t only if they are n o t able to justify the w ork as eco­ nom ically necessary for family survival. In general, whereas the balance o f pow er for A frican A m erican husbands is challenged by financial insecurities, that balance for W hite husbands is threatened by participation in traditionally fem inine tasks in the hom e. O rbuch and Eyster (1997a) argued th at Black husbands in the Early Years o f M arriage Project were m ore aware o f gender inequality and thus less threatened by involvem ent in fem ale-typed tasks at h om e than were white husbands. This argu­ m ent is supported by H unter and Sellers (1998), w ho contended th at African A m erican m en are likely to su p p o rt an egalitarian gender-role ideology because of th eir “unique social lo cation— that is, their sim ultaneous experience o f race o p ­ pression and gender privilege, in additio n to less gender asym m etry in the eco­ nom ic and social organization o f A frican-A m erican families” (p. 2). O rbuch and Eyster (1997a) also provided evidence th at African A m erican wives have signifi­ cantly greater resources (e.g., incom e and education) relative to their husbands than do W hite wives. T his greater econom ic independence may explain why Afri­ can A m erican husbands’ participation in hom e labor was positively related to m ar­ ital well-being for African A m erican wives b u t n o t for W hite wives. African A m er­ ican wives may expect their husbands’ assistance at hom e. In te g r a tio n W ith So cia l N e tw o rks. There has been additional su p p o rt for the conclusion th at the well-being and stability o f African A m erican couples de­ pend significantly on the integration w ith their families, to a greater extent than for

38

O R B U C H , VEROFF, H U N T E R

W hite couples. Veroff (in press) found in Black males b u t in no o th er group evi­ dence that integration w ith in-laws contributed to com m itm ent to m arriage by the seventh year. Tim m er, Veroff, and H atchett (1996) specifically found th at am ong Black couples, a close relationship betw een the couple and the h u sb a n d ’s family in the first year o f m arriage predicted greater happiness in the th ird year, particularly w hen the couple was stressed by a com bination o f low incom e and high family re­ sponsibilities w ith children. The h u sb an d ’s family am ong Blacks can act as a p a r­ ticularly im p o rta n t buffer for m arital stability. T im m er an d Veroff (1997) found in-law buffering m uch m ore prom inent in predicting the risk for divorce in Black couples com pared to W hite couples. Thus, we can conclude th at the role o f the ex­ tended fam ily is m ore critical for the well-being o f African A m erican couples than for W hite couples. Any analysis o f why African A m erican young m en and w om en are less prone to get m arried in the first place and m ore prone to divorce if u n h a p ­ pily m arried in the second place has to take into account the im portance o f Black extended family netw orks. These netw orks m ight give som e su p p o rt for n o t m a r­ rying if children are bo rn o u t o f wedlock (see Stack, 1974) and for divorcing rather than sticking it o u t w hen there is trouble in the m arriage. H u s b a n d s’ In d e p e n d e n c e a n d C ontrol. The im portance o f independence for W hite m en in m aintaining their m arriage is underw ritten fu rth er in VerofPs (in press) findings that for only W hite m en is n o t being very disclosing to their wives a factor that correlates well w ith their com m itm en t to their m arriages in the seventh year. Such was not the case for the Black husbands, although interdependence was n o t absent in W hite m ales’ com m itm ent. For both Blacks and W hites, Veroff found th at the strongest p red icto r o f com m itm en t was how affirm ed husbands felt by th eir wives. We had previously suggested th at for W hite m arriages there is an ideal o f interdependence alongside the image o f m ale independence. Thus, there no do u b t is a defensive adherence to signs o f independence in W hite husbands that m ay be critical to the balance o f pow er in W hite m arriages. A lthough these find­ ings m ay appear in conflict w ith G ottm an (1994), w ho found that husbands w ho engage in stonew alling (i.e., m en w ho w ithdraw from conflict) are in unhappy m arriages, we m ust rem em ber that his m easure o f w ithdraw al is a behavioral as­ sessm ent. In contrast, Veroffs (in press) self-disclosure m easure is a self-report m easure m ostly dealing w ith talking about problem s. T he m ore husbands th in k they talk about problem s, the m ore they m ay actually have been confronted w ith problem s from their spouse. T h e M e a n in g o f P a ren th o o d . Findings from the Early Years o f M arriage data also su p p o rt the n o tio n that the presence, absence, a n d tim ing o f children have differential effects for Black and W hite m arriages. O rb u ch and Eyster ( 1997b) exam ined the stories th at individual spouses and couples told regarding the expe­ rience o f becom ing a parent. They argued th a t African Am erican couples are m uch less likely to think a b o u t if and when to have children than W hite couples are. In

2.

BLACK COUPI.F.S, W H I T E C O U P L E S

39

contrast, the them es o f w hether the b irth was planned and w hether the couple de­ sired a child w hen they becam e pregnant were salient them es in the b irth n a rra ­ tives o f W hite m arried couples. O rbuch, Eyster, and Veroff (1997) also found th a t p a renthood differentially affects m arital quality for African A m erican and W hite couples. W hite husbands and wives with children in the first year o f m arriage re p o rt less m arital quality in the third year o f m arriage than do those couples w ith o u t children. F urtherm ore, parenthood in the first year leads to a decrease in m arital quality from Year 1 to Year 3 for W hite husbands and wives. T his was n o t the case for Black husbands and wives. In the early years o f m arriage, children have negative effects on the quality o f m arriages for W hites b u t n o t for Blacks. We speculate th a t the m eaning o f children a nd its connection to the psychological w ell-being o f m arriages differ in the two cultures. We argue that any future analysis o f the risk factors associated w ith d i­ vorce and lower m arital quality m ust consider the different m eanings o f p aren t­ hood and children for African A m erican and W hite m arriages.

CONCLUSION Recently there has been a growing interest in discovering w hat m akes a m arriage w ork am d the risk factors associated w ith divorce and low m arital quality. Much o f this literature has been geared tow ard developing universal assum ptions about how to keep a m arriage intact and happy. This vast literature advances the notio n that two sets o f factors affect the psychological well-being and stability o f m a r­ riages over tim e: structural issues connected to the couple’s lives and the inter­ personal o r interactional processes w ithin the m arriage. However, the literatures on African A m erican and on W hite couples vary in th eir em phasis on stru ctu ral or interactive processes. Specifically, studies o f Black m arriages and racial differences in m arital stability are m ore likely to focus on structural processes. Given this general approach in the m arriage literature, the goals o f this chapter were twofold. First, we focused on ways in w hich m arital quality and stability de­ pend on the social context o f race and ethnicity, specifically the cultural norm s and expectations associated w ith being African A m erican o r W hite in A m erican soci­ ety. Second, in exam ining the interactional styles o f relating that m ay be im p o rtan t to predicting w hich m arriages survive and w hich are m ore resilient to stress, we looked for sim ilarities and differences in African A m erican and W hite couples. We spotlighted findings from one unique longitudinal study o f m arriage, the Early Years o f M arriage Project at the University o f M ichigan. O u r inquiry into the risks that are present in unhappy and unstable m arriages identified few inter­ changeable factors b u t m any parallel factors that take on different guises in the two co-cultural contexts. We identified five p ro m in e n t interactional them es that we a r­ gued differentially express them selves in Black or W hite m arriages: (a) supportive wives, (b) interpersonal com patibility, (c) m ale achievem ent and power, (d) inte­

40

O R B U C H , VEROFF, H U N T E R

gration o f social netw orks, and (e) h u sb a n d ’s independence and control. African A m ericans and W hites attach differentia] m eaning to such interactional p h e n o m ­ ena as wives’ n u rtu ran c e o f their husbands, being independent or interdependent w ith one another, sexual relationships in m arriage, m arital com patibility, financial security, husbands’ participation in household activities, resolving conflicts about m oney m atters, and the m eaning or experience o f becom ing a parent. All these phenom ena and issues are significant risk factors for one group or the other, or both. O u r discussion an d interpretation o f these interactional risk factors in Black and W hite couples were inform ed by an exam ination o f gender-pow er dynam ics th at differentially play them selves out in the two cultures. We argue th at the expe­ rience o f being m ale or female has a unique m eaning depending on the cultural social context. T hese m eanings are im p o rta n t to gaining a better understanding o f the m ale-fem ale interactional processes in Black an d in W hite m arriages that p rom ote or inhibit happiness and divorce. As social researchers we need to exam ­ ine m arital interactional processes through b o th cultural and gendered lenses. Similarly, even structural factors c ontributing to m arital risks or resiliencies are best interpreted w ithin a cultural and gendered context. T hus the m eaning o f parental status, incom e, educational background o f m other, an d social netw ork ties is different in the two groups (and for husbands and wives), which leads to dif­ ferential connections to m arital risk in the two groups. It w ould not be a wild extrapolation to suggest th at ju st as there are differential m eanings to these risk variables for the two groups so w ould there be differential m eanings on w hat such variables m ight im ply for the risks faced by children o f these m arriages as their young lives develop. Given that there are gender d if­ ferences in the effects o f divorce and rem arriage on children (D em o & Acock, 1988; H etherington, 1989; O rbuch, T h o rn to n , & C ancio, 1997), these m arital risks m ight also have different im plications for daughters and sons. In addition, our analyses strongly suggest that the im pact o f divorce and unhappy m arriages on children has to be considered w ithin the contexts o f African A m erican and W hite cultures.

ACKNOW LEDGM ENTS T he research in this chapter was supported by N IM H G rant M H41253. T he au­ thors would like to th an k Libby D ouvan, Sandra Eyster, H alim ah Hassan, m em bers o f the Early Years o f M arriage Project Sem inar and the Family Studies Sem inar at the University o f M ichigan, and the editor o f this volum e for their helpful com ­ m ents and suggestions.

2.

BLACK C O U PL E S , W H I T E CO U PL ES

41

REFERENCES Acitelli, L. K., Douvan, E., 8c Veroff, J. (1997). The changing influence of interpersonal perceptions on marital well-being among Black and White couples. Journal of Social and Personal Relationships, 14, 291-304. Ade-Ridder, L., & Brubaker, T. H. (1983). The quality of long-term marriages. In T. H. Brubaker (Ed.), Family relationships in later life (pp. 21-30). Beverly Hills: Sage. Adelmann, P. K., Chadwick, K., 8c Baerger, D. R. (1996). Marital quality of Black and White adults over the life course. Journal o f Social and Personal Relationships, 13, 365-385. Amato, P. R. (1993). Childrens adjustment to divorce: Theories, hypotheses, and empirical support. Journal o f Marriage and the Family, 55,23-38. Amato, P. R., & Keith, B. (1991). Separation from a parent during childhood and adult socioeconomic attainment. Social Forces, 7 0 ,187-206. Amato, P. R., 8c Rodgers, S. T. (1997). A longitudinal study of marital problems and subsequent divorce. Journal o f Marriage and the Family, 59,612-624. Belsky, J., Spanier, G., 8c Rovine, M. (1983). Stability and change in marriage across the transition to parenthood. Journal of Marriage and the Family, 47,455-462. Billingsley, A. (1988). Black families in White America. New York: Simon 8c Schuster. Billingsley, A. (1992). Climbing Jacob's ladder: The enduring legacy o f African-American families. New York: Simon 8c Schuster. Bowman, P. (1988). Research perspectives on Black men: Role strain and adaption across the adult life cycle. In R. L. Jones (Ed.), Black adult development and aging (pp. 117-150). Berkeley, CA: Cobb and Henry. Broman, C. L. (1988). Household work and family life satisfaction of Blacks. Journal o f Marriage and the Family, 50, 743-748. Broman, C. L (1993). Race differences in marital well-being. Journal of Marriage and the Family, 55, 724-732. Carstensen, L. L (1991). Selectivity theory: Social activity in life-span context. In K. W. Schaie (Ed.), An­ nual Review o f Gerontology and Geriatrics (pp. 195-217). New York: Springer. Carstensen, L. L., Gottman, J. M., 8c Levenson, R. W. (1995). Emotional behavior in long-term marriage. Psychology and Aging, 10,140-149. Collins, P. H. (1990). Black feminist thought. Boston: Unwin Hyman. Coon> H. M., 8c Veroft', I. (1996). Marital norms and marital stability: Understanding subcultural differ­ ences. Unpublished manuscript, University of Michigan, Ann Arbor. Cowan, P., 8c Cowan, C. (1994, September). VVTiere s the romance? What happens to marriage when part­ ners become parents. Paper presented at the conference on intimate relationships at Iowa State Uni­ versity, Ames, Iowa. Crohan, S. E. (1992). Marital happiness and spousal consensus on beliefs about marital conflict: A lon­ gitudinal i nvestigation. Journal o f Social and Personal Relationships, 9,89-102. Davis, T. J. (1995). The occupational mobility of Black males revisited: Does race matter. The Social Sci­ ence Journal, 32(2), 121-135. Demo, D. H., 8c Acock, A. (1988). The impact of divorce on children. Journal of Marriage and the Fam­ ily, 50,619-648. Demo, D. H., 8c Acock, A. (1996). Family structure, family process, and adolescent well-being. Journal of Research on Adolescence, 6 ,457-488. Elder, G., 8c Caspi, A. (1988). Economic stress in lives: Developmental perspectives. Journal o f Social Is­ sues, 44,25-45. Epps, E. (1985). Educational and occupational aspirations and early attainment of Black males and fe­ males. Atlanta, GA: Southern Educational Foundation.

42

ORBUCH, VEROFF, HUNTER

Fine, M. A. (1993). C urrent approaches to understanding family diversity: An overview o f the special issue. Family Relations, 42(3), 235-237. G ottm an, J. M. (1994). W hat predicts divorce? Hillsdale, NJ: Lawrence F.rlbaum Associates. H atchett, S., Veroff, J., & Douvan, E. (1995). M arital stability am ong Black and W hite couples in early m arriage. In M. B. Tucker 8c C. M itchell-K ernan (Kds.), The decline in marriage am ong AfricanAmericans (pp. 177-218). New York: Sage. Henderson^King, D. H., & Veroff, J. (1994). Sexual satisfaction and m arital well-being in the first years o f m arriage. Journal o f Social and Personal Relationships, 1 1 ,509-534. H etherington, E. M. (1989). C oping w ith family transitions: W inners, losers, and survivors. Child De­ velopment, 6 0 ,1-24. H etherington, E. M. (1993). An overview o f the Virginia longitudinal study o f divorce and rem arriage with a focus on early adolescence. Journal o f Family Psychology, 7, 39-56. H etherington, E. M., 8c Clingem peel, W. G. (1992). Coping w ith m arital transitions: A family systems perspective. Monographs o f the Society for Research on Child Development, 7(2-3, Serial No. 227). Hossain, Z., 8c Roopnarine, J. L. (1993). Division o f household labor and child care in dual-earner African-Am erican families with infants. Sex Roles, 2 9 ,571-583. H unter, A. G.> & Davis, J. E. (1992). C onstructing gender: An exploration o f Afro-A m erican men's con­ ceptualization o f m anhood. Gender and Society, 6 ,464-479. H unter, A. G ., 8c Sellers, S. L. (1998). Feminist attitudes am ong A frican-Am erican wom en and m en. Gender and Society, 12,81-99. H uston, T. L., & C horost, A. F. (in press). Behavioral buffers on the effect o f negativity o n m arital satis­ faction: A longitudinal study. Personal Relationships. Johnson, L. B. (1990). T he employed Black: T he dynam ics o f work-family tension. In H . E. C heatham and J. B. Stewart (Eds.), Black families (pp. 217-233). N ew York: Transaction. Kane, E. W. (1992). Race, gender, and attitudes toward gender stratification. Social Psychology Quarterly, 55,311-320. Kochm an, T. (1981). Black and W hite styles in conflict. Chicago: University o f Chicago Press. Lichter, D. (1988). Racial differences in underem ploym ent in A m erican cities. American Journal o f Soci­ ology, 93, 771-792. Lichter, D. T., McLaughlin, D. K., Kephart, G., & Landry D. J. (1992). Race and the retreat from m ar­ riages: A shortage o f m arriageable men? American Sociological Review, 5 7 ,781-799. Littlejohn-Blake, S. M., 8c D arling, C. (1993). U nderstanding the strengths o f A frican-Am erican fami­ lies. Journal o f Black Studies, 23, 460-471. M arkm an, H. J. (1991). C onstructive m arital conflict is not an oxym oron. Behavioral Assessment, 13, 83-96. M arkm an, H . J., Renick, M. J., Floyd, F. J., Stanley, S. M ., 8c Clem ents, M. (1993). Preventing m arital dis­ tress through com m unication and conflict m anagem ent training: A 4- and 5-year follow-up. Jour­ nal o f Consulting and Clinical Psychology, 61(1), 001-008. M cLanahan, S., 8c Sandefur, G. (1994). Growing up with a single parent: W hat hurts, what helps. C am ­ bridge, MA: H arvard University Press. O rbuch, T. L., 8c Custer, L. (1995). T he social context o f m arried w om en’s w ork and its im pact on Black husbands and W hite husbands. Journal o f Marriage an d the Family, 5 7 ,333-345. O rbuch, T. L., 8c Eyster, S. L. (1997a). Division o f household labor am ong Black couples and W hite cou­ ples. Social Forces, 76 ,301-332. O rbuch, T. L., 8c Eyster, S. L. (1997b). Becoming a parent: The social context o f story-telling. M anuscript under review. O rbuch, T. L., Eyster, S. L.,& Veroff, J. (1997). The early years o f marriage for Black couples a nd W hite cou­ ples. Paper presented a t the annual m eeting o f the International N etwork o n Personal Relation­ ships, University o f M iami, O xford, O H. O rbuch, T. L., H ouse, ]. S., Mero, R. P., 8c Webster, P. S. (1996). Marita! quality over the life course. Social Psychology Quarterly, 59(2), 162-171.

2.

BLACK C O U P L E S , W H I T E CO U P L E S

43

Orbuch, T.L., Thornton, A., & Cancio, J. (1997). The impact of divorce, remarriage and marital quality on the relationships between parents and their children. Manuscript under review. Orbuch, T. L., Veroff, J., 8c Holmberg D. (1993). Becoming a married couple: The emergence of m ean­ ing in the first years o f marriage. Journal o f Marriage and the Family, 55, 815-826. Seltzer, J. A., & Bianchi.S. M. (1988). Children’s contact with absent pa rents. Journal o f Marriage and the Family, 50 ,663-677. Smith, A. W. (1991). Personal traits, institutional prestige, racial attitudes, and black students’ academic performance in college. In W. Allen, E. Epps, 8c N. Haniff (Eds.), College in Black and White: AfricanAmerican students in predominantly white and in historically black public universities (pp. 111-126). Albany, NY: SUNY Press. Stack, C. B. (1974). All our kin: Strategies for surviving in a black community. New York: Harper 8c Row. Staples, R. (1991). Black male genocide: The final solution. In B. Bowser (Ed.), Black male adolescents: Parenting and education (pp. 39-57). Lanham, MD: University Press of America. Staples, R., 8c Johnson, L. B. (1993). Black families at the crossroads. New York: Jossey-Bass. Sutherland, L., Went, D., 8c Douvan, E. (1990). Traditionalism in gender roles among black and white newlyweds. Unpublished manuscript, University of Michigan, Ann Arbor. Taylor, R. J., Chatters, L. M., 8c Jackson, J. S. (1993). A profile of familial relations among three-genera­ tion Black families. Family Relations, 42,332-341. Taylor, R. J., Chatters, L. M., Tucker, M. B., 8c Lewis, E (1991). Developments in research on Black fam­ ilies: A decade in review. Journal o f Marriage and the Family, 52,993-1014. Taylor, R. J., 8c Zhang, X. (1990). Cultural identity in maritally distressed and non-distressed black cou­ ples. The Western Journal of Black Studies, 14, 205-213. Timmer, S. G., & Veroff, J. (1997). Discontinuity o f divorce: Family ties and the marital happiness of new­ lywed couples form intact and non-intact families. Manuscript under review. Timmer, S. G., Veroff, J., & Hatchett, S. (1996). Family ties and marital happiness: The different marital experiences of black and white newlywed couples. Journal o f Social and Personal Relationships, 13, 335-359. Tucker, M. B., 8c Mitchell-Kernan, C. (Eds.). (1995). The decline o f marriage among African-Americans. New York: Sage. Veroff, J. (in press). Marital commitment in the early years of marriage. In W. Jones 8c J. M. Adams (Eds.), Handbook o f interpersonal commitment and relationship stability. New York: Plenum. Veroff, J., Douvan, L., 8c Hatchett, S. (1995). Marital instability. Westport, CT: Greenwood. Veroff, J., Douvan, E., Orbuch, T. L., & Acitelli, L. K. (1998). Happiness in stable marriages: The early years. In T. N. Bradbury (Ed.), The developmental course of marital dysfunction (pp. 152-179). New York: Cambridge University Press. Vinokur, A. D., 8c Van Ryn, M. (1993). Social support and undermining in close relationships: Their in­ dependent effects on the mental health of unemployed persons. Journal o f Personality and Social Psychology, 65,350-359. Walker, H. A. (1988). Black-White differences in marriage and family patterns. In S. M. Dornbusch 8c M. H. Strober (Eds.), Feminism, children and the new family (pp. 87-112). New York: Guilford. Webster, P., Orbuch, T. L, 8c House, J. (1995). Effects of childhood family background on adult marital quality and perceived stability. American Journal of Sociology 101,404-432. White, L„ 8r Edwards, J. N. (1990). Emptying the nest and parental well-being: Evidence from national panel data. American Sociological Review, 55, 235-242. Williams, D. R., Lavizzo-Mourey, R., 8c Warren, R. C. (1994). The concept o f race and health status in America. Public Health Reports, 109>26-41. Willie, C. V. (1985). Black and White families: A study in complementarity. New York: General Hall.

This page intentionally left blank

PART

II

Child Adjustment in Different Family Forms

This page intentionally left blank

3

Multiple Risks and Adjustment in Young Children Growing up in Different Family Settings A British Community Study o f Stepparent, Single Mother, and Nondivorced Families

Kirby Deater-Deckard University o f Oregon

Judy Dunn Institute o f Psychiatry, London

P a r e n t a l s e p a r a t i o n , divorce, remarriage, and family reconstitution have becom e com m on experiences in children’s and parents’ lives. The idyllic image of the tw o-parent, m ultiple-child nuclear family o f the 1950s has been replaced w'ith a m ore complex yet m ore realistic view o f parenting and families in the 1990s. The intricate networks of family relationships that are form ed, nurtured, and som e­ tim es broken over the lifecourse o f a family that experiences parental separation and remarriage present a complex array o f risks as well as protective factors that are, in part, responsible for the wide range o f individual differences seen in chil­ dren’s post-divorce and post-rem arriage adjustm ent. There remains no doubt that experiencing such family changes during childhood and adolescence carries some liability for em otional and behavioral problem s (Amato & Keith, 1991; Booth & D unn, 1994; Buchanan, Maccoby, & D ornbusch, 1996). However, the m agnitude o f this effect and the diversity in children’s adjustm ent forces us to consider the true complexity of individual and family psychological processes that are im pli­ cated. This a im — the exploration of m ultiple pathways o f risk for children living in different family contexts following parental separation and family reconstitu­ tio n — is the im petus for this chapter.

47

48

DEATER-DECKARD A N D D U N N

In Britain, changes in family structures and rates o f divorce, rem arriage, and non m arital cohabitation m irro r the sam e secular trends in m ost oth er W estern in ­ dustrialized nations. D ata from the Office for N ational Statistics (ONS) in the United Kingdom dem onstrate th at 12.5% o f children (1 in 8) in the UK will, at som e tim e d u rin g childhood or adolescence, live in a household w ith a birth p a r­ ent w ho has form ed a new p a rtnership either through cohabitation or rem arriage (Haskey, 1994). For the m ajority o f children w hose parents divorce, these changes occur early in life. In the ONS analyses, nearly three qu arters o f the children who were living in stepfam ilies had entered this household structure before the age o f 10 years. F urtherm ore, m ost children w ho have parents w ho divorce live in a single­ m o th e r household, at least tem porarily. T hus, increasing n um bers o f children in the UK experience parental separation and rem arriage and the recom binations of sibling relationships that are form ed. In order to understand m ore about the im ­ pact o f these changes on children’s adjustm ent, we explore preschool children’s de­ velopm ental outcom es in different types o f families (e.g., divorced, single parent, rem arried), taking the o p p o rtu n ity to use inform ation from a large-scale co m m u ­ nity study o f families in England, the Avon Longitudinal Study o f Pregnancy and C hildhood (ALSPAC; see G olding, 1996).

RISK MODELS: ACCUMULATION AND EQ U IFIN ALIT Y Risk and resilience m odels o f psychopathology represent a parsim onious a p ­ proach to gaining som e insight in to the distal and proxim al processes th at operate to produce individual differences in children’s post-divorce and post-rem arriage adjustm ent. A nu m b er o f studies have dem onstrated the utility o f these m ultiple and cum ulative risk m odels o f psychopathology (for exam ple, B iederm an et al., 1995; lessor, Van D en Bos, V anderrym , Costa, 8c T urbin, 1995; Liaw 8c BrooksG unn, 1994; Sameroff, Seifer, Baldwin, 8; Baldwin, 1993; Shaw 8c Emery, 1988). We define risk factors as aspects o f the individual and his o r h e r external envi­ ro n m en t that are linked, presum ably in a causal way, to p o o re r social-em otional outcom es (G arm ezy 8c Rutter, 1983). A lthough there are nu m ero u s psychosocial risk factors that have been identified for social-em otional m aladjustm ent in child­ h ood and adolescence, these can be represented in various broad dom ains (DeaterD eckard, Dodge, Bates, 8c Pettit, in press), including b u t certainly n o t lim ited to (a) aspects o f the child; (b) sociocultural factors, including com m unity and neigh­ b o rh o o d contexts; a n d (c) the hom e environm ent, including paren t attributes and parenting behavior. Child risk factors include characteristics o f the child, such as adverse tem pera­ m ent (e.g., irritability; see R othbart 8c Bates, in press), his o r h e r genetic m ake­ up (Plom in, 1994), and being m ale (Zahn-W axler, 1993). These endogenous risk factors are thought to influence the child’s behavior, longitudinally and crosssituationally, an d to interact w ith external o r exogenous risks found in the other dom ains. We em phasize here that by endogenous risk, we m ean neither intractible

3.

M U L T I P L E RI SKS A N D C H I L D R E N ’S A D J U S T M E N T

49

nor determ inistic characteristics, but rather that aspects o f the individual are stable and increase the liability for problem s in adjustm ent. Sociocultural or demographic risks can include poverty (Huston, McLoyd, & Coll, 1994), unstable or adverse household characteristics (e.g., living in a single-m other hom e; Achenbach, How­ ell, Quay, & Conners, 1991), unsafe neighborhoods (B urton, Price-Sparlen, & Spencer, in press), the num ber o f stressful life events faced by the child (Abidin, Jenkins, & McGaughey, 1992), and parental stress and social isolation (DeaterDeckard, in press). Home or fam ily factors include conflict (and in extreme cases, violence) between parents (Abidin et al., 1992; Shaw & Emery, 1988), harsh and negative parenting (Dodge, Pettit, 8c Bates, 1994), conflicted sibling relationships (D unn,Slom kow ski, Beardsall, & Rende, 1994), and parental psychopathology in ­ cluding depression (Downey 8c Coyne, 1990). O f course, these risk factors covary so that on average children w ho have one or two o f these risk factors are likely to have o ther risk factors as well. It is im portant to consider m ultiple risk factors and the effects of these risk processes within the context o f other risk factors because it is possible that p articular sets o f risk factors provide unique or differential predictions of individual differences in children’s social-em otional adjustm ent (Biederman et al., 1995;Liaw8c Brooks-Gunn, 1994). Thus, there are at least three testable and com peting hypotheses about the nature and effects o f m ultiple risk factors. The generic m ultiple-risk hypothesis is that the presence or absence of risk in any o f these three dom ains (i.e., sociocultu­ ral, endogenous, parenting) accounts for all o f the prediction o f children’s adjust­ m ent problems. If this is the case, then these dom ain-specific risks are essentially redundant w ith each other. In contrast, the specific m ultiple risk hypothesis is that the underlying risk pro­ cesses indexed by each of these dom ains o f risk provide unique statistical pre­ diction o f children’s social-em otional adjustm ent. If the latter hypothesis is true, then each dom ain o f risk should independently increase the prediction o f em o­ tional and behavioral problem s. Support for the m ultiple risk hypothesis, however, does not rule out the possibility that these m ultiple, dom ain-specific sets o f risk factors operate in the sam e way to produce similar outcom es, a principle called equifinality (Cicchetti & Rogosch, 1996). We need then also to explore whether particular m ultiple risk pathways explain m ost o f the variance in children’s socialem otional adjustm ent, or alternatively, if each m ultiple risk pathway is associated with similar outcomes. Individual differences in children’s adjustm ent may be predicted not from p ar­ ticular sets o f m ultiple risks b u t from the accum ulation of risk that is independ­ ent o f the presence or absence of particular risk factors. The cumulative risk hy­ pothesis states that the num ber of risk factors, and not necessarily the particular dom ain of risk, is im portant for understanding the underlying risk process (see Sam eroff et al., 1993). Thus, the addition of m ultiple risk factors (regardless of dom ain or the specific content o f the risk factors) contributes to a general level o f stress and adversity for the child, which in turn predicts social-emotional adjustm ent.

50

DF. ATER- DECKARD A N D D U N N

THE ROLE OF FAMILY CONTEXT ON RISK PROCESSES These m ultiple and cumulative risk models can be particularly useful in describing the processes that lead to adjustm ent problem s for children following a parental divorce, separation, or remarriage. These risk processes may operate in a very sim ­ ilar way for all children, regardless of the particular family and broader cultural context w ithin which they are em bedded. Alternatively, risk processes may be specific to particular contexts. For example, it is possible that differential parental treatm ent of siblings plays a m ore im portant role in the developm ent o f emotional and behavioral problem s am ong stepsiblings living in reconstituted families than it does for full siblings in nondivorced families (see Deater-Deckard, D unn, O ’Connor, & Golding, 1997; Mekos, H etherington, & Reiss, 1996).

AIMS These m ultiple risk models to be described here can be used to answer questions about children’s and parents’ psychosocial adjustm ent in the different family con­ texts that arise following parental separation o r divorce and parental remarriage and family reconstitution. In this chapter, we attem pt to answer four questions about the risk processes that may lead to problem s in children’s social and em o­ tional developm ent in stepfamilies and single m other families; 1. Multiple risk. How do the know n risk factors for problem s in children’s social-em otional developm ent work together to produce observed indi­ vidual differences in children’s adjustm ent? 2. Cum ulative risk. Do these risk factors operate in a cum ulative fashion, so that the num ber o f risk factors, rather than the specific content o f these risks, predicts children’s adjustm ent? 3. Equifinality. Are there m ultiple pathways (i.e., different com binations o f risk factors) to similar outcomes, o r are there one or two predom inant pathways to problem s in adjustm ent? 4. Contextualism. Do these m ultiple and cumulative risk models operate similarly or differently in nondivorced, single-m other, stepfather, and stepm other families, and if so, what explains this context “effect” on de­ velopm ental process?

THE AVON LONGITUDINAL STUDY OF PREGNANCY AND CHILDHOOD The Avon Longitudinal Study of Pregnancy and C hildhood (ALSPAC) is an ongo­ ing epidemiological study of nearly 14,000 wom en who were pregnant and gave

3.

M U L T I P L E R I S K S A N D C H I L D R E N ’S A D J U S T M E N T

51

birth to the target children during a 21-m onth period (from April 1,1991, to De­ cem ber 31, 1992) in Avon county, England (see Golding, 1996, for a detailed de­ scription). Approximately 85-90% o f the eligible population o f wom en partici­ pated in the study. This large com m unity sample is representative o f Avon county (which includes the city of Bristol), and the sample also resembles the British pop­ ulation generally, with the exception o f an underrepresentation o f ethnic m inority groups (Baker, M orris, & Taylor, 1997). The rates of stepparent, single parent and intact or nonstep families are similar to those found in the UK population, with a slight underrepresentation o f single m others, m ost likely because the participants were recruited during a pregnancy (O ’C onnor et al., in press). A ttrition has been low— the m ajority (75%) o f the study participants have rem ained in the study through its first 5 years. The m others and their partners completed various ques­ tionnaires about themselves, the target children, and their older siblings, beginning prior to the target child’s birth and continuing throughout the child’s first 4 years o f life. The sample for the analyses discussed here included just over 6,000 families with complete data pertaining to household structure and various sociodem o­ graphic indicators, parenting and the hom e environm ent, and child attributes such as gender, tem peram ent and social-em otional adjustm ent. Target children and their older siblings were classified into four family types, based on the household com position o f biological and step relationships. Nonstep families were those families in which all children were biologically related to both parents living in the home. Stepfather families were those in which at least one older sibling was the biological child of the m other an d a stepchild to her resident partner. These families could also include children who were biologically related to both parents. Stepmother or “other” stepfamilies were those families in which at least one o f the older siblings was a stepchild o f the m other. This category also in­ cluded complex o r blended stepfamilies, where both parents had stepchildren in the hom e. Finally, single-mother families were those in which the target child and older sibling were biologically related to a single m other who had either never m arried or cohabited or who was currently separated or divorced from the chil­ dren’s father. We assessed the target children’s social-em otional adjustm ent when they were 4 years old, using the Strengths and Difficulties Q uestionnaire (SDQ: G oodm an, 1997), a 25-item questionnaire that assesses m ultiple dom ains o f children’s posi­ tive and negative adjustm ent (see G oodm an, 1997, for details regarding reliability and validity). Because we were interested in exploring general functioning, we fo­ cused on the 20-item Total Problems score (scale alpha = .80). This scale includes item s pertaining to children’s em otional problem s, hyperactivity, conduct pro b ­ lems, and peer relationship difficulties. We estim ated children’s risk for adjustm ent problem s using four indicators from each of three dom ains— sociodem ographic risks, parenting, and child (en­ dogenous) risks. All risk factors were based on m others’ reports at various tim e points. All measures had acceptable internal consistency, and composites contain­

52

DEATER- DECKARD A N D D UN N

ing m ultiple item s were com puted by sum m ing items or scale scores that were first standardized.

Sociodemographic Risk Risk factors in this dom ain were assessed when the children were 21 m onths old and included socioeconomic status, neighborhood characteristics, children’s ex­ posure to violence, and ethnic group mem bership. Socioeconomic circumstances (SES) were estimated by sum m ing standardized scores for three indicators of hous­ ing conditions— crowding, hom e ownership (rented versus m ortgaged/ow ned), and type o f hom e (from low-incom e housing, apartm ent, sem i-detached o r d u ­ plex hom e, to detached hom e). Neighborhood risks were com puted by sum m ing four standardized indicators— m others’ concerns about vandalism and property damage, violence (e.g., muggings), burglaries, and disturbances from youths (orig­ inally scored: 0 = no opinion o r not a problem ; 1 = m inor problem ; 2 = m ajor p rob­ lem). Exposure to violence was indexed by a measure of violence between the m other and her partner, represented by sum m ing two standardized item s— hit­ ting o r slapping, and throw ing or breaking objects (originally scored: 0 = no p art­ ner or did not happen; 1 = m other did or partner did; 2 = both m other and partner did). Ethnicity was coded as 0 = ethnic m ajority (Caucasian), and 1 = ethnic m i­ nority (e.g., Afro-Caribbean, African, Black British, Asian). For m ore precise de­ tails about these measures, refer to D unn, Deater-Deckard, Pickering, O 'C onnor, & Golding (in press).

Parenting Risk The four risk factors in this dom ain were assessed at various tim e points and in­ cluded m aternal negativity toward the target child (48 m onths), m aternal depres­ sive sym ptom s (21 m onths), use o f spanking (18 m onths), and parenting stress (21 m onths). Maternal negativity was m easured by sum m ing four standardized items from an 8-item positivity/negativity questionnaire designed to tap m others’ negative and aversive feelings toward the target child (alpha = .62; originally scored as yes or no). M other’s depressive symptoms were assessed using the 10-item Edinburgh Postnatal Depression Scale (EPDS; Cox, Holden, & Sagovsky, 1987), a valid and reliable instrum ent that is applicable outside the postnatal period (Thorpe, 1993). These items assess the com m on sym ptom s o f depression, such as m ood, self-blame, and vegetative states (originally scored: 0 = not at all to 3 = as m uch as I ever did). Use o f spanking or smacking was m easured with the item, “W hen she/he has tem per tantrum s how often do you: sm ack/spank child?” M others responded to this item: 1 = never/no tantrum s, 2 = som etim es, 3 = often. Finally, parenting stress was assessed by sum m ing eight standardized items from a questionnaire designed to measure m others’ feelings about stress and inadequacy in the parenting role (originally scored: 1' = feel never to 4 = feel exactly). Example

3.

M U L T I P L E R I S K S A N D C H I L D R E N ’S A D J U S T M E N T

53

items included “whining makes parent want to hit child,” "having a young child is absolutely exhausting,” and “a m other can feel exasperated when she wants to calm her child down and nothing works” (see Deater-Deckard, 1997, for details about this measure).

Child Risks The four child risk factors included child gender, frequency o f tem per tantrum s (21 m onths),difficult tem peram ent (6 m onths), and illness (6 m onths). Child gen­ der was coded: 0 = girl, 1 = boy. Frequency o f temper tantrums was assessed by a sin­ gle item and coded: I = never, 2 = < once a week, 3 = at least once/week, 4 = most days, 5 = > once/day. Difficult temperament was m easured using a sum m ed com ­ posite o f six tem peram ent scales from a modified version o f the Carey Infant Tem­ peram ent Scale (Carey & McDevitt, 1977). Each of 88 items was coded: 1 - almost never, to 6 = almost always. The difficult temperament com posite was determ ined from results o f a higher order factor analysis o f the nine subscales (factor loadings > .4). The difficult temperament composite included rhythmicity, approachability, adaptability, m ood/em otionality, persistence, and distractibility. Finally, illness was assessed by sum m ing two standardized item s— the health of the child in the first m onth o f life and the health of the child in the past m onth (originally coded: 1 = very healthy, 2 = m inor problem s, 3 = som etim es quite ill, 4 = m osdy unwell).

Analysis Method Hierarchical regression analysis and cluster analysis are useful tools for testing these various com peting hypotheses. To test the com peting generic and specific m ultiple risk hypotheses, risk factors w ithin the three specified dom ains (i.e., soci­ ocultural, endogenous, parenting) can be forced to enter the regression equation as sets o f predictors. Through an iterative process using different hierarchical order­ ing of risk factors, the overlapping and unique prediction from each dom ain of risk can be estimated. The cumulative risk hypothesis can be tested by adding the num ber o f risk factors for each child to a hierarchical regression model. If the indi­ vidual risk factors provide unique increm ents to problem s in the children’s adjust­ m ent beyond the num ber-of-risks variable, the content of risk rather than solely the accumulated num ber o f risks faced by the children, m ust be considered as im ­ p o rtant to the process.

Multiple Risk Factors The bivariate correlations between the 12 risk factors and childrens SDQ Total Problem scores are shown in Table 3.1. We estim ated these correlations for the whole sample, as well as separately for each family type, in order to describe pat­ terns in the covariation o f m ultiple risks and child adjustm ent across family type. T he sociodem ographic risk factors were modestiy correlated with child adjust-

54

DEATER-DECKARD A N D D UN N

TABLE 3.1 Bivariate Correlations Between Multiple Risk Factors and 4-year-old Children’s Strengths & Difficulties Questionnaire Total Scores, by Family Type All Families (n = 6,022)

Nonstep (n= 4,710)

Stepfather (n = 654)

-.15*** .13*** .09*** .01

-.12**' .12*** .08*** .00

-.16**’ .17*** .13*** .05

Parenting: 5. negativity 6. depression 7. spanking 8. parenting stress

.48*** .26*** .11*** .29***

.48**’ .26*** .10*** .29***

.47*’ * .22*** .13*** .25***

.49*** .13 .19* .26**

.52*** .23*** .1 2 " .30*"

Child: 9. gender1* 10. tantrums 11. temperament 12. illness

.09’ " .23*** .21**' .15” '

.09*** .22*** .21*** .15’ "

.05*’ * .25'** .20*** .10***

.17 .22* .26** .19*

.11** .24*** .19"* .17***

Sociodemographic: I.SES 2. neighborhood 3. violence 4. ethnicity*

Stepmother (n = 107)

-.14 .01 .11 ___ C

Single (n= 551)

-.2 0 " * .1 2 " .09* -.03

Note. *p< .05. **p< .01. ***p< .001. n = minimum sample size based on pairwise deletion o f missing data. *0 - ethnic majority, 1 = ethnic minority. b0 - girl, 1 = boy. cNo variance in ethnicity for this group.

m cnt. For the whole sam ple, children w ith higher SDQ Total Problem s scores were from lower SES families, living in m ore dangerous neighborhoods, and had higher a m o u n ts o f exposure to violence in the hom e. E thnicity was unrelated to SDQ Total scores. Overall, this pattern o f correlations was very sim ilar for children in all four family types. The parenting and hom e risk factors were m odestly to substantially correlated w ith child adjustm ent. For the w hole sam ple, children w ith higher SDQ scorcs had m ore negative relationships w ith their m others, had m others w ho had m ore de­ pressive sym ptom s, were m ore likely to have been punished physically, and had m others w ith higher levels o f parenting stress. Again, this p attern o f covariation betw een parenting risk factors and SDQ adjustm ent scores was very sim ilar for children across all four family types. T he characteristics o f the children were also related to the children’s SDQ Total Problem s scores. C hildren w ith m ore behavior and em otional problem s were m ore likely to be boys and to have had histories o f tem per tan tru m s, difficult tem ­ p eram ent, and physical illness. As w ith the sociodem ographic and p arenting risk dom ains, the pattern o f the correlations betw een child adjustm ent and child risk factors was sim ilar for children in nonstep, stepparent, and single-m other families.

3.

M U L T I P L E R I SKS A N D C H I L D R E N ’S A D J U S T M E N T

55

We used hierarchical m ultiple regression analyses to estim ate the total explained variance in children’s SDQ Total scores and the unique statistical prediction from each o f the three dom ains o f m ultiple risk factors. The unique prediction for each d om ain was estim ated by entering the dom ain o f interest on the last step in the equation. For exam ple, to estim ate the unique prediction o f the parenting risk fac­ tors, predictors were entered into the equation in two steps: (a) sociodem ographic and child risk factors; (b) parenting risk factors. We conducted these analyses for the whole sam ple as well as separately w ithin each o f the four family types. Overall, statistical prediction o f children’s SDQ Total scores was very good, explaining 29% (p < .001) o f the variance for the w hole sam ple. The full m odel predicted SDQ scores equally well in all four family types: nonstep, 29%; stepfa­ ther, 28%; stepm other, 29%; single m other, 32%. The relative predictive weight o f each dom ain varied substantially. For the en­ tire sam ple, p arenting and hom e risk factors had the largest unique prediction o f children’s SDQ Total scores, explaining 18% o f the variance. C hild risk factors ex­ plained only 3% o f the unique variance and sociodem ographic risks only 1% or the unique variance for the w hole sam ple. Thus, 21 % o f the explained variance was a ttributable to unique com ponents in these three dom ains o f risk. Subtracting this from the 29% total explained variance, about 7% was attributable to predictive variance that was n o t unique (e.g., overlapping variance) to any particular dom ain o f risk. This pattern was quite consistent across the four fam ily types. Parenting and hom e risk factors accounted for 15% to 20% o f the u nique variance in children’s SDQ scores in the four family types, followed by 2% to 5% for child risks, and less than 1% to 2% for sociodem ographic risks. Given th at the total explained variance was sim ilar across the four family types, the a m o u n ts o f overlapping predictive variance betw een the three dom ains were also nearly identical across the family types (6 to 7%). To sum m arize, each dom ain o f m ultiple risk factors provided unique statistical prediction o f children’s SDQ Total scores, although the vast m ajority o f this unique prediction was found in the parenting and hom e dom ain. Furtherm ore, som e, but by no m eans the m ajority, o f the explained variance was overlapping betw een the three dom ains o f m ultiple risk factors. Finally, these m ultiple risk m odels operated in a very sim ilar way in all four family contexts. T his suggests that, on average, the risk processes that are im plicated in the developm ent o f behavioral and em otional problem s in the preschool period are the sam e in families w here parents have d i­ vorced and in those w here they have not; they are sim ilar too in tw o-parent and single-parent families.

Cumulative Risk Factors Next, we developed a cum ulative risk m odel based on the sam e set o f m ultiple risk factors identified in the previous analyses. In order to c onstruct the cum ulative risk

DKATER-DECKARD A ND D U N N

56

m odel, we dichotom ized each c ontinuous risk factor by selecting those cases 1 SD above the whole sam ple m ean as being at risk. This included about 15% o f the whole sam ple on average, although for any given variable the extrem e gro u p could include as little as 9% o f the whole sam ple o r as m uch as 28% o f the whole sam ple, depending on the distribution characteristics (i.e., skewness and kurtosis) o f the given variable. Two risk factors were n o t dichotom ized using a threshold score be­ cause they were nom inal-level variables: child gender (boys being at greater risk) and ethnicity (ethnic m inorities being at greater risk). Table 3.2 includes the prevalences (i.e., percentages o f the sam ple) o f each o f the 12 risk factors w ithin the three risk dom ains for the w hole sam ple as well as w ithin each o f the four family types. W ith the exception o f gender and ethnicity, these risk percentages are arbitrary, so the com parison o f interest is n ot the actual prevalence o f risk w ithin the whole sam ple or any subgroup o f families b u t betw een each of the four family types. In o rder to test w hether the prevalence o f these risk factors varied as a function of fam ily type, we m ade all possible com parisons (six in total) using x2 tests. All group com parisons that were significant at p < .05 o r less are show n in Table 3.2. In general, across the sociodem ographic and p arenting risk dom ains, nonstep families had the lowest prevalence rates, stepfather and stepm other families had TABLE 3.2 Prevalence of Multiple Risk Factors as Percentage of Sample, by Family Type

Sociodemographic: l.lowSES 2. neighborhood 3. violence 4. ethnicity* Parenting: 5. negativity 6. depression 7. spanking 8. parenting stress Child: 9. gender4 10. tantrum s 11, temperament L2. illness

All

Nonstep

Stepfather

Stepmother

Single

Tukey tests (p < .05)

13 11 10 2

9 10 9 1

25 15 13 2

19 15 9 1

35 20 10 4

N S= .05). In particular, although beneficial to both, authoritative p arenting was m ore beneficial to adolescents from intact hom es th an from single-parent hom es. TABLE 4.7 Impact of Neglectful Parenting on Adolescent Adjustment Adjustment Indices

Whole sample Intact (4,037) Single (1,037) Ecological niches White, working class Intact (1,192) Single (294) White, middle class Intact (1,207) Single (195) Black, working class In ta c t(120) Single (147) Black, middle class Intact (52) Single (65) Asian, working class In ta c t(289) Single (46) Asian, middle class In ta c t(291} Single (36) Hispanic, working class Intact (304) Single (96) Hispanic, middle class Intact (50) Single (21)

Distress

Self-Esteem

GPA

Delinquency

Substance Use

.0751* .0636*

-.1568* -.1728*

-.1549* -.1218*

.1965* .1766*

.1750* .1761*

.0797* .0665

-.1187* -.2141

-.1259* -.1205*

.1714* .1289*

.1999* .2303*

.1539* .1203*

-.1995* -.1903*

-.2077* -.2724*

.2461' .2675*

.2445* .2423*

.0698 .0693

-.1866* -.2371*

-.1721* -.0811

.1329 .1455

.2077* .0778

-.2375* .0427

-.1137 -.0998

.0783 -.0567

.2379 .4233*

.0549 .2815*

.0397 .0800

-.0784 -.1887

- . 1055' -.1487

.1978' .1217

.1145* .1780

.0366 .3021*

-.1415* .2239

-.0909 -.0026

.1628* .1824

.1687* .2638

.0908 -.0315

-.1 200' .0314

-.2204* -.0510

.2051* .1975

.1809* .1424

.2796* -.0213

-.1849 -.3533

-.1986 -.4960*

.1203 .2542

.3905* .3435

*/>< .05; Bold type indicates that the effect sizes differ statistically at p< .05.

84

A VE N EV OL I, S ESSA, S T E I N B E R G

A u th o r ita r ia n P aren tin g . Table 4.5 displays results o f the com parisons o f effect sizes betw een intact and single-parent hom es across ecological niches. In a few cases, a u th o ritarian p arenting is m ore harm ful in one family type than in the other. First, parental au th o ritarian ism is related to greater psychological distress in Asian m iddle-class and H ispanic working-class, single-parent families than in in ­ tact families. Second, a u th o ritarian parenting is m ore dam aging to the self-esteem o f adolescents from single-parent than from tw o-parent families am ong W hite working-class, African A m erican working-class, A frican A m erican m iddle-class, and Asian w orking-class families (n o t all are statistically different). The opposite is tru e for W hite m iddle-class families: A uthoritarianism is m ore d etrim ental to the self-esteem o f adolescents from intact families than those from single-parent families. In term s o f school grades, au th o ritarian parenting is significantly m ore harm ful to the grades o f adolescents living in intact, African A m erican, m iddle-class hom es th an adolescents living in single-parent, m iddle-class hom es. O n the oth er hand, auth o ritarian parenting is significantly m ore harm ful to the grades o f Asian w ork­ ing-class adolescents living in single-parent versus intact hom es. In term s o f a d o ­ lescent delinquency, it is especially notable that the relations betw een a u th o rita ri­ anism and delinquency are in the opposite direction for intact and single-parent African A m erican w orking-class families. O nly one o f the p lanned com parisons o f effect sizes in intact versus single­ parent families w ithin ecological niches reached the .05 level o f significance. In particular, we found that a uthoritarianism is m ore harm ful in term s o f psycholog­ ical distress and self-esteem and less beneficial in term s o f delinquency and su b ­ stance use to the overall adjustm ent o f adolescents from H ispanic working-class, single-parent hom es (z = 1.6, p = .05). P erm issive P arenting. M irroring the trends o f parental authoritarianism , parental perm issiveness appears to be m ore beneficial in one family type over the o th er only w ithin H ispanic working-class families o r only w ithin specific o u t­ comes. N e g le c tfu l P a ren tin g . In only one com parison of effects did the relation be­ tween neglectful p arenting and adolescent adjustm ent differ betw een intact and single-parent hom es. In Asian m iddle-class families, neglectful p arenting is related to lower adolescent self-esteem in intact hom es b u t to higher self-esteem in single­ p aren t hom es. P lanned com parisons indicated that neglectful parenting is m ore d etrim ental to overall adolescent adjustm ent in intact H ispanic working-class families than in single-parent families (2 = 1.87, p = .03) and is m ore d etrim ental to overall adjust­ m ent in African A m erican m iddle-class, single-parent families versus intact fam i­ lies ( 2 = 2.00, p = .02).

4.

P A R E N T I NG AND A DO L E S CE N T A D I U S T M E N T

85

DISCUSSION AND CONCLUSIONS Adolescents in contem porary society experience a broad range o f contextual influ­ ences that affect their psychosocial adjustm ent and behavior. In this chapter, we exam ined how some of these contextual variables (i.e., family structure, ethnicity, and social class) m oderate the influence of parenting on adolescents’ em otional and psychosocial functioning. The results o f this research both validate and ex­ pand previous research in three general areas. First, our findings confirm that adolescents experience different parenting styles in tw o-parent versus single-parent homes. Regardless of ethnicity and social class, tw o-parent families tend to be m ore authoritative, m ore authoritarian, less permissive, and less neglectful than single-parent families. Although the differ­ ences in parenting style between tw o-parent and single-parent families are gener­ ally small, they are consistent with other investigators’ observations that single p ar­ ents exhibit dim inished parenting, especially with respect to behavioral control and lim it-setting (i.e., m ore permissive and less involved parenting; Hetherington et al., 1982, 1989; Wallerstein, 1983). Indeed, the m ultiple dem ands and stresses (e.g., working and caring for children) placed on single parents may contribute to their decreased capacity to provide sufficient m onitoring and structure for their children. In contrast, the com bined effort of two parents in intact homes, who of­ ten share child rearing responsibilities, may be additive in term s o f how it con­ tributes to an increased capacity for active and involved parenting. Second, o u r findings suggest that, with only a few exceptions, the relation be­ tween parenting style and adolescent adjustm ent is m ore similar than it is different across ecological contexts. Consistent with the results o f previous research, two broad generalizations were confirm ed in the present study. First, we find that au­ thoritative parenting is related to adaptive adolescent adjustm ent virtually regard­ less of family structure, ethnicity, or social class (see Steinberg et al., 1991). Spe­ cifically, adolescents from authoritative hom es are well-adjusted emotionally, do well in school, have healthy self-esteem, and are less likely to engage in deviant behavior. In contrast, we find that neglectful parenting, characterized by parental disengagement, is associated w ith em otional distress and deviance, low self­ esteem, and poor academic achievem ent across all ecological niches. Across m ost ecological contexts, authoritarian parenting is associated with psychological dis­ tress, low self-esteem, and poor school perform ance. Permissive parenting, al­ though associated with m ore adaptive adolescent developm ent than authoritarian parenting, is still less effective than authoritative parenting. T here are two notable exceptions to the pattern o f relations between parenting and adjustm ent across ecological contexts, however. First, am ong African Am eri­ can middle-class adolescents, authoritative parenting has a m odestly less bene­ ficial influence on their self-esteem and academic perform ance and is actually

86

AVENEVOLI, SESSA, S TE IN B ER G

associated with dim inished psychological well-being and greater involvement in deviant and delinquent behaviors. The interpretation o f this finding is unclear, given the general absence o f research or theory on parenting in this particular eco­ logical niche. It is possible that African American m iddle-class parents are at­ tem pting to practice a som ewhat different, and perhaps less effective, form of authoritative parenting, one that combines two som ewhat contradictory orien­ tations to socialization. O n the one hand, m em bership in the m iddle class may incline these African Am erican parents toward the use o f psychologically based discipline techniques such as reasoning and explanation, which are fundam ental com ponents of authoritative parenting and widely touted as elements o f good parenting by middle-class child-rearing experts (Hoff-Ginsberg 8c Tardif, 1995). O n the other hand, however, the African American heritage condones the use of stricter and harsher approaches to punishm ent, including physical discipline (Garcia Coll, Meyer, 8c Brillon, 1995). In the effort to blend these sets o f practices within an overall climate o f authoritativeness, the effectiveness o f this sort o f par­ enting may be dim inished. The second exception to the general pattern o f relations between parenting and adjustm ent is found in the dom ain o f school perform ance. The relation between parenting, characterized as authoritative, authoritarian, or permissive, and GPA is inconsistent across ecological contexts. For example, although authoritative par­ enting is associated with good perform ance in school for alm ost all adolescents, the benefits of authoritative parenting on school perform ance are greater am ong adolescents from particular ethnic and social class backgrounds (i.e., am ong White, working- and middle-class youth). In addition, authoritarian and perm is­ sive parenting are associated with relatively poorer school perform ance in m ost niches bu t with better school perform ance in other adolescents (e.g., with perm is­ siveness am ong Hispanic m iddle-class youth and w ith authoritarian parenting am ong Asian middle-class youth). A final general conclusion is that, although family structure appears to m oder­ ate the relation between parenting and adolescent adjustm ent w ithin some ecolog­ ical niches (e.g., African American and Hispanic w orking class), its m oderating in­ fluence generally holds little practical significance. Specifically, in all cases in which differential effects o f parenting on adolescent adjustm ent were found, the effects varied only in m agnitude and not in direction. T hat is, when differences in the link between parenting and adjustm ent are found in tw o-parent versus single-parent homes, these differences do not suggest that one style o f parenting is beneficial in one family structure but harm ful in another. Rather, the results suggest that a par­ ticular parenting style simply may be relatively m ore beneficial in one group than in another. For example, authoritative parenting buffers against involvement in delinquent behavior am ong African American middle-class adolescents in both tw o-parent and single-parent homes. However, there is a relatively stronger re­ lation between authoritative parenting and adolescent delinquent behavior in single-parent homes com pared to two-p'arent homes.

4.

PARENTING AND ADOLESCENT ADJ USTME NT

87

As with all research, the present study has som e im portant lim itations. First, due to concerns over having an adequate sample size w ithin particular ecological niches, we com bined divorced, widowed, and never m arried parents in our single­ parent family structure category. Previous research indicates that it is im portant not only to distinguish am ong types of single-parent families (e.g., H etherington, 1972) but to quantify the tim e since separation in cases of divorce (e.g., Wallerstein, C orbin, & Lewis, 1988). Related to this issue, we did not examine parenting and adjustm ent in families transform ed by remarriage. The presence of a step­ parent can influence the parenting style experienced by the adolescent and conse­ quently that child’s adjustm ent. Because we partitioned our sample into ethnic and social class ecological niches, however, there was not sufficient sample size to examine variations w ithin the single-parent category or to include rem arried fam ­ ilies in the study. Second, our data are lim ited by the self-report nature o f the measures. Because all of the data come from adolescents’ self-reports, we can only say that youngsters who characterize their parents in certain ways show particular patterns of behav­ ior and psychological functioning. W hat this may suggest is that youngsters’ sub­ jective experience of parental behavior is an im portant influence on their own de­ velopm ent and well-being. Although it may be inform ative to investigate whether parents’ actual behavior toward their adolescents is associated in similar ways to adolescent adjustm ent, we do not believe that objective assessments of parental be­ havior are the only valid indicators of what takes place in the family. In our view, both subjective and objective m easures of parenting provide an im portant window on children’s experience in their families. In conclusion, what can we say about parenting and adolescent adjustm ent in tw o-parent and single-parent families? O ur results indicate, consistent w ith previ­ ous research, that single parents are less likely than m arried parents to use the warm, firm, and dem ocratic parenting style that we know is associated with m ore adaptive adolescent developm ent. However, it is im portant to bear in m ind that this difference in parenting style between single- and tw o-parent families is rela­ tively small in m agnitude and that family structure, per se, exerts a relatively m o d ­ est effect on parenting style. Furtherm ore, there is likely considerable variation in authoritativeness within the population o f single-parent families; we suspect that less effective or dim inished parenting would be m ost evident during the tim e when the family is transform ed from a tw o-parent to a single-parent family. The m ost im portant conclusion from this set o f analyses, however, is that fam ­ ily structure does not m eaningfully m oderate the relation between parenting and adjustm ent across a wide variety o f socioeconomic and ethnic groups. Adolescents from one- o r tw o-parent hom es who experience authoritative parenting are better adjusted than peers w ho experience any o f the other parenting styles traditionally studied in socialization research. Virtually regardless o f ethnic or social class back­ ground, across one- and tw o-parent hom es alike, adolescents whose parents are warm and dem ocratic and who provide limits and structure report less psycholog-

88

AVENEVO LI, SESSA, S TE INBE RG

ical distress and m ore p ositive self-esteem , perform better in sch ool, and engage in less d elin q u en t b ehavior than peers.

REFERENCES Amato, P. R.> 8c Keith, B. (1991). Parental divorce and the well-being of children: A meta-analysis. Psy­ chological Bulletin, 110,26-46. Aseltine, R. H. (1996). Pathways linking parental divorce with adolescent depression. Journal o f Health and Social Behavior, 3 7 ,133-148. Baldwin, A., Baldwin, C., & Cole, R. E. (1990). Stress-resistanl families and stress-resistant children. In J. E. Rolf, A. S. Masten, D. Cicchetti, K. N. Wechterlein, & S. Weintraub (Eds.), Risk and protective fac­ tors in the development o f psychopathology (pp. 257-280). New York: Cambridge University Press. Baumrind, D. (1972). An exploratory study of socialization effects on Black children: Some BlackWhite comparisons. Child Development, 43,261-267. Baumrind, D. (1991). Parenting styles and adolescent development. In J. Brooks-Gunn, R. Lerner, & A. C. Petersen (Eds.), The encyclopedia o f adolescence (pp. 746-758). New York: Garland. Bronfenbrenner, U. (1979). The ecology of human development: Experiments by nature and design. Cam­ bridge, MA: Harvard University Press. Bronfenbrenner, U. (1986). Ecology of the family as a context for human development. Developmental Psychology, 22,723-742. Bronfenbrenner, U. (1989). Ecological systems theory. Annals of Child Development, 6, 187-249. Bronfenbrenner, U., 8c Crouter, A. C. (1983). The evolution of environmental models in developmental research. In W. Kessen & P. H. Mussen (Eds.), History, theory, and methods: Handbook o f child psy­ chology (Vol. 1, pp. 357-414). New York: Wiley. Chao, R. K. (1994). Beyond parental control and authoritarian parenting styles: Understanding Chinese parenting through the cultural notion of parenting. Child Development, 6 5 ,1111-1119. Children’s Defense Fund. (1995). The state o f America's children yearbook. Washington, DC: Author. Cohen, J. (1977). Statistical power analysis for the behavioral sciences. New York: Academic. Deater-Deckard, K., Dodge, K. A., Bates, J. E., & Pettit, G. S. (1996). Physical discipline, among African American and European American Mothers: Links to children’s externalizing behaviors. Develop­ mental Psychology, 3 2 ,1065-1072. Dornbusch, S. M., Carlsmith, J. M., Bushwall, S .}., Ritter, P. L., Leiderman, H., Hastorf. A. H., & Gross, R. T. (1985). Single parents, extended households, and the control of adolescents. Child Develop­ ment, 56,326-341. Dornbusch, S. M., Ritter, P., Liederman, P., Roberts, D.,8c Fraleigh, M. (1987). The relation of parenting style to adolescent school performance. Child Development, 58, 1244-1257. Emery, R. E. (1982). Interparental conflict and the children of discord and divorce. Psychological Bul­ letin, 92, 310-330. Emery, R. E. (1988). Marriage, divorce, and childrens adjustment. Newbury Park, CA: Sage. Featherman, D., Spenner, K., 8c Tsunematsu, N. (1988). Class and the socialization of children: Con­ stancy, change, or irrelevance. In E. M. Hctherington, R. Lerner, 8c M. Perlmutter (Eds.), Child de­ velopment in life-span perspective (pp. 67-90). Hillsdale, NJ: Lawrence Erlbaum Associates. Forehand, R., Long, N., & Brody, G. (1988). Divorce and marital conflict: Relationship to adolescent competence and adjustment in early adolescence. In E. M. Hetherington & J. D. Arestah (Eds.), Im­ pact o f divorce, single parenting, and slepparenting on children (pp. 135-154). Hillsdale, Nj: Lawrence Erlbaum Associates. Forgatch, M. S., Patterson, G. R., 8< Skinner, M. L. (1988). A mediational model for the effect of divorce on antisocial behavior in boys. In E. M. Hetherington 8c j. D. Arestah (Eds.), Impact o f divorce, single parenting, and stepparenting on children (pp. 135-154). Hillsdale, Nj: Lawrence Erlbaum Associates.

4.

PARENTING AND ADOLESCENT ADJUSTM ENT

89

Furstenberg, F. F. (1990). Coming of age in a changing family system. In S. S. Feldman 8c G. R. Elliot (Eds.), A t the threshold: The developing adolescent (pp. 147-170). Cambridge, MA: Harvard Univer­ sity Press. Garcia Coll, C., Meyer, E. C„ 8c Brillon, L. (1995). Ethnic and minority parenting. In M. Bornstein (Ed.), Handbook of parenting: VoL 2. Biology and ecology of parenting (pp. 189-209). Hillsdale, NJ: Law­ rence Erlbaum Associates. Glick, P. C., 8c Lin, S. (1986). Recent changes in divorce and remarriage. Journal o f Marriage and the Familyj 48, 737-739. Gold, M. (1970). Delinquent behavior in an American city. Belmont, CA: Brooks/Cole. Greenberger, E., Steinberg, L., 8c Vaux, A. (1981). Adolescents who work: Health and behavioral conse­ quences of job stress. Developmental Psychology, 17,691-703. Guidubaldi, J., Cleminshaw, H. K., Perry, J. D., Nastasi, B. K., 8c Lightel, J. (1986). The role of se­ lected family environment factors in children’s post-divorce adjustment. Family Relations, 35,141— 151. Hernandez, D. J. (1988). Demographic trends and the living arrangements of children. In E. M. Hether ington 8c J. D. Arasteh (Eds.), Impact of divorce, single parenting, and stepparenting on children (pp. 3-22). Hillsdale, NJ: Lawrence Erlbaum Associates. Hetherington, E. M. (1972). Effects o f father absence on personality development in adolescent daugh­ ters. Development Psychology, 7,313-326. Hetherington, E. M. (1989). Coping with family transitions: Winners, losers, and survivors. Child De­ velopment, 6 0 ,1-14. Hetherington, E. M., Cox, M., & Cox, R. (1982). Effects of divorce on parents and children. In M. Lamb (Ed.), Nontraditionalfamilies (pp. 233-288). Hillsdale, NJ: Lawrence Erlbaum Associates. Hetherington, E. M., Stan ley-Hagan, M., 8c Anderson, E. R. (1989). Marital transitions: A child’s per­ spective. American Psychologist, 44, 303-312. Hoff-Ginsbcrg, E., 8c Tardif, T. (1995). Socioeconomic status and parenting. In M. Bornstein (Ed.), Handbook o f parenting: Vol. 2. Biology and ecology o f parenting (pp. 161-188). Hillsdale, NJ: Law­ rence Erlbaum Associates. Lin, C. C., 8c Fu, V. R. (1990). A comparison of child-rearing practices among Chinese, immigrant Chi­ nese, and Caucasian-American parents. Child Development, 61,429-433. Maccoby, E., 8c Martin, J. (1983). Socialization in the context of the family: Parent-child interaction. In P. H. Mussen (Series Ed.) 8c E. M. Hetherington (Vol. Ed.), Handbook o f child psychology: Vol. 4. So­ cialization, personality, and social development (4th ed., pp. 1-101). New York: Wiley. McLanahan, S., 8c Sandefur, G. (1994). Growing up with a single parent. Cambridge, MA: Harvard Uni­ versity Press. McLoyd, V. (1990). The impact of economic hardship on Black families and children: Psychological dis­ tress. Parenting, and socioemotional development. Child Development, 6 1 ,311-346. Nye, F. I. (1957). Child adjustment in broken and in unhappy unbroken homes. Marriage and Family Living, 19, 356-361. Ogbu, J. U. (1981). Origins of human competence: A cultural-ecological perspective. Child Develop­ ment, 52,413-429. Parke, R. D., 8c Buriel, R. (1998). Socialization in the family: Ethnic and ecological perspectives. In W. Damon (Series Ed.) & N. Eisenberg (Vol. Ed.), Handbook of child psychology: Vol. 3: Social, emo­ tional, and personality development. New York: Wiley. Patterson, G., 8c Stouthamer-Loeber, M. (1984). The correlation of family management practices and delinquency. Child Development, 5 5 ,1299-1307. Radloff, L. S. (1977). The CES-D scale: A self-report depression scale for research in the general popula­ tion. Applied Psychological Measurement, 1,385—401. Rodgers, R. R. (1966). Cornell Parent Behavior Description— An interim report. Unpublished m anu­ script, Cornell University, Ithaca, NY. Rosenberg, M. (1965). Society and the adolescent self-image. Princeton: Princeton University Press.

90

AVENEVOLI* SESSA, S TE I N B E R G

Rosenthal, R., & Rosnow, R. (1984). Essentials o f behavioral research: Methods and data analysis. New York: McGraw-HilL Rosnow, R., 8c Rosenthal, R. (1989). Statistical proceduresand the justification of knowledge in psycho­ logical science. American Psychologist, 44, 1276-1284. Schaefer, E. (1965). Children’s reports of parental behavior: An inventory. Child Development, 36, 413—424. Sorrentino, C. (1990). The changing family in international perspective. Monthly labor Review, 113, 41-58. Steinberg, L. (1990). Interdependency in the family: Autonomy, conflict, and harmony. In S. Feldman 8c G. Elliot (Eds.),/If the threshold: The developing adolescent (pp. 255-276). Cambridge, MA: Harvard University Press. Steinberg, L., Dornbusch, S. M., & Brown, B. B. (1992). Ethnic differences in adolescent achievement: An ecological perspective. American Psychologist, 47, 723-729. Steinberg, L., Mounts, N. S., Lamborn, S. D., 8c Dornbusch, S. M. (1991). Authoritative parenting and adolescent adjustment across varied ecological niches. Journal o f Research on Adolescence, /> 19-36. Wallerstein, J. S. (1983). Children of divorce: Stress and developmental tasks. In N. Garmezy & M. Rut­ ter (Eds.), Stress, coping, and development in children (pp. 265-302). New York: McGraw-Hill. Wallerstein, J. S., Corbin, S. B., 8c Lewis, J. M. (1988). Children of divorce: A 10-year study. In E. M. Hetherington 8c J. D. Arasteh (Eds.), Impact o f divorce, single parenting, and stepparentingon children (pp. 197-214). Hillsdale, NJ: Lawrence Erlbaum Associates. Weinberger, D., Tublin, S., Ford, M., 8c Feldman, S. (1990). Preadolescents’social-emotional adjustment and selective attrition in family research. Child Development, 61, 1374-1386. Weinmann, L , Steinberg, L., 8c Dornbusch, S. M. (1990). Divorce, remarriage, and adolescent adjustment. Unpublished manuscript, Temple University. Weston, S., 8c Weston, W. (1987, October). Education and the family. Mimeo available from the Office of Educational Research and Improvement, U.S. Department of Education, Washington, DC.

PART I I I

Family Functioning and Child Adjustment in Divorced and Single-Parent Families

This page intentionally left blank

5

Should We Stay Together for the Sake of the Children? E. M av is H e th e rin g to n University o f Virginia

o f t h e m o s t com m only asked questions ab o u t divorce is w hether it is b etter to stay in an u nhappy conflictual m arriage for the sake o f the children o r to get divorced. In attem p tin g to answ er that question researchers have used two m ain strategies. First, a few investigators have com pared the adjustm ent o f chil­ dren w hose parents later divorce to those w ho do n o t divorce. The assum ption in these studies is that m arital conflict and dysfunction will have affected the adjust­ m en t o f children before divorce and th a t m any o f the adverse effects on children’s adjustm ent attrib u ted to divorce are related to earlier m arital conflict. C hildren w hose parents later divorce show m ore problem s th an those in families th at later rem ain intact; however, even with these predivorce levels o f adjustm ent controlled there are significant differences in the adjustm ent o f children in divorced and no n divorced families (A m ato & Booth, 1996; A m ato & Keith, 1991a; Block, Block, & G jerde, 1986; C herlin et al., 1991). A second strategy involves exam ining the adjustm ent o f children in high-conflict nondivorced families and in divorced families. Although there is a substantial liter­ ature on the effects o f m arital conflict and divorce on children’s adjustm ent, there are few studies th a t directly com pare children in nondivorced families with acri­ m onious m arital relations w ith those in divorced families. M oreover, those that m ake this com parison often use m arital dissatisfaction as a proxy for m arital con­ flict (e.g., Sim m ons & Associates, 1996), although m arital conflict is a better pre­ dictor o f children’s adjustm ent th an is general m arital distress o r dissatisfaction (Em ery & O ’Leary, 1984; Johnson 8c O ’Leary, 1987). T he few studies that have m ade direct com parisons betw een high-conflict nondivorced families and divorced fam ­ ilies suggest th at the changes and stresses accom panying divorce and life in a single­ p arent, usually m other-headed, household may co n trib u te m ore adversely to cer­

O ne

93

94

HETHERINGTON

tain dom ains o f child adjustm ent than does an unhappy or conflictual m arriage (A m ato & Booth, 1996; Sim m ons et al., 1996). However, these studies do n o t co n ­ sider the level o f m arital conflict in divorced families, although conflict betw een divorced spouses often rem ains high o r escalates follow ing divorce (M accoby & M nookin, 1992; M asheter, 1991). It is im p o rta n t to assess the experiences and quality o f relationships in the divorced families as well as in nondivorced families, and the changes in risks and resources o r protective factors preceding and follow ­ ing divorce. T he response to divorce is influenced by the quality o f family relations in the predivorce m arriage, the circum stances o f m arital dissolution, and the experiences and changes th at follow divorce. It involves an interplay betw een individual char­ acteristics o f parents and children, family relationships, a n d extrafam ilial factors that serve to su p p o rt or to u nderm ine the well-being o f family m em bers as they negotiate the changes and challenges associated w ith divorce. A djustm ent to d i­ vorce is a process that takes place over tim e. In the first year following divorce, n o ­ table disru p tio n s in family roles and relationships, in parenting, and in children’s adjustm ent occur. Restabilization o f the family and a new hom eostasis in family functioning usually are attained 2 to 3 years following divorce, accom panied by im provem ent in p a ren t-ch ild relations and in the adjustm ent o f children. Two o f th e m ain challenges confronting parents following divorce are to m in i­ m ize the a m o u n t o f conflict to which the child is exposed and to m aintain a u th o r­ itative parenting characterized by high w arm th, com m unication, responsiveness, control and m onitoring, and low coerciveness. The effects on child adju stm en t of other stressors often accom panying divorce are frequently m ediated o r m oderated by the quality o f parenting.

THREE STU D IES OF MARITAL TRAN SITIO N S In this chapter the effects on children’s adjustm ent o f conflict and p arenting in nondivorced an d divorced families are exam ined. The data are taken from three longitudinal studies o f divorce and rem arriage. In the Virginia Longitudinal Study o f Divorce an d Remarriage (H etherington, 1993, in press), children whose parents had divorced o r w ho were in nondivorced families were exam ined in six waves from age 4 to young adulthood. A lthough the study started w ith 144 families, it was expanded in subsequent waves an d by young adulth o o d included 450 families equally divided between nondivorced, divorced, and rem arried families. T he sec­ o n d study, the H etherington and Clingem peel study o f divorce and rem arriage (H etherington & Clingem peel, 1992), exam ined the effects o f m arital transitions on early adolescent adjustm ent and included 202 nondivorced, divorced, o r re ­ cently rem arried families studied three times. Finally, the N onshared E nvironm ent study (Reiss et al., 1994) studied 720 nondivorced families an d stepfam ilies and two sam e-sex adolescent sibling pairs in each family twice, 4 years apart.

5.

S H O U L D W I STAY T O G E T H E R ?

95

All studies used m ultim easure, m ultim ethod, m ulti-inform ant assessments in­ cluding observations to gather inform ation on family functioning and children’s adjustm ent. Assessment o f com m on constructs and m any shared measures were used in all studies, often perm itting the com bination of data from the three stu d ­ ies. This chapter includes measures o f m arital conflict; parenting including warm th, negativity/coercion/conflict, m onitoring and control, authoritativeness and parentifkation; and child adjustm ent including externalizing, internalizing, cognitive agency, social responsibility, social competence, self-esteem, and associations with delinquent peers. Some analyses include families and offspring draw n from all three studies, whereas others include data from only the Virginia Longitudinal Study; therefore, sample sizes vary widely. All o f the divorced families are m othercustody families. This chapter begins w ith a brief discussion o f child adjustm ent in high-conflict nondivorced and divorced families, followed by an exploration of some family processes associated w ith the adjustm ent o f children in these families.

THE ADJUSTMENT OF CHILDREN IN HIGH-CONFLICT NONDIVORCED FAMILIES AND IN DIVORCED FAMILIES O n average, children in divorced families in com parison to those in nondivorced families, and children in families with high m arital conflict in com parison to those with low m arital conflict, show m ore problem s in adjustm ent (see Am ato & Keith, 1991a; Davies & Cum m ings, 1994; H etherington & Stanley-Hagan, 1997, in press, for reviews). The type o f adjustm ent problem s in children associated with m arital conflict and with divorce are similar, with the largest effects obtained with exter­ nalizing disorders, lack o f self-regulation, low social responsibility, and cognitive agency, and to a lesser extent with internalizing, social agency, and self-esteem. Difficulties in relations with parents, siblings, peers, and teachers also are associ­ ated with both m arital conflict and divorce. Some o f these childhood problem s are sustained into adolescence; however, adolescence may trigger behavior problem s in children in divorced families who previously have appeared to be functioning reasonably well (H etherington, 1993). Recent work suggests that problem s in adjustm ent may remain in young adult­ hood, long after parental divorce has occurred; these m ay include problem s in re­ lationships in the workplace and with family and spouses, higher rates o f marital instability and divorce, m ore behavior problems, lower socioeconomic attain­ m ent, and lower reported well-being (Am ato & Keith, 1991b; Amato, Loomis, & Booth, 1995; Booth & Amato, 1994; H etherington, in press; McLanahan & Sandefur, 1994). Although on the average there is about a twofold o r greater increase in pro b ­ lems such as externalizing disorders, delinquency, teen pregnancy, truancy, school dropout, substance abuse, idleness, and welfare dependency in adolescents from

HETHERINGTON

96

divorced families in com parison to nondivorced families (H etherington, 1993; H etherington & Clingempeel, 1992; McLanahan 8c Sandefur, 1994; Sim m ons & Associates, 1996), the vast m ajority o f children from divorced families eventually emerge as reasonably com petent individuals (H etherington, 1993; H etherington & Clingempeel, 1992). The increase in total behavior problem s in high-conflict, n on­ divorced families is som ewhat lower than those in divorced families, but the m a­ jority o f these offspring also are resilient and eventually able to cope with their adverse life situations. This emphasis on resiliency is not m eant to m inim ize the distress, apprehen­ sion, confusion, and anger children experience in response to m arital conflict or divorce. Marital conflict and divorce are reported by children and parents to be two o f their m ost stressful life events (Davies & Cum m ings, 1994; H etherington, 1993). It is m eant to underscore the adaptability and resiliency o f children and parents in coping with stressful life transitions.

Factors Contributing to Adjustment Divorce may offer children and parents an escape from conflict and the o p portu­ nity for m ore fulfilling harm onious personal relationships; however, it also m aybe associated with an increase in risk factors that make successful parenting and the developm ent o f com petent children m ore difficult. Both parents and children re­ port higher rates o f negative life events in divorced than in nondivorced families. Am ong the negative experiences associated with divorce are econom ic declines leading to m ultiple changes in residence, loss of friends, moves to poorer neigh­ borhoods with fewer resources, inadequate schools and delinquent peer groups, lack o f social supports, family conflict, parental depression, inept nonauthoritative parenting, and loss or dim inished contact with the noncustodial parent. All o f these variables have been associated w ith adjustm ent problem s in children follow­ ing divorce. In this chapter the focus is on family relations and parenting and how they differ in high-conflict nondivorced families and in divorced families because in m ost o f our work we have found that these variables m ediate o r m oderate many of the effects o f other stressors on child development. The effects of conflict, authoritative parenting, and parentification are dis­ cussed. Authoritative parenting is parenting that is high on w arm th, involvement, responsiveness, com m unication, control and m onitoring, and low on negativity and coercion. Parentification occurs when there is role reversal in supportive caretaking relations and the child takes on instrum ental o r em otional support of the parent. C onflict Repeated exposure to angry, acrim onious, unresolved m arital conflict may lead to heightened em otional arousal, sensitization to conflict, an inability to regulate negative affect and behavior, and a lack of em otional security in children (Davies 8c

5.

97

S H O U L D W E S TAY T O G E T H E R ?

C um m ings, 1994; G rynch & Fincham , 1993). Conversely, children m ay learn to regulate their em otions an d solve interpersonal problem s by observing their p a r­ ents successfully negotiate m arital conflict a n d solve problem s in the conjugal rela­ tionship. All children are n o t equally vulnerable to the effects o f m arital conflict, and all types o f m arital conflict are n o t equally corrosive to children’s adjustm ent. Intelligent, socially m atu re children w ith an easy tem peram ent and an internal locus o f control are better able to deal b o th w ith parental conflict and w ith divorce (H etherington, 1989, 1991). In o u r w ork we have found that preadolescent boys are m ore vulnerable to the adverse effects o f both conflict and divorce but that gender differences are found less often in adolescence (H etherington, 1989,1991, 1993). Table 5.1 show s the correlations o f different types o f m arital conflict and the child’s cognitions o r feelings about the conflict w ith total behavior problem s on the C hild Behavior Checklist (Achenbach 8c Edelbrock, 1983) for 100 fifteen-yearolds from nondivorced families and 100 from divorced m other-custody families from the Virginia L ongitudinal Study o f Divorce a n d Rem arriage. All conflict is n o t the sam e in its consequences for child adjustm ent. Encapsulated conflict to w hich the child is n o t directly exposed has no effect on children’s well-being.

TABLE 5.1 Correlations Between Different Typesof Marital Conflict andTotal Behavior Problems inAdolescents Nondivorced (n = 100)

Typeof marital conflict 1. Encapsulated II. Direct Themes Economic Household tasks Personal behavior ofspouse Childbehavior or discipline Strategies Symbolicaggression Physical aggression Reasoningnegotiation Degreeofresolution III. Child response Guilt Shame Threat Caught in themiddle Disengagement

.20* .36**

.33** .52**

.21* .41'* -.25** -.28"

.24* .53*’ -.29** -.37"

.27** .31“

.36** .46**

Onlysignificant correlations arereported. .05. **p< .01

Note. *[}
LC-ND B>G HC-ND>LC-ND LC-ND>D> HC-ND G>B LC-ND>HC-ND G>B LC-ND>HC-ND LC-ND>HC-ND

100

HETHERINGTON

A sim ilar pattern emerges in parenting with parents in high-conflict nondivorc­ ing families show ing m ore inept parenting than those in low-conflict families. The m eans for parents in divorcing families were interm ediate but often not signifi­ cantly different from the other two groups. M others in high-conflict nondivorcing families, in com parison to those in lowconflict nondivorcing families, are lower in warm th, m onitoring, and control and higher in negative coercive parenting (see Table 5.3). M others who later divorce also are higher in negativity and lower in control than those in low-conflict n o n ­ divorcing m arital relationships, but they have greater control than those in highconflict nondivorcing families. Fathers in high-conflict nondivorcing families show lower positivity and con­ trol and m ore negativity than those in low-conflict families (see Table 5.4). Fathers who later divorce also exhibit less warm th than those in low-conflict nondivorcing

TABLE 5.3 Mothers' ParentinginCouples WhoWill Later Divorceand Those inHigh- and Low-Conflicl NondivorcingFamilies

Divorcing (D) Boy

Girl

High-Conflict Nondivorcing (HC-ND) Boy

Girl

Lcnv-Conflict Nondivorcing (LC-ND) Boy

Significant Contrasts

Girl

Positivity/Warmth Negativity/Conflict

-.11 -.04 .16 .05

-.28 -.14 .21 -.01

.07 .10 -.08 -.25

Monitoring

-.14

.03

-.36 -.15

.06 .14

Control

-.13

.05

-.29 -.13

.13 .17

LC-ND>HC-ND D,HC-ND>LC-ND B>G LC-ND>HC-ND G>B LC-ND>D> HC-ND

TABLE 5.4 Fathers ParentinginCouplesWhoWill Later Divorceand Those inHigh- andLow-Conflict NondivorcingFamilies

Divorcing (D) Boy

Girl

High-Conflict Nondivorcing (HC-ND) Boy

Girl

Low-Conftia Nondivorcing (LC-ND) Boy

Girl

Positivity/Warmth

.09 -.04

-.05 -.22

.45 .36

Negativity/Conflict

.02 .14

.11 .29

-.13 -.01

-.16 .03 .02 .08

-.17 -.04 -.15 -.06

-.20 .07 .14 .11

Monitoring Control

Significant Contrasts

LC-ND>HC-ND HC-ND:B>G HC-ND>LC-ND HC-ND:G>B G>B LC-ND>D> HC-ND

5.

S H O U L D W E STAY T O G E T H E R ?

101

families. T his was the only difference betw een the divorcing group and either of the other groups found for fathers. Som e gender differences also were obtained. Both m others and fathers m o n ito r girls’ behavior m ore than boys’ behavior, and m others have m ore conflictual rela­ tions w ith boys than with girls. In addition, un d er high m arital conflict, fathers are less w arm and m ore negative to girls than to boys. This is a finding wc tend to see m ore often w ith younger children than w ith adolescents and concurs w ith the w ork o f Phil Cow an and his colleagues (Cow an, Cow an, & Schultz, 1996). In sum m ary, in com parison to low -conflict families, children in high-conflict families are less well adjusted and their parents exhibit less authoritative, m ore conflictual and coercive parenting. There are som e scattered differences in the a d ­ ju stm en t o f children an d in the quality o f p arenting in harm o n io u s nondivorcing families and those in later divorcing families, b u t m any fewer th an those w ith conflictual nondivorcing families. This finding suggests that before the divorce the family dynam ics leading to divorce are less deleterious th an those associated w ith extrem ely high levels o f overt m arital conflict. C on flict a n d C h ild r e n ’s A d ju s tm e n t A fte r D ivorce. T he few studies that have com pared conflictual non divorced families w ith divorced families ignore the level o f conflict in the divorced families. We tried to rectify th is situation by exam ­ ining the adjustm ent o f children in high- and low-conflict divorced and n o n d i­ vorced families. F urtherm ore, we exam ined the divorced families in the first 2 years after divorce an d 2 or m ore years after divorce w hen considerable restabilization in divorced families had occurred. Figures 5.1 and 5.2 show som ew hat different pictures for the adju stm en t of boys and girls follow ing divorce. In the first 2 years follow ing divorce, a higher pe r­ centage o f girls in b o th high- and low -conflict divorced families score above the clinical cutoff on the Child Behavior Checklist th an those in high-conflict n o n ­ divorced families, and fewer girls in the low-conflict nondivorced families score above the cu to ff than in any o f the oth er three groups. In the first 2 years following divorce, a greater percentage o f boys from high-conflict divorced families and a lower percentage o f boys from low -conflict nondivorced families are above the clin­ ical cutoff than in the other groups. The low -conflict divorced a n d high-conflict nondivorced groups are not significantly different from each other. The divorce per se, regardless o f the level o f conflict initially, seem s to be affecting the girls, whereas boys are responding differentially b o th to divorce an d to the level o f con­ flict w ithin the divorced families. Two years after divorce, conflict in a divorced family has m ore adverse effects o n b o th boys and girls than conflict in a n o n ­ divorced family. However, the adjustm ent o f girls in low-conflict nondivorced and divorced families is similar, whereas m ore boys in low-conflict divorced families are still show ing serious problem s than those in low-conflict nondivorced families. If conflict is going to continue, it is better for children to rem ain in an acrim o ­ nious tw o-parent household than to divorce. If there is a shift to a m ore h arm onious

102

HETHERINGTON Less than 2 y ears after di voice

HCD [

LCD

IIC N -D

T w o w m ore years after divorce

LC N -D

HCD

LCD

H C N -D

L.CN-D

» H » jh C iy ilk i D n o a c d ______ o t o w C onftcl Dtxorccd ______ a H qjt Cofifbci N oft-dnorecd _____ p t ow C onflct N on-dnorccd __ j

FIG. 5.1. Percentageof girlsscoringabove theclinical cutoffon total behavior problems. Less than 2 years after divorco

HCD

i

LC D

■ High Conflict Divorced

HCN -D

Two or m ore years a lte r divorce

L.CN-D

B Low Conflict D ivorced

HCD

LCD

H C N -D

□ High C o r J k t Kun-divo/CCd

LCN -D

□ L ow Conflict N on-divoiccd

H1G.5.2. Percentageof boysscoringabove theclinical cutoffon total behavior problems. household a divorce is advantageous to both boys an d girls. However, even with low acrim ony betw een the divorced couple, boys in divorced families are disadvantaged over those in nondivorced families. T he gender differences m aybe a ttributable to boys being m ore sensitive to the relative unavailability o f fathers and to the fact that in these m other-custody families divorced m others and preadolescent boys are m ore likely to be involved in difficult coercive paren t-ch ild relationships.

5.

S H O U L D W E STAY T O G E T H E R ?

103

C o n flic t a n d P arenting. The previous discussion o f m arital conflict and child adjustm ent is oversimplified. T he m arital relationship is often viewed as the cornerstone o f good family functioning, and certainly a positive m arital relation­ ship is associated w ith authoritative parenting and positive p a ren t-ch ild and sib­ ling relationships (Davies & C um m ings, 1994). In contrast, m arital conflict is as­ sociated w ith low parental w arm th and m onitoring, w ith ineffective, inconsis­ tent control, and w ith high negativity, coerciveness, and conflict in p a ren t-ch ild and sibling relationships (Cow an et al., 1996; Davies & C um m ings, 1994; Sim ­ m ons & Associates, 1996). The key to the link betw een m arital conflict and child adjustm ent seems to be parenting. The direct effects o f m arital conflict on child a d ­ ju stm en t are usually found to be sm all o r nonexistent (Davies & C um m ings, 1994). T hey largely are m ediated or m oderated by the quality o f parenting. For younger children the effects o f parenting on adjustm ent are direct and pow er­ ful. For adolescents a substantial p o rtio n o f the effects o f p arenting are m ediated by the peer group, as discussed in a later section o n m odeling in this chapter. An ANOVA involving family type (high-conflict divorced, low-conflict divorced, high-conflict nondivorced, low-conflict nondivorced), sex o f the child, and m o th er and father authoritative parenting com bination was perform ed on scores for total behavior problem s on the C hild Behavior Checklist (Achenbach & Edelbrock, 1983). Significant m ain effects were obtained for family types, gender, and p a ren t­ ing com bination. However, these effects were qualified by com plex higher order interactions. Table 5.5 presents the m eans and significant contrasts for boys and girls w ith four different com binations o f parenting in high-conflict and low -conflict d i­ vorced and nondivorced families. If authoritative parenting, characterized by high w arm th, responsiveness, com m unication, and m on ito rin g can be m aintained, the effects o f m arital conflict are greatly reduced. In nondivorced families there is a protective effect against the adverse effect o f m arital conflict o f even one p arent being authoritative b u t a greater effect o f b o th being authoritative. M oreover, u n d er high conflict, girls in nondivorced families benefit m ore from authoritative m others than fathers. In divorced families, however, the p arenting o f the custodial m o th e r is m ore salient than th at o f the noncustodial father for b o th boys and girls. U nder conditions o f high conflict betw een the divorced spouses and a n o n a u th o ritative custodial parent, an authoritative noncustodial p arent is an ineffective b u f­ fer. U nder low conflict, an authoritative father can to som e extent protect sons but n o t daughters from the adverse consequences o f the m o th e r’s n onauthoritative parenting. This is p a rt o f the reason why a significant association betw een fre­ quency o f visitation by the noncustodial paren t and child adjustm ent is rarely found. T he effects o f visitation depend on the circum stances o f visitation and the qualities o f the visiting parent. U nder conditions o f low-conflict visitation by a reasonably w ell-adjusted (i.e., nondepressed, non-antisocial), authoritative n o n ­ custodial p arent can have benefits for the adjustm ent o f the child, and these bene­ fits vary w ith the sex o f the child and parent. U nder these conditions authoritative

104

HETHERINGTON

TABLE 5.5 Total Behavior Problems inAdolescents in High- and Low-Conflict Nondivorced and Divorced Families Under Different ParentingCombinations Nondivorced (N D ) Authoritativeness o f Parent

High conflict Boys(B) Girls (G) Lowconflict Boys (B) Girls(G)

MJFJ MJFf

Divorced M other Custody (D)

MfFJ MTF)

MJFJ MJFT MJFJ MfFT

62.7 59.5

54.2 52.4

56.6 46.1

42.1 40.0

74.2 67.9

69.4 64.3

58.2 54.7

54.6 53.9

56.1 53.2

41.9 44.1

44.2 42.1

37.4 35.3

66.7 58.5

56.8 57.8

47.3 43.9

40.5 39.5

Significant contrasts B>G HC> LC D> ND MJFJ >M|F|,M|F| >MfFf HC-D> HC-ND HC-ND-B: MJFJ >MJF|, MjFJ >M|Ff

IIC-ND-G: MJFJ >MJFf >MfF| >MfFf LC-ND: MJFJ >MJFf, MfFJ >MfFf HC-D: MJFJ, MJFf >MfFJ,MfFf LC-D-B: MJFJ >MJFf >MfFJ >MfFf LC-D-G: MJFJ., MJFf >MfFJ.Mf, FJ LC-B: D-MJFJ, MJFf >ND-MJFJ.MJFf LC-G: D-MJFf >ND-MJFf

noncustodial fathers are associated w ith decreased antisocial behavior and higher social responsibility in sons b u t have no effect on the adjustm ent o f daughters. T h e R ela tio n A m o n g M a rita l C onflict, P arental A n tiso c ia l B ehavior, P arenting, A n tiso c ia l Peers, a n d E x tern a lizin g in A dolescents A m ediational m odel o f factors that contribute to externalizing in adolescent boys and girls in divorced and nondivorced families was exam ined w ith structural equation m odeling. Figure 5.3 presents the m odel for fathers and externalizing in boys (N = 138 nondivorced, N = 123 divorced families), Fig. 5.4 for fathers and girls ( N = 136 nondivorced, 121 divorced families), Fig. 5.5 for m others and boys (N = 139 nondivorced, 118 divorced families), and Fig. 5.6 for m others and girls {N = 136 nondivorced, 116 divorced families). D otted lines indicated nonsignificant paths, and solid lines significant paths. Ideally m others and fathers w ould have been included in the sam e m odel, but this was precluded by sam ple size. In a d d i­ tion, because o f sam ple size, m ultiple indicators o f a c onstruct were com posited, and all constructs were treated as observed rather than as latent variables. T he x 2 were satisfactory and the goodness o f fit for all m odels were .97 o r above. In all m odels for both m others and fathers and boys and girls in divorced and nondivorced families the effects o f m arital conflict are indirect and are m ediated by increases in negative coercive relations w ith children and declines in a u th o rita ­ tive parenting. Parental antisocial behavior show s a sim ilar b u t usually sm aller re­ lationship w ith parenting; however, fathers’ antisocial behavior in nondivorced

KICaL 5 J. M jriiJ cor.fltcc. foihtrs* Afitksocul twh*¥iar. pAcerMing, aiuiic-zliJ pet:s, jrui externJinnf; in adotnccnl boy*. Nffltdn'urcrd frir.iin arr on left diwrced Fu nilin are on riphl. Doited lino- ire

iiiMfpi tfindri |u l ) v 'j K .n5

F IG . 5 .4 ,

M a r it a l c c m ilic l. f a t h e r s 'j n l i n x i z l b e h a w r . p a r e n l u ^ j r a i w K i a l p * e r s . u k * c if c r n a lr c in K

in ir t k f c H C iK i « l » - N O t ld f r O r t t d

On M l ; d i't o r t & d fM fii| i« . * 1? a n rifc^l D o l t a l l i n f * * ? ?

i n i i g n i l k a n t p i'. t a mp < JOG.

105

f K r. 5 1

MarHa’ co n flict, m o th e r? ' anfcisodaJ behavior, p a te n tin g . jn n s o c i J p e frs . a n d rc re m a lir

irj in aclnlrKfnl hcrp. ' 4gndiwc'i»v'd fa-rnilin 11; nn lrfh fnQrced farnilip; gr« OH rigbl- nvnlt^j linn ill* trxbgitifLcacI ^iItJL mp K .0-5-.

FIC^. 'i t i

M j r i la l -: ::itl:iU, imilfscr:. aril u.-:i: iaJ ItftuVKir, J u r f r iin g , a r iliiiu lii jsrtjjL I n d ei£ .01;

***/>< .001.

readiness: Cognitive Stim ulation was significantly, and W arm th m arginally signifi­ cantly, predictive o f higher school readiness scores. Are the specific dimensions o f parenting im portant regardless o f the num ber o f other protective factors? Does the num ber ofprotective factors continue to be o f im por­ tance above and beyond specific parenting behaviors? In a final set o f analyses, we asked w hether the specific dim ensions o f paren tin g behavior co ntinue to be im ­ p o rta n t even controlling for the n u m b er o f o th er protective factors present. In the sam e way, we asked w hether the n u m b er o f protective factors continues to predict developm ent even controlling for the influence o f specific dim ensions o f p a ren t­ ing. For these analyses, we deleted from the protective factors sum m ary the vari­ ables related to relationships w ithin the family (i.e., the Q uality o f R elationship a n d Father Involvem ent variables). F indings presented in Table 9.5 show th a t w hen all the variables o f interest are considered sim ultaneously, H arsh Discipline and the protective factors su m m ary co ntinue to be uniquely predictive o f children’s behavior problem s. For the Bracken school readiness assessm ent, the coefficients for Cogn itive Stim ulation and the p ro ­ tective factors su m m ary approach b u t d o not reach statistical significance.

216

ZASLOW

F T A I..

T A B L E 9. 5

Regression Coefficients for Unique Effect of Parenting and Protective Factors on Child Development Outcomes, Controlling for Baseline Characteristics Predictor Variable

Behavior Problems Index

Covariates Experimental/Control group status Child age Philadelphia site Portland site Number of children Child gender Race/ethnicity

1.77 .01 -.47 .23 -.09 -1.20 5.03**

Parenting Composites Harsh Discipline Warmth Cognitive Stimulation

2.60** .78 .12

Protective Factors Index* Protective Factors Summary Total R2

-1.96*** .16***

Bracken

-.64 .03 -.16 -1.50* -1.32** .50 -.1 2 .14 .50 .43 .23 .13**'

Source: New Chance data from baseline enrollment form, 18-m onth follow-up survey, coded observational study data from 21 -m onth visit, and 42-month follow-up survey. Note. Sample size = 247. All variables were entered simultaneously in this model; thus coefficients represent the unique effect of each factor. Statistical significance is indicated as: *p< .05, **p< .01, ***/?< .001. Coefficients are unstandardized. Race/ethnicity was coded so that 0 = Black and 1 = NonBlack. 'The Protective Factors Index used in this analysis excludes the category for positive parent-child relationships.

D ISCU SSIO N T his study provides a particularly stringent exam ination o f the protective factors perspective for the developm ent o f children in stressful circum stances due to the n a tu re o f the present sam ple. In m any studies o f risk a n d resilience, although fam ­ ilies can all be characterized as facing stress, there is variation in term s o f such key characteristics as extent o f poverty, m arital status, and m aternal education. In the present study, all m o th ers had dropped o u t o f school, all were receiving Aid to Fam ilies w ith D ependent C hildren at the start o f the study, all were teenagers upon the b irth o f th eir first children, and nearly all were u nm arried. O u r findings indicate th at even in this extrem ely disadvantaged sam ple, fam i­ lies nevertheless have im p o rta n t resources that can buffer th eir children from stress. D espite the relative hom ogeneity in th eir econom ic circum stances, these families varied substantially in term s o f the n u m b er o f protective factors that were present. F urtherm ore, protective factors functioned cum ulatively, such that chil­

9.

PR O TEC TIV E

FA C T O R S IN L O W - I N C O M E C H I L D R E N

217

dren’s developm ent in both the behavioral and cognitive dom ains was m ore posi­ tive am ong children whose families had m ore protective factors. We also see clear affirm ation for the view that children draw upon resources at m ultiple levels o f the ecological m odel. In o u r analyses, protective factors from each o f the four levels or categories that we considered (i.e., child characteristics, m aternal functioning, parent-child relationships, and external supports) emerged as significant individual predictors o f developm ent. Indeed, only a single variable from am ong the eight we considered in the protective factors sum m ary failed to serve as a significant predictor: T he m easure o f father involvement predicted nei­ ther childrens behavior problem s n o r school readiness. Why m ight it be the case, in the present analyses, that father involvement was unrelated to the child outcom es, particularly given previous research (Greene & M oore, 1996) pointing to the im portance of father involvement in welfare fam i­ lies? The present study focused on adolescent m others, whereas the previous re­ search focused on a sample o f m others w ho were all 20 years o f age or older. In a d ­ dition, the associations shown previously between father involvement and child outcom es were docum ented in analyses w ithin a single tim e period (i.e., associ­ ations were contem poraneous), whereas the present analyses are longitudinal. It is possible that for the young m others in the present sample, who were themselves undergoing im portant developm ental transitions (Q uint & Egeland, 1995), the re­ lationships with the children’s fathers are also changing over time. Perhaps concur­ rent but not longitudinal predictions would have shown associations with the child outcom es under these circumstances. The work o f Greene and M oore (1996) also suggests that rather than sum m ing across all form s o f father involvement, as we have done here, it may be im portant to look separately at the com ponents of visitation, inform al financial support, and the provision o f child support. Finally, it may be that we need to examine measures o f the quality o f father involvement in com bination with m easures o f the extent o f father involvement, and we m ay re­ quire father report data in addition to o r instead of m other report data. In our analyses, we see that child characteristics and the quality o f the m o th e rchild relationship were im portant to developm ent in both the cognitive and be­ havioral dom ains, whereas m easures o f the m others’ functioning and of the broader social context emerged as im portant only to the children’s behavioral adjustm ent. As suggested by previous research, w ithin this sample, protective factors appear to function som ewhat differentially for differing outcom es o f interest. Dion (1997) noted that research on the developm ent o f children in poverty tends to bypass the possible influence o f child characteristics, despite the fact that poverty may increase the prevalence o f such problem s as chronic health conditions and conditions that indicate risk for attentional disorders. The present analyses, working w ithin the protective factors framework, affirm the im portance o f taking child characteristics into account. It is especially interesting that positive sociabil­ ity predicted school readiness and no t just behavioral adjustm ent, and similarly that attentiveness predicted both outcom es.

218

ZASLOW

ET AL.

We have noted that parenting behavior may serve a pivotal role am ong the dif­ ferent protective factors, integrating the influences o f such variables as child char­ acteristics and social support. In the present analyses, we see m ultiple indications o f the im portance o f parenting behavior, or rather m ore specifically, m aternal behavior with the child. Like the child characteristics, the variable describing the dyadic quality o f the m other-child relationship predicted both developmental outcom es. Further, we saw that all three specific dim ensions of parenting we con­ sidered, Cognitive Stim ulation, W arm th, and Harsh Discipline, were significant predictors o f developm ent. Finally, even in analyses that controlled for back­ ground characteristics and the presence o f other protective factors, parenting be­ havior emerged as a significant predictor. Am ong the specific dim ensions, the absence o f harsh interactions emerged as particularly robust in these analyses controlling for other protective factors. This is in agreem ent with the conceptual perspective and research of McLoyd (McLoyd, 1990; McLoyd, Jayaratne, Ceballo, & Borquez, 1994), in which harsh treatm ent is viewed as a key m eans by which parental distress in response to econom ic depriva­ tion is conveyed to the child. This finding has em erged in work ranging from con­ sideration o f the im plications o f the Great Depression for children, to work on poverty am ong rural families, and as McLoyd has noted, specifically in research on African American families in poverty. These descriptive findings may be used as a resource for considering how to en­ hance the developm ent o f children facing the double stressors o f poverty and birth to a teenage m other. They raise the possibility that it may be im portant to address m ultiple levels in the ecological m odel in order to enhance children’s developm ent. A lthough previous interventions for adolescent m others in poverty have been clear as to the im portance o f enhancing parenting behavior and providing broader social support to the young m others, interventions have not routinely screened the children for early indications o f attention deficit o r other learning disorders for which early intervention could be extremely im portant. Similarly, few interven­ tions have provided m ental health com ponents that are in keeping w ith the perva­ siveness o f depression docum ented in samples o f welfare families or the evidence o f the severity of sym ptom s in som e m others. O ur findings m ay also be used to target those aspects o f interventions focusing on parenting behavior. A critical first step may be to work with young m others on reducing harsh and punitive inter­ actions. Indeed, this may be a prerequisite to increasing the supportiveness and stim ulation in interactions (Egeland 8c Zaslow, 1995). Finally, although the findings o f this study can be seen as a source o f optim ism in that they reveal im portant sources o f strength for children in extremely disad­ vantaged families, we m ust conclude with a sobering caveat. T he children in the New Chance sam ple were at high risk for problem s in developm ent. For example, the m ean score on the Bracken m easure o f school readiness for the children in this sam ple corresponds to the 16th percentile nationally (with only 12.4% o f the chil­ dren scoring above the national average). A lthough the presence o f m ultiple pro­

9.

PR O TEC TIVE

FA C TO R S IN

LO W -IN C O M E C H ILD R E N

219

tective factors was clearly and strongly predictive o f higher scores on the m easure o f school readiness, for the children in the present analyses, it did n o t assure devel­ o p m en t at the level o f natio n al norm s. T he 31 children in o u r sam ple w ith six to eight protective factors, on average, still scored only at the 32nd percentile on the m easure o f school readiness. T his is double the average percentile score for chil­ dren in the observational study sam ple, b u t it does not nearly m eet the average percentile (50% ) for the national standardization sam ple. T hus, it is critical to th in k n o t only o f the resources already available w ithin these families that can buffer the children’s developm ent b u t also o f how to strengthen an d b roaden the set o f protective factors.

ACKN O W LEDGM ENTS T his chapter was w ritten w ith funding from the Family and C hild W ell-Being Re­ search N etw ork o f the N ational Institute o f C hild H ealth a n d H u m an D evelop­ m ent (N IC H D G rant 1 U01 HD 3093 0-01). T he New C hance O bservational Study was carried o u t w ith fu n d in g from th e Foundation for Child D evelopm ent, the W illiam T. G ran t F oundation, and an anonym ous funder. T he interdisciplinary team for the New C hance O bservational Study was assem bled by the M anpow er D em onstration Research C o rp o ra tio n (M D RC). We gratefully acknow ledge the w ork o f Robert G ranger in c o ordinating the overall w ork o f the team . H ans Bos and Janet Q u in t o f M DRC also helped assure coordination o f the research w ith the larger New C hance Evaluation. In dependent research and evaluation consultant Carolyn Eldred was in stru m en tal in adapting the observational procedures o f the New Chance O bservational Study for fielding in a survey research context. This chapter reflects m any helpful suggestions from a careful review by Kristin M oore. We thank Jennifer Sargent o f C hild Trends for her w ork in helping to brin g this m anuscript to closure.

REFERENCES Alexander, K. L., & F.ntwistle, D. R. (1988). Achievement in the first 2 years o f school: Patterns and processes. Monographs o f the Society for Research in Child Development, 53 {2, Serial No. 218). Baldwin, W.,& Cain, V. (1980). The children of teenage parents. Family Planning Perspectives, 12,34-43. Baumrind, D. (1971). H armonious parents and their preschool children. Developmental Psychology, 4, 99-102. Baumrind, D. (1989). Rearing competent children. In W. Damon (Ed.), Child development today and tommorrow (pp. 349-378). San Francisco: Jossey-Bass. Booth, C. L., Mitchell, S. K., Barnard, K. E., & Spieker, S .). (1989). Development of maternal social skills in multiproblem families: Effects on the mother-child relationship. Developmental Psychology, 25, 403-412. Bradley, R. H., & Caldwell, B. M. (1981). The HOME inventory: A validation o f the preschool scale for black children. Child Development, 52,708-710.

220

Z A S L O W ET AL.

Bradley, R. H., & Caldwell, B. M. (1984). The relation of infants’ home environment to achievement test performance in first grade: A follow-up study. Child Development, 55,803-809. Bronfenbrenner, U. (1986). Ecology of the family as a context for human development: Research per­ spectives. Developmental Psychology, 22, 723-742. Brooks-Gunn, J.>8c Furstenberg, F. F., Jr. (1986). The children of adolescent mothers: Physical, academic and psychological outcomes. Developmental Review, 6 ,224-251. Bureau of the Census. (1995). Statistical brief. Mothers who receive AFDC payments: Fertility and socio­ economic characteristics (SB/95-2). Washington, DC: U.S. Government Printing Office. Bus, A. G., Van Ijzendoorn, M. H., 8c Pellegrini, A. D. (1995). Joint book reading makes for success in learning to read. A metaanalysis on intergenerational transmission of literaccy. Review of Educa­ tional Research, 65,1-21. Caldwell, B. M., 8c Bradley, R. H. (1984). Home Observation for Measurement o f the Environment. Little Rock, AR: University of Arkansas. Comstock, G. W., 8c Helsing, K. J. (1976). Symptoms o f depression in twro communities. Psychological Medicine, 6, 551-563. Crockenberg, S. (1987). Predictors and correlates of anger toward and punitive control of toddlers by adolescent mothers. Child Development 58,964-975. De Temple, J. (1994). Book reading styles o f low-income mothers with preschoolers and childrens later lit­ eracy skills. Unpublished doctoral dissertation, Harvard University, Cambridge, MA. De Temple, J. (1997, April). Mothers’ book reading with young children: An observational study within the New Chance Demonstration Project. In L. Sherrod (Chair), Mother-child relations in the context o f contrasting programs for welfare families. Symposium conducted at the biennial meeting of the Society for Research in Child Development, Washington, DC. De Temple, J., 8c Snow, C. (1998). Mother-child interactions related to the emergence of literacy. In M. J. Zaslow & C A. Eldred (Eds.), Parenting behavior in a sample o f young single mothers in povertyr. Re­ sults of the New Chance Observational Study. New York: Manpower Demonstration Research Cor­ poration. Dion, M. R. (1997, April). Implications of child characteristics for children’s developmental outcomes in a welfare sample. In K. Moore (Chair), Factors in the development o f children in welfare families: An ecological perspective. Symposium conducted at the biennial meeting of the Society for Research in Child Development, Washington, DC. Dodge, K. A., Pettit, G. S., 8c Bates, J. E. (1994). Socialization mediators o f the relation between socio­ economic status and child conduct problems. Child Development, 65,649-665. Dubow, E., 8c Luster, T. (1990). Adjustment of children born to teenage mothers: The contribution of risk and protective factors. Journal o f Marriage and the Family, 52 ,393-404. Duncan, G. J.> 8c Brooks-Gunn, J. (1997). Income effects across the life span: Integration and interpre­ tation. In G. Duncan 8c J. Brooks-Gunn (Eds.), Consequences o f growing up poor{ pp. 596-610). New York: Russell Sage Foundation. Duncan, G., Brooks-Gunn, J., 8c Klebanov, P. K. (1994). Economic deprivation and early childhood de­ velopment. Child Development, 65,296-318. Egeland, B.> Carlson, E., 8c Sroufe, L. A. (1993). Resilience as process. Development and Psychopathology, 5,517-528. Egeland, B.> 8c Zaslow, M. (1995, March). The effects of the New Chance program on m other-child in­ teraction. In A. Huston (Chair), Effects o f welfare on children and parenting processes. Symposium conducted at the biennial meeting of the Society for Research in Child Development, Indianapolis* IN. Elardo, R. D., 8c Bradley, R. H. (1981). The Home Observation for Measurement of the Environment (HOME) scale: A review of research. Developmental Review, 1,113-145. Eldred, C. A. (1998). Implementing observational research within a survey context. In M. J. Zaslow 8c C. A. Eldred (Eds.), Parenting behavior in a sample o f young single mothers in poverty: Results of the New Chance Observational Study. New York: Manpower Demonstration Research Corporation.

9.

P R O T E C T I V E FACTORS IN L O W - I N C O M E C H I L D R E N

221

Erickson, M. F., Egeland, B., & Pianta, R. (1989). The effects o f maltreatment on the development of young children. In D. Cicchetti 8c V. Carlson (Eds.), Research and theory: Child maltreatment (pp. 647-684). London: Cambridge University Press. Furstenberg, F. F., Brooks-Gunn, J., 8c Chase-Lansdale, L. (1989). Teenage pregnancy and childbearing. American Psychologist, 44,313-320. Garmezy, N. (1985). Stress-resistant children: The search for protective factors. In J. E. Stevenson (Ed.), Recent research in developmental psychopathology (pp. 213-233). Oxford: Pergamon. Garmezy, N. (1991). Resiliency and vulnerability lo adverse developmental outcomes associated with poverty. American Behavioral Scientist, 34,416-430. Garmezy, N., Masten, A. S., & Tellegen, A. (1984). The study of stress and competence in children: A building block for developmental psychopathology. Child Development, 55,9 7 -1 11. Greene, A. D., 8c Moore, K. A. (1996). Nonresident father involvement and child outcomes among young children in families on welfare. Paper presented at the Conference on Father Involvement, sponsored by the National Institute of Child Health and Human Development’s Family and Child Well-Being Network, Bethesda, MD. Harrington, G., Block, J. H., 8c Block, J. (1978). Intolerance of ambiguity in preschool children: Psychometric considerations, behavioral manifestations, and parental correlates. Developmental Psychol­ ogy, 14, 242-256. Klerman, L. V. (1991). The association between adolescent parenting and childhood poverty. In A. C. Huston (Ed.), Children in poverty. Cambridge, England: Cambridge University Press. Lindblad-Goldberg, M., 8c Dukes, L. J. (1985). Social support in Black, low-income, single-parent fam­ ilies: Normative and dysfunctional patterns. American Journal o f Orthopsychiatry, 55,42-58. Luster, T., 8c McAdoo, H. P. (1994). Factors related to the achievement and adjustment of young African-American children. Child Development, 6 5 ,1080-1094. Luthar, S. S., 8c Zigler, E. (1991). Vulnerability and competence: A review of research on resilience in childhood. American Journal of Orthopsychiatry, 61,6-22. Masten, A. S., Garmezy, N., Tellingen, A., Pellegrine, D. S., Larkin, K., 8c Larsen, A. (1988). Competence and stress in school children: The moderating effects of individual and family qualities. Journal of Child Psychology and Psychiatry, 29,745-764. Maynard, R. A. (1997). The costs of adolescent childbearing. In R. Maynard (Ed.), Kids hawng kids: Eco­ nomic costs and social consequences of teen pregnancy (pp. 285-337). Washington, DC: The Urban Institute. McLoyd, V. C. (1990). The impact o f economic hardship on black families and children: Psychological distress, parenting, and socioemotional development. Child Development, 61,311-346. McLoyd, V. C., Jayaratne.T. E., Ceballo, R., 8c Borquez, J. (1994). Unemployment and work interruption among African-American single mothers: Effects on parenting and adolescent socioemotional func­ tioning. Child Development, 65, 562-590. McLoyd, V., 8c Wilson, L. (1991). The strain of living poor: Parenting, social support, and child mental health. In A. C. Huston (Ed.), Children in poverty: Child development and public policy (pp. 105— 135). New York: Cambridge University Press. Moore, K. A. (1986). Children o f teen parents: Heterogeneity o f outcomes. Final Report to the Center for Population Research, National Institute for Child Health and Human Development, Grant No. H D -184262. Moore, K. A., Morrison, D. R., 8c Greene, A. D. (1997). Effects on the children born to adolescent m oth­ ers. In R. A. Maynard (Ed.), Kids having kids: Economic costs and social consequences o f teen preg­ nancy (pp. 145-180). Washington, DC: The Urban Institute. Moore, K. A., Myers, D., Morrison, D. R., Nord, C. W., Brown, B., 8c Edmonston, B. (1992). Age at first birth and later poverty. Paper presented at the National Institute for Child Health and Human De­ velopment Conference on outcomes of early childbearing: An appraisal of recent evidence, Bethesda, MD. Moore, K. A., Zaslow, M. /., Coiro, M. J., Miller, S. M., 8c Magenheim, E. B. (1995). How well are they far­

222

Z A S L O W ET AL.

ing? AFDC families with preschool-age children in Atlanta at the outset o f the JOBS Evaluation. Wash­ ington, DC: U.S. Department of Health and Human Services, Office of the Assistant Secretary for Planning and Evaluation. Pearlin, L. L, Menaghan, E. G„ Lieberman, M. A., 8c Mullan, J. T. (1981). The stress process. Journal of Health and Social Behavior, 22,337-356. Polit, D. F. (1996). Parenting and child outcome measures in the New Chance 42-month survey. New York: Manpower Demonstration Research Corporation. Quint, J. C., Bos, J. M„ & Polit, D. F. (1997). New Chance: Final report on a comprehensive program for young mothers in poverty and their children. New York: Manpower Demonstration Research Corpo­ ration. Quint, J. C., 8c Egeland, B. (1995). Two generation programs for families in poverty: A new intervention strategy. In S. Smith (Ed.), Two-generation programs for families in poverty {pp. 199-225), Norwood, NJ: Ablex. Quint, J. C., Fink, B. L., & Rowser, S. L. (1991). New Chance: Implementing a comprehensive program for disadvantaged young mothers and their children. New York: Manpower Demonstration Research Corporation. Quint, J. C , Musick, J. S. & Ladner, J. A. (1994). Lives o f promise, lives o f pain: Young mothers after New Chance. New York: Manpower Demonstration Research Corporation. Quint, J. C., Polit, D. F., Bos, H., & Cave, G. (1994). New Chance: Interim findings on a comprehensive pro­ gram for disadvantaged young mothers and their children. New York: Manpower Demonstration Re­ search Corporation. Radloff, L. S. (1977). The CF.S-D Scale: A self-report depression scale for research in the general popu­ lation. Applied Psychological Measurement, 1, 385-401. Rahe, D. F. (1984). Interaction patterns between children and mothers on teaching tasks at age 42-months: Antecedents in attachment history, intellectual correlates, and consequences on children's socioemotional functioning. Unpublished doctoral dissertation. University of Minnesota. Rutter, M. (1979). Protective factors in children’s responses to stress and disadvantage. In M. W. Kent & ]. Rolf (Eds.), Primary prevention ofpsychopathology, Vol. Ill: Social competence in children (pp. 4974). Hanover, NH: University Press of New England. Saluter, A. (1996). Marital status and living arrangements March 1995 (Update). Current Population Re­ ports, P20-494. Washington, DC: U.S. Bureau of the Census, U.S. Department of Commerce. Stevens, J. H. (1988). Social support, locus of control, and parenting in three low-income groups of mothers. Black teenagers, black adults, and white adults. Child Development, 59,635-642. Unger, D. G.,8c Wandersman, L. P. (1988). The relation of family and partner support to the adjustment of adolescent mothers. Child Development, 59,1056-1060. U.S. House o f Representatives, Ways and Means Committee. (1996). Green Book, 1996. Print WMCP: 104-14. Washington, DC: U.S. Government Printing Office. Ventura, S. J., Peters, K. D., M artin, }. A., & Maurer, J. D. (1997). Births and deaths: United States, 1996. Monthly Vital Statistics Report (Vol. 46, No. 1, Supp 2). Hyattsville, MD: National Center for Health Statistics. Weinfield, N. S., Egeland, B., 8c Ogawa, J. R. (1998). Affective quality of mother-child interaction. In M. Zaslow 8c C. Eldred (Eds)., Parenting behavior in a sample o f young single mothers in poverty: Re­ sults o f the New Chance Observational Study (pp. 71-113). New York: Manpower Demonstration Research Corporation. Weinfield, N. S., Ogawa, J. R., llennighausen, K. H., 8c Egeland, B. (1997). Helping mothers, helping children? Predicting child outcomes from m other-child interaction in the New Chance Demon­ stration. In L. Sherrod (Chair), Mother-child relations in the context o f contrasting programs for wel­ fare families. Symposium conducted at the biennial meeting of the Society for Research in Child Development, Washington, DC. Weiss, B., Dodge, K., Bates J., 8c Pettit, G. (1992). Some consequences of early harsh discipline: Child ag­ gression and a maladaptive social information processing style. Child Development, 6 3 ,1321—1335.

9.

PR OTE CTI VE FACTORS IN L O W -I N C O M E C H IL D R E N

223

Werner, E. (1989). High-risk children in young adulthood: A longitudinal study from birth to 32 years. American Journal of Orthopsychiatry, 59, 72-81. Werner, E. (1993). Risk, resilience and recovery: Perspectives from the Kauai Longitudinal Study, Devel­ opment and Psychopathology, 5, 503-515. Werner, E., & Smith, R. (1982). Vulnerable but invincible: A longitudinal study o f resilietit children and youth. New York: McGraw-Hill. Whitehurst, G. J., Galco, F. L., Lonigan C. J., Fischel, J. E. (1988). Accelerating language development through picture book reading. Developmental Psychology, 2 4 ,552-559. Zaslow, M. J., & Eldred, C. A. (Eds.). (1998). Parenting behavior in a sample of young single mothers in poverty: Results o f the New Chance Observational Study. New York: Manpower Demonstration Re­ search Corporation. Zill, N. (1985). Behavior problem scales developed from the 1981 Child Health Supplement to the National Health Interview Survey. Washington, DC: Child Trends. Zill, N., Moore, K. A., Smith, E. W, Sticf, T., & Coiro, M. J. (1995). The life circumstances and develop­ ment of children in welfare families: A profile based on national survey data. In L. P. Chase-Lansdale & J. Brooks-Gunn (Eds.), Escape from poverty: What makes a difference for children? (pp. 38-59). New York: Cambridge University Press.

This page intentionally left blank

PART IV

Family Functioning and Child Adjustment in Repartnered Relationships and in Stepfamilies

This page intentionally left blank

JO

Contexts as Predictors of Changing Maternal Parenting Practices in Diverse Family Structures A Social Interactional Perspective o f Risk and Resilience

D av id S. D e G a rm o M a rio n S. F o rg a tch Oregon Social Learning Center

S t u d i e s c o m p a r i n g the social and academ ic adjustm ent o f youngsters living in diverse family structures provide convincing evidence that divorce and repart­ nering are transitions that place youngsters at risk. Such family structure transi­ tions are associated w ith increased risk for conduct problem s, school problems, and peer rejection (Brody, N eubaum , & Forehand, 1988; Capaldi 8c Patterson, 1991; Chase-Landsdale, Cherlin, & Keirnan, 1995; Forgatch, Patterson, 8c Ray, 1996; Furstenberg 8c Seltzer, 1986; H etherington 8c Clingempeel, 1992; Zill, M orri­ son, & Coiro, 1993). W hy do m ost youngsters w eather these family transitions relatively unscathed while a substantial subset suffers? O ne answer appears to lie w ithin the parent-child context. Effective parenting practices contribute signifi­ cantly to differential adjustm ent for the children who live in divorced and repart­ nered families (Anderson, Lindner, 8c Bennion, 1992; Bray 8c Berger, 1993; Capaldi & Patterson, 1991; Forgatch 8c DeG arm o, 1997; H etherington, 1993; H etherington 8c Clingempeel, 1992; Vuchinich, Vuchinich, 8c Wood, 1993). Some m others are able to m aintain effective parenting practices even while gaining or losing an inti­ m ate partner. This ability to rem ain constant in parenting may reflect a balance in protective and risk factors w ithin the m o th er’s social environm ent. From a social interactional perspective, the contexts surrounding interpersonal interactions affect the m icrosocial behaviors that are exchanged between individu-

227

FIG 10.1.

228

Maintaining balance on the parenting path.

10.

C O N T E X T S AS P R E D I C T O R S O F P A R E N T I N G

229

als (Patterson, in press; Patterson & Forgatch, 1990). As these exchanges become patterned and habitual, they can generalize to other social settings. Patterned be­ haviors are assumed to form w ithin influential social settings. For m ost children, the family is the prim ary socializing setting (Patterson, 1982). Parents teach their children how to interact with others through their use o f parenting practices, such as discipline and interpersonal problem solving. For example, coercive discipline practices m ay teach youngsters to em ploy coercive tactics w ith peers and teachers (Patterson, 1986; Snyder, Peterson, 8c St. Peter, 1997). Effective problem solving, on the o ther hand, may teach children how to resolve conflict and get along with o th ­ ers (Forgatch, 1989; Forgatch et al., 1996; Forgatch 8c D eG arm o, 1997). Family structure transitions generate profound changes in social context for parents and children. If these potentially disruptive changes take place with few protective factors, the quality o f parenting practices m ay decline. T hroughout this chapter, we exam ine how changes in the m other’s social environm ent are related to the quality of her discipline strategies and interpersonal problem solving practices. W ithin a social interactional framework, we sum m arize recent findings that sug­ gest how certain contexts m ay protect or interfere w ith parenting practices. Figure 10.1 illustrates the dynamics o f weathering storm y transitions for divorced and repartnered mothers. Transitions generate a m ultitude o f risk factors, m aking it difficult for a parent to m aintain balance on the path of parenting during transi­ tions. Protective factors can provide the balance necessary to stay on course.

A SOCIAL INTERACTIONAL VIEW OF RISK Family structure transitions are em bedded within the environm ent in which the family lives (Capaldi 8c Patterson, 1991). Several studies show that divorce and repartnering are transitions associated with background factors that interfere with effective parenting (Bank, Forgatch, Patterson, 8c Fetrow, 1993; Conger, Patterson, 8c Ge, 1995; Forgatch et al., 1996; H etherington, Stanley-Hagan, 8c Anderson, 1989; Patterson 8c Forgatch, 1990; Simons, Beaman, Conger, 8c Chao, 1993). These back­ ground factors include stress, m aternal depression, socioeconom ic disadvantage, and antisocial parents. The social interactional m odel argues that the association between such contextual risk factors and negative child outcom es is mediated by parenting practices in all family structure types (Forgatch et al., 1996; Patterson, 1986). Some parents m aintain effective child-rearing practices even while transi­ tioning in and ou t of intim ate relationships. W hat contexts contribute to this sort of parenting stability? In the first part of this chapter, we conceptualize family structure transitions as a dynam ic contextual factor th at has a profound im pact on the family social envi­ ronm ent. In an at risk population-based sample, we evaluate parenting practices across four family structure types (i.e., nuclear, first-tim e single m other, first-tim e stepfainily, m ultiple transitions). We then show how the process o f coupling and

230

D

eG

ARMO

a n d

f o r g a t c h

uncoupling for these families affects their parenting across tim e. In the second part of the chapter we examine social support as a protective factor enhancing the qual­ ity o f parenting during divorce and repartnering.

The LIFT Study Data for this p art of the chapter are from the LIFT study (Linking Interests o f Fam ­ ilies and Teachers). LIFT was p art o f a population-based random ized intervention trial sponsored by the N ational Institute o f M ental Health center for the preven­ tion o f conduct disorders (Reid, 1993). Details on the design, sample characteris­ tics, and intervention outcom es are described elsewhere (Reid, Eddy, & Fetrow, in press). T he preventative intervention was designed for all first and fifth grade ele­ m entary school boys and girls and their families living in at-risk neighborhoods characterized by high rates o f delinquency. T he LIFT sam ple was drawn from 12 eligible schools. O f the 762 families con­ tacted in these schools, 12% declined to participate, resulting in 671 first and fifth graders (51% were girls). P articipants tended to be W hite (89%), to be in the m id­ dle to lower socioeconom ic classes, and to have com pleted high school or som e college education. Approximately 25% received som e type o f financial assistance. O f those 671 participating families, our sample for this report was drawn from 621 families, excluding single fathers and stepm other families because our focus was on the relation between transitions and m aternal parenting. T he data for the c u r­ rent analyses were collected across 2 years with four assessments (i.e., BaselineTime 1, a 6-m onth follow -up— Time 2, a 1-year follow -up— Tim e 3, and a 2-year follow -up— Tim e 4). The final longitudinal sam ple consisted o f 444 families with complete data on contextual predictors and observed parenting variables. The family structures were distributed evenly across grade and gender with no differ­ ential rates o f m issing longitudinal d a ta .x 2 (3) = 2.98, p = .26.

Transitions We defined family structure types as a com bination o f family configuration and the cumulative num ber o f transitions at baseline. Nuclear families consisted o f two biological parents living together with the focal child. No prior separations o r d i­ vorces had occurred for either parent. Single-mother families included a focal child living with a biological m other w ho had no cohabiting intim ate partner (i.e., never-m arried single m others and once separated or divorced m others). First step­ families included a single m other who cohabited with a spouse or intim ate partn er and the m other’s biological child. Multiple-transition families included a focal child with a biological m other w ho had experienced several separations o r re­ unions with the sam e or different partners. In our analysis there were 344 nuclear families (55% ), 84 single-m other families (14% ), 91 first-tim e stepfamilies (15%), and 102 m ultiple-transition families (16%).

10.

C O N T E X T S AS P R E D IC T O R S OF P A R E N T IN G

231

Parenting Practices Two aspects o f parenting practices were m easured using direct observations of m o th e r-ch ild interactions in the laboratory, conflict b o u ts initiated by the m o th er and family problem -solving outcom e. For the first graders, there were four m o th e rchild in teraction tasks: a 5-m inute co m m unication task, a 5-m inute social teach­ ing task, a 5-m inute problem -solving task, and a 10-m inute clean-up task, for a total o f 25 m inutes. For the fifth graders, there were three interaction tasks: a 5-m inute fam ily w arm -u p task (m o th e r-ch ild , o r m o th e r-fa th e r-c h ild for tw oparent fam ilies), a 10-m inute problem -solving task w ith m o th e r and child only, and a 10-m inute family problem -solving task, for a total o f 25 m inutes.

Conflict Bouts. This m icrosocial m easure reflected a form o f coercive disci­ pline in which a m o th er initiated a conflict bo u t. A b o u t was an exchange o f aversive behaviors in w hich the child had to respond w ith an aversive behavior w ithin 12 seconds. T he b o u t ended when there was a period o f 12 o r m ore seconds w ith­ o u t aversive behavior. T he frequency o f b outs was scored w ith the negative content and affect codes from the Interpersonal Process C ode (IPC; Rusby, Estes, & Dishion, 1991). Ten percent o f the interactions were random ly selected and coded by indep en d en t observers for interobserver reliability. Average percent agreem ent was 85% for co ntent and 86% for affect. Kappa was .70 for co n te n t and .60 for affect. Problem -Solving Outcome. M others and children chose topics from the Is­ sues Checklist (adapted from Prinz, Foster, Kent, & O ’Leary, 1979) listing com ­ m on conflicts. Dyads were asked to atte m p t to resolve issues rated as “ hottest.” A six-item scale (e.g., quality o f solutions, likelihood o f follow through, and family seem ed satisfied w ith resolution) was based on IPC coder ratings after scoring discussion o f each issue d u rin g the problem solving tasks (alphas ranged from .85 to .89). C ontextual P redictor Variables The contextual variables were m o th e r’s age, m aternal depression using the C en­ ter for E pidem iological Study Depression scale (CES-D; Radloff, 1977), n u m b e r o f parental arrests based o n c o u rt records, and socioeconom ic status (SES) using the H ollingshead Index o f Social Status (H ollingshead, 1975).

The Relation Between Family Transitions and Parenting Practices We first exam ined the m ean scores for parenting practices across tim e by family stru ctu re using repeated-m easures ANOVA. M ean trends in the data for m otherinitiated conflict bouts are illustrated in Fig. 10.2 in the form o f bar graphs. First, there was a significant decrease for the sam ple as a whole, F (3, 444) = 4.33, p =

232

D eG A R M O

an d

forgatch

.005, illustrated by the general decrease in bars for each fam ily stru ctu re going from left to right. Second, nuclear families had fewer m oth er-in itiated conflict b outs over the 2-year period com pared to o th er fam ily types on the average, F (3, 446) = 8.78, p < .001. All o f the bars for nuclear families fell below 1.5 bouts, whereas the o th e r families were higher on the average. The ANOVA revealed no differences in the rate o f decrease co m paring each family stru c tu re type. T he m ean tren d s for family problem -solving o utcom e are illustrated in Fig. 10.3. N uclear fam ilies again fared better on the average over tim e com pared to no n -n u clear fam ilies, F (3, 458) = 4.31, p = .005. T he levels for nuclear families were rath er stable, w ith a m ean score o f approxim ately 2.9. In general, single­ m o th e r fam ilies an d the m ultiple-transition families show ed a linear increase in problem solving; however, the ANOVA revealed no m ean tren d for the sam ple o r significant differences over tim e by family stru ctu re type. In sum m ary, b o th Figs. 10.2 a n d 10.3 show ed th at nuclear families exhibited better p a ren tin g over tim e th an o th er fam ily types. T he graphs illustrate m ean trends; however, they do n o t show the influence o f o th er contextual factors related to parenting, nor, m ore im p o rta n t, the im pact o f any change in fam ily stru c tu re th a t occurred d u rin g the study. To show these p a r­ ticular effects we em ployed grow th m odels that included predictors o f the rates o f change using hierarchical linear m odeling (HLM ; Bryk & R audenbush, 1992; R audenbush, 1995). HLM grow th m odels com bine individual and gro u p levels o f analysis, th u s allow ing for the assessm ent o f both in tra-individual change and inter-individual differences in change. In general, least-squares regression curves are fit to the repeated m easures o f each family. T he individual grow th curves then becom e the o utcom e focus o f analysis. Estim ates in the m odel for each individual are derived from b o th the inform ation on each fam ily’s grow th curve and in fo r­ m ation from the covariates. T he final estim ates are weighted com binations o f each fam ily’s grow th curve param eters and the predictors o f change. The first step in the analysis tests an unconditional m odel o r the estim ated m ean levels before entering p red icto r variables. C ontrolling for error, the esti­ m ated average n u m b er o f conflict b outs was 1.61 for the sam ple, a n d the average score for problem solving was 2.78. T he sam ple significantly declined in conflict b o u ts w ith a m ean slope o f -.1 3 2 (p< .05) and significantly im proved in problem solving w ith a m ean slope o f .034 (p < .05). The estim ated HLM m ean levels p ro ­ duced a slightly different result than the repeated-m easures ANOVA that showed no m ean trend for problem solving. We then m odeled covariates o f grow th by entering the contextual predictors o f baseline levels a n d change in parenting practices. Results o f the p aren tin g grow th m odels are presented in Table 10.1 in the form o f unstandardized beta coefficients. At baseline, Table 10.1 shows th at family structure, m aternal depression, a n d p a ­ rental arrests w ere all significant risk factors associated with m ore conflict bouts. H igher SES and m aternal age were m arginally associated w ith fewer conflict bouts. For problem solving, SES was significantly associated w ith hig h er scores at baseline

HtJ 30.2. Mem Iih k Js for mi bated conflict tarts from Tir»* I r-n Tirr? 4 tyr fa-mily slTvciijrf Sypc in LIFT fn — -m l

]l I I ■ ' '

. -

. u

m

. l i

tJ

nuclcuf

s in g le

fittistep

multiple

mnUicT M«*fl let-rtds ifeii family f.ji.lJi-jii solving cmicoom-fromT u n e 1 to Tim e A b f (unity sm K tune type in U t T { h

FIG IQ-3,

23J

234

D eG A R M O

a nd

forgatch

TABLE 10.1 Contexts as Predictors of Growth in Parenting Practices Controlling for Grade and Gender of Child Mother-Initiated Conflicts

Problem Solving Outcome

Baseline status as DV Intercept Family structure M other’s age SES Maternal depression Parent arrests

2.045*** .131* -.023' -.010' .016* .066*

2 .3 6 0 "’ -.034 .009' .006* -.007* -.012

Rate o f change as DV Intercept Family structure Mother s age SES Maternal depression Parent arrests

-.177 -.010 .009 .000 -.014* .022

.095 .023 -.003 .001 -.000 -.007

Note. DV = dependent variable; n - 444; ***p< .001; **p< .0!;*/>< .05;'p < .10.

a nd m aternal depression w ith low er scores at baseline. O lder m others also scored m arginally better on problem -solving outcom e. For change in paren tin g over the 2-year interval, higher levels o f m aternal de­ pression a t baseline were associated with decreases in conflict bouts, whereas none o f the predictors was associated w ith grow th in problem -solving outcom e. In su m ­ m ary, the contextual variables explained m ore o f the initial status differences in p arenting for the LIFT sam ple than change over tim e. We next w anted to address the question o f fam ily stru c tu re transitions th at oc­ c u rred d u rin g the 2-year study period. We m odeled additional transitions as a tim e-varying covariate, m eaning this variable could change w ith tim e. In o th er words, the m odel shows the effect o f change in fam ily stru c tu re as it covaries w ith change in parenting practices. T he tim e varying m odel produced a significant effect for additional transitions on conflict bouts ((3 = .090, p < .01) and a signifi­ cant effect for problem solving outcom e ((3 = -.0 3 4 , p < .05). These coefficients m ean th at although paren tin g im proved over tim e for the sam ple, family tra n ­ sitions th at occurred had a negative im pact on the quality o f discipline strate­ gies and problem solving. The risk o f having transitions is illustrated in Figs. 10.4 and 10.5. Taking in to account the effect o f im provem ent over tim e an d the detrim ental effect o f transitions d u rin g the study, the final slopes indicated th at transitions had a negative im pact on p a ren tin g by slowing dow n the rate o f im provem ent. For ex­ am ple, Fig. 10.4 shows the rate o f decrease for m oth er-in itiated conflict b o u ts was steeper than for families th at transitioned. Similarly, Fig. 10.5 show s that the rate o f

I * |2

5h 3

r* r c 2

I is L i

i=r r" 3 * * -

I| s i

T- ■"

estim ated gw roth in family problem solving outcom e

• lK»

p » v i |f U i - . m i |i o u i

JOf

HAITI?

qw urf

ijhUti | >«lfl &*q:iUFj| p U t1 .*}dlL*7? bXJI

D U

1Z*T1

Jlpnis tfuunp HJI JQf S:f>«]

p ^ iru iijv H

T -flf

estim ated grow th in m olher-initiatcd conflict bouts

236

D

e

GARMO

AND

FORGATCH

change for families that transitioned was flat com pared to the im provem ent for the w hole sample.

The Relation Between Contexts and Transitions In a final m odel, we w anted to investigate contextual factors th at were related to the risk family stru c tu re transition. T he results are displayed in Table 10.2. Significant risk factors associated w ith the n u m b er o f transitions at the study’s baseline were m aternal depression, parental arrests, younger m aternal age, and lower SES. For change over the 2-year period, younger m others as well as those w ho were d e ­ pressed were at risk for additional transitions. We also exam ined tran sitio n s d u r­ ing the study by family stru c tu re at baseline. N uclear fam ilies experienced fewer transitions d u rin g the study com pared to non -n u clear families, F (3 ,6 1 7 ) = 8.06, p < .001. In total, 11% o f the nuclear families experienced a tran sitio n , 33% o f the single-m other families, 20% o f the stepfam ilies, and 37% o f the m u ltip le-tran si­ tio n families. These data indicated that a p rio r history o f tran sitio n was a risk fac­ to r for fu rth er family d isruption. Indeed, entering fam ily stru c tu re as a predictor o f grow th (n o t show n) was the greatest risk factor. In sum m ary, data from the LIFT sam ple show ed th a t fam ily stru c tu re ac­ c ounted for differences in effective parenting practices. N uclear families and m ore stable families over tim e had b etter p arenting com pared to families w ith tran si­ tions. Baseline fam ily stru c tu re type did not p redict grow th in conflict o r problem solving. O ne m ight conclude there was little risk for parenting over tim e. However, this interpretation w ould only be p art o f the picture. We fo u n d an inform ative ap-

TABLE 10.2 Contexts as Picdictors of Family Transitions Controlling for Grade and Gender Family Transitions During Study Baseline status as DV Intercept Mother s age SES Maternal depression Parent arrests

2.811*** -.056” * -.014* .030*** .092**

Rate o f change as DV Intercept Mother’s age SES Maternal depression Parent arrests

.215* -.005* .001 .004* .010

Note. DV - dependent variable; n = 444; ***p < .001; **p < .01; *p< .05; lp < . 10.

10.

CONTEXTS

AS P R E D I C T O R S O F

PA REN TIN G

237

p roach was to conceptualize and m odel transitions as a process. The longitudinal m odel o f transitions indicated th at any d isru p tio n in the organization o f family life had a detrim ental im pact on p arenting practices. O u r analyses reflect an attem p t to move beyond a static view o f fam ily structure based on com parisons o f family typologies. The field o f family research has p ro ­ vided convincing evidence o f the linear an d cum ulative risks for parenting th a t are associated w ith family structure. However, family stru c tu re typologies m ay m ask the dynam ic n ature and heterogeneity w ithin transitional families th a t m ay ac­ c o u n t for variation o f individual differences. For exam ple, detailed family stru c ­ ture histories for the LIFT families indicated that for fam ilies w ho had experienced any transition (45% ), there were 89 different pathw ays to th eir baseline family stru ctu re (e.g., going from nuclear family, to single-m other family, to stepfam ily, a nd then back to single-m other family before baseline). C onceptualizing tran si­ tions as a process m ay be a m o re inform ative way to predict m icrosocial aspects re­ lated to child outcom es. For exam ple, we considered single m o th ers w ith m ultiple transitions as different than once-divorced m others. By using m ethods and typol­ ogies th a t capture process we can better un d erstan d a d ju stm en t outcom es. This first section o f the c hapter has exam ined how tran sitio n s an d contexts were risk factors for parenting practices. We tu rn now to the study o f social contexts that en ­ hance paren tin g quality.

A SOCIAL I NTERACTI ONAL V I EW OF PARENTING RESI LI ENCE Social su p p o rt has long been studied as a protective factor that contributes to a p e rso n s w ell-being in the face o f stress (Barrera, 1986; C ohen 8c Wills, 1985; Coyne & Downey, 1991; Rook, 1987) and as a source o f resiliency for adolescents (Wills, Blechm an, 8c M cN am ara, 1996; Wills, Vaccaro, 8c M cN am ara, 1992). T he concept o f resiliency is often discussed as strengths w ithin individuals th at help them over­ com e adversity. T he social interactional perspective looks for these strengths w ithin the social environm ent, for exam ple, contingent and supportive behavioral exchanges betw een adults o r betw een parents and children. We have been p a rticu ­ larly interested in how adult su p p o rt functions to help m others m aintain quality p a ren tin g practices as they m ake family stru ctu re transitions. Divorce research show s th a t su p p o rt a n d com panionship from close relatio n ­ ships buffer stress for single w om en and single m others (Aseltine 8c Kessler, 1993; H ughes, G ood, 8c C andell, 1993; Kitson, 1992; Sim ons et al., 1993). C hanging parental dem ands are a m ajor source o f stress for divorced m others, w ho are likely to be custodial caretakers. Effective social su p p o rt can benefit the q uality o f p a r­ en tin g for these m others (Belsky 8c Vondra, 1989; Holloway 8c M achida, 1991; Sim ons 8c Johnson, 1996). However, the relation betw een social su p p o rt and p a r­ en tin g is not well understo o d for a n u m b er o f reasons.

238

DEGARMO AND

FO RGATCH

A social interactional view o f the role o f support in parenting m ust take into ac­ count several factors. These include the personal characteristics o f the interactants, how these characteristics interrelate and contribute to support as an outcom e, how support functions as a predictor o f relevant outcom es (in this case parenting qual­ ity), and how larger contexts (e.g., SES) influence support processes. Personal characteristics can lead to the loss o f support or to the availability o f support givers who have little of use to offer. A stress m aintenance fram ework suggests how an individual can becom e entrapped in a stress-generating process (Patterson & For­ gatch, 1990). For example, in an analysis o f m others studied shortly after separa­ tion, m aternal stress, irritability, and depressed m ood were associated with irritable interactions observed between the m others and their confidants. These negative exchanges resulted in poor problem -solving outcom es for the discussions, and poor problem -solving outcom es contributed to future stress and reductions in fu­ ture support contacts (Patterson 8c Forgatch, 1990). A sim ilar perspective, the sup­ port erosion m odel (Kaniasty & Norris, 1993), states that a person’s heightened distress can be aversive and prom ote increased stress for support providers, result­ ing in loss in the a m ount or the quality of support given. Longitudinal studies have shown that em otional distress and negativity erodes a partner’s support for pa­ tients with physical health im pairm ents (Bolger, Foster, Vinokur, 8c Ng, 1996; Lane 8c Hobfoll, 1992). A fascinating tu rn in support research was taken with the position that not all support is helpful. Costs can be incurred when the support is given in the context of po o r quality relationships (Burleson, 1990; La Gaipa, 1990). Support providers who are themselves distressed or socially unskilled can increase stress levels for the recipient (Belle, 1982; Patterson 8c Forgatch, 1990; Sarason, Sarason, Hacker, 8c Basham, 1985). Bad tim ing in provision o f support and support that is incongruent with the needs o f the situation are not useful (C utrona, 1990). Too little em phasis on the study o f characteristics o f support providers has lim ited our u n ­ derstanding o f factors that prom ote or that erode quality support (DeG arm o 8c Forgatch, 1997b; Duffy, 1993; Hobfoll, 1990). There has been som e disagreem ent about w hether the relation between support and parenting is direct or indirect. In the m ost comprehensive review to date on social support and quality o f parenting, Simons and Johnson (1996) specified both direct and indirect effects o f social support on parenting, depending on who pro­ vides the support. Direct effects o f support on parenting tend to be hypothesized for spousal support, whereas social support from nonspousal relationships is as­ sum ed to be indirect (Belsky, 1984). Cox, Owen, Lewis, and H enderson (1989), for example, using observational measures found a direct relation between spousal support and quality parenting for nuclear families. Simons et al. (1993) found indirect effects o f nonspousal support on parenting through psychological well­ being using self-reports o f single m others. The m arital partner in nuclear families is seen as the prim ary support person. Because couples share responsibility for socializing children, partners are each

10.

C O N T E X T S AS P R E D IC T O R S OF P A R E N T IN G

239

other’s prim ary source o f parental support. Belsky and Vondra (1989) suggested that when m arital support is absent, the social netw orks becom e m ore im portant for single m others and are the principal source of parental support. Sim ons and Johnson (1996), however, argued that nonspousal support is likely not direct be­ cause it is unlikely to substitute for o r function as p arenting support from spouses, and therefore its im pact on parenting is indirect through the m other’s well-being. O ur position is in keeping with that o f Belsky and Vondra (1989) that divorced m others m aintain quality parenting by drawing on the support o f their confidants. Reliance on self-report has also limited the study o f support process (Coyne 8c Downey, 1991; C utrona, 1989; Schwarzer 8c Leppin, 1991). W ithin the field o f di­ vorce, this is particularly relevant, because divorced wom en tend to be stressed or depressed. The w ell-known cognitive distortions associated with depression make such individuals poor reporters o f the quality o f support they receive as well as the quality of their own parenting practices. We have addressed these m easurem ent issues in our studies. Most o f o u r models are based on observations of what tra n ­ spired between m others and their support confidants. Characteristics o f both par­ ticipants in the support relationship (i.e., m other and confidant) were assessed using m ultiple m ethods and in some cases m ultiple agents. We defined social sup­ p o rt as b oth quality and am ount o f supportive behaviors provided by a confidant during attem pts to solve the m other’s ow n parenting and personal problem s. The social interactional view specifies a direct effect o f social support on par­ enting when considering the characteristics o f the su p p o rt provider and the m eas­ urem ent of support and parenting. We argue that a proximal m easure o f social sup port in relation to parenting will have a direct effect. For example, support p ro ­ vided in the context o f discussing and trying to resolve parenting and personal problem s is m ore proximal to parenting practices than are global assessments of sup port networks or perceived availability. Self-reported measures o f available su p p o rt m ay be m ore likely to be associated with self-reported m easures o f well­ being. In fact, on this issue, Dunkel-Schetter and Bennett (1990) argued (1) th a t the co n tex t o f su p p o rt receipt is extrem ely im p o rta n t to consider, a nd (2) that the level o f specificity o f m easurem ent for stressors, receipt o f su p p o rt, and w ell-being sh o u ld be c o m p a ra b le .. . . In ad d itio n , it is likely th at the effects o f re­ ceived su p p o rt will be m o re ap p aren t in a sam ple th a t shares a c o m m o n negative life event th an in c o m m u n ity a n d stu d en t sam ples ex periencing diverse stresses. Before these problem s can be properly addressed, the conceptualization and m easurem ent o f received su p p o rt m u st be fu rth er developed, (p. 288)

Raising children in divorced and reconstituted families is an extraordinarily complex problem . We are convinced that m uch o f the variability in parenting quality in such families is explained by m others’ abilities to find helpm ates and to engage them in the struggle. For this reason, we have conducted a series o f studies to understand how this process operates. In this effort we have focused on three m ain questions: How do personal characteristics o f divorced m others and their confidants relate to the support process? How does confidant support relate to m a­

240

D eGARMO

and

forgatch

ternal parenting practices? and How do contextual factors, confidant support, and problem solving relate to parenting and child outcomes? To address these ques­ tions, data from the Oregon Divorce Study were employed.

The Oregon Divorce Study The Oregon Divorce Study (ODS) was conducted in two phases with independent samples o f recently separated single m others and their sons. The first phase (ODSI) began in 1984 and was a passive longitudinal study designed to develop an d test m odels relevant to children’s adjustm ent following m arital separation. Phase two (OD S-II) began in 1992 and was a random ized experimental longitudinal study. An intervention based on social learning and social interactional perspectives was carried out. The program focused on im proving child outcom es and preventing children’s adjustm ent problem s by strengthening m aternal parenting practices. Participants in both samples were recruited w ithin the same m edium -sized Pacific Northwest m etropolitan area. Assessment procedures were sim ilar and used extensive m ultiple-agent and m ultiple-m ethod m easurem ent. In both studies, each m other selected a confidant w ho was either a friend, family m em ber, or intim ate partn er to participate in the study. Data in this chapter were obtained from laboratory interviews and ques­ tionnaires with m others, telephone interviews with the m others and their con­ fidants, observations o f problem -solving interactions between m others and confi­ dants in the laboratory, observations o f several structured laboratory tasks be­ tween the m others an d focal children, and questionnaires filled out by teachers.

ODS-I Participants Participants were 196 recently separated single m others and their sons. Families were recruited through divorce application records, m edia advertisem ents, and flyers distributed throughout the com m unity. In eligible families, m others had been separated from their partners w ithin the prior 3 to 15 m onths, resided with their sons, who were in Grades K -7, and did not cohabit with new partners. The racial/ethnic com position o f the boys was representative o f the local com m unity: 89% W hite, 1% African American, 3% Latino, 3.6% Native Am erican, and 3.6% other. D em ographics are provided in Table 10.3. Generally, families were living in poverty, having experienced a dram atic drop in m edian annua! incom e from $ 15,000 before separation to $7,200 afterward. The m ean annual incom e for ODS-I m others was som ewhat lower than that reported for fem ale-headed households with children in the county w here the study was conducted, according to 1980 census data ($6,600 and $9,400, respectively). Al­ though m ost o f the wom en were employed and quite well educated, 55% o f the families were receiving public assistance. Confidant data in ODS-I were gathered in Year 4 (referred to here as T3). At T3, 40% o f the m others were cohabiting with a new partner an d o ne h alf o f the part-

10.

CONTEXTS

AS P R E D IC T O R S

241

OF PA R E N T IN G

TABLE 10.3 Demographic Characteristics of the Oregon Divorce Study ODS-I Began 1984 fn = 196) Education Some high school High school graduate High school plus College graduate and beyond Occupation Unskilled Semiskilled Clerical/skilled Medium business/minor professional Major business/major professional Employed outside the home Receiving public assistance Median annual income Mean mother age Mean boy age Average months separated Average number o f children

ODS-II Began 1992 fn = 238)

8% 20% 55% 17%

4% 20% 58% 18%

3% 20% 32% 39% 6% 60% 55% $7,200 33.1 8.7 6.4 2.2

30% 22% 23% 23% 2% 67% 76% $12,500 34.8 7.8 9.2 2.1

nered m o th ers were rem arried. Forty-seven percent o f the confidants were friends, 24% were family m em bers, and 29% were partners. Sixty-six p ercent were female. D etailed dem ographic characteristics o f the su p p o rt confidants in T3 o f O D S-I are provided by D eG arm o and Forgatch (1997b).

ODS-1I Participants Participants were 238 recently separated single m others and their sons recruited using the sam e m ethods as OD S-I. To be eligible in phase two, m o th ers had to be separated from th eir partn ers w ithin the p rio r 3 to 24 m onths, reside w ith their sons, w ho were in G rades 1-3, and n o t cohabit w ith new p artn ers. T he racial/ ethnic com position o f the boys was representative o f the local com m unity: 86% W hite, 1% African A m erican, 2% Latino, 2% Native A m erican, an d 9% o th er m i­ nority groups. Sim ilar to the O D S-I sam ple, families experienced a sharp decline in m edian annual incom e after separation, from $25,000 to $12,500, an d 76% o f the families w ere receiving public assistance. T he m ean annual incom e for OD S-II m others was sim ilar to th at reported for fem ale-headed households w ith children in the county, according to 1990 census ($14,900 and $15,300, respectively). O D S-II m o th ers were sim ilar to th eir O D S-I c o u n terp arts in their education b u t n o t in their occupation. W om en in O D S-II tended to w ork m ore in lower status jobs com pared to w om en in OD S-I (respectively, 29% and 3% for unskilled jobs, and 45% and 25% for professional-level jobs).

242

DEGARMO

a n d

f o r g a t c h

Measures M ultiple m ethods were used to build constructs from b o th sam ples o f ODS for m aternal distress, confidant negativity, social su p p o rt (observed a n d perceived), adult problem -solving outcom e, p arenting (aversive discipline and problem solv­ ing), and antisocial behavior o f the focal boy. Unless specifically noted, the m eas­ u rem en t for bo th sam ples was the same.

M aternal Distress. T his construct was m easured w ith three self-report in d i­ cators scaled to the sam e m etric. Family stress was assessed w ith the Family Events Checklist (Patterson, 1982). Life events stress sum m ed negative events rated on the Life Experience Survey (Sarason, Johnson, & Siegel, 1978). Depressed mood was m easured by the CES-D (Radloff, 1977). For chronic distress, the indicators were averaged to com pute a distress score at baseline, 1 year later, and 4 years after baseline. C onfidant N egativity. T his construct evaluated negative em otional qualities and personality traits o f the confidant w ith three indicators. Intervieu'er report o f irritability rated how angry o r irritable the confidant appeared d u rin g telephone interview s m ade once a week over 6 weeks. Irritability self-report was a selfreported telephone interview item describing the confidant’s m o o d in the last day. Each telephone indicator was aggregated from the m ean o f the six calls. Depressed m ood was m easured by the CES-D, filled out on the day o f the m o th e r-co n fid a n t problem -solving task p rio r to the interaction. O bserved C onfidant Support. This construct was m easured by three indica­ tors obtained from observations o f the m o th e r-co n fid a n t problem -solving task at T3 o f the OD S-I study using a set o f microsocial coders and a set o f global raters. M others selected two topics to discuss from an issues checklist o f com m on c o n ­ cerns for single m others regarding p arenting issues and personal problem s. M o th ­ ers selected one topic from each area and were videotaped discussing them w ith the confidants for 10 m inutes. Interpersonal support evaluated the a m o u n t and quality o f su p p o rt as rated w ith the Problem Solving System (PSS; Forgatch 8c Lathrop, 1988). A m ean score was com puted from five item s follow ing each topic (e.g., how supportive the confidant was to the m other, how often the confidant offered em otional su p p o rt). Confidant likeability was one item rated by PSS coders, “H ow w ould you like this person for your su p p o rt person?” Em otional support was the rate per m inute o f supportive behaviors scored w ith th e m icrosocial coding system , Specific Affects (SPAFF; G o ltm an , 1989). T he in dicator score was com ­ p u ted by dividing the frequency o f su p p o rt behaviors by the total tim e. Perceived Support. This co n stru ct had three indicators m easuring the m o th ­ ers’ evaluations o f the confidants and their help w ith problem s. Relationship sup­

10.

C O N T E X T S AS P R E D IC T O R S

OF PA REN TIN G

243

port was assessed w ith six item s obtained from the m others in face-to-face in te r­ views ab o u t the confidants (e.g., understands and sym pathizes w hen you are upset, w ould go o u t o f their way to do you a favor). Supportive affect was m easured by two item s from the repeated telephone calls (i.e., how h a ppy-enthusiastic an d how calm is the confidant). Help w ith problems was m easured by an item asked d u ring the repeated telephone interview s (i.e., how helpful were discussions w ith the c o n ­ fidant for solving problem s).

A du lt Problem -Solving Outcome. This c onstruct was a score from the two m o th e r-co n fid a n t problem -solving interactions rated by coders using the PSS ra t­ ing system. The scale used the sam e six item s as m o th e r-ch ild problem -solving outcom e. Parenting Practices. T his co n stru ct was m easured w ith indicators o f prob­ lem solving outcome and aversive discipline using sim ilar m easures described for the LIFT analysis. For OD S-I, problem solving used the sam e scale score as LIFT and included one m o th e r issue an d one child issue. The discipline m easure was the m ean o f the conflict b o u t score described previously for LIFT plus a global rating o f how often the m o th e r w ithdrew from interaction in a negative or avoidant way. For O D S-II, problem -solving outcom e was scored w ith three m other-picked is­ sues, and discipline used conflict bouts and a scale score for global ratings o f disci­ pline (e.g., nagging, inconsistent, inappropriate). Boy Antisocial Behavior. This c onstruct reflected a variety o f antisocial b e ­ haviors rated by parents, teachers, and the boys. Teacher and p arent reports were collected using the C hild Behavior Checklist (A chenbach & Edelbrock, 1983). Sep­ arate child and paren t reports were collected from six telephone interview s regard­ ing the boys’ m isbehavior at hom e. Personal Characteristics as Predictors of Support T he first m odel addressed the first question o f how m aternal and confidant ch ar­ acteristics affect the su p p o rt th at transpires d u rin g discussions a b o u t a m o th e r’s p aren tin g and personal problem s. Results o f the social interactional m odel includ­ ing chronic m aternal distress (i.e., average levels over 4 years) and confidant nega tivity a t T3 are displayed in Fig. 10.6 w ith standardized paths. This m odel, w hich is adapted from D eG arm o and Forgatch (1997b), controlled for relationship in ti­ macy. New p a rtn e r was a variable coded 1 for those m o th ers w ho were repartnered a t T3 and 0 for those w ho were single at T3. C hronic m aternal distress was associated w ith higher levels o f confidant nega­ tivity (P = .28). T his p ath supports the no tio n o f stress contagion o r spillover w ithin the erosion m odel. A nother explanation for this path is the hom ophily hy­ pothesis th at suggests persistently distressed m others select or are selected by con-

DflGARMO A N D F O K C a T CH

24 4

QB5-I Tim-1 *5 Tim^)

IIG 10.fi. ODS-1 p rw ra mocel prediaing aipporlrw conidint bthavion, controlling (or r*■ .05 pcoJ tn>m P f G*rffli> I |M7b) wjih pfrtiiijiKin-

"ji kinship intimacy, i ’ (fi7. N = ]}#) = 34,6i-. /»= .07; t.wnpi:»|iv>< hi in < .05) higher than stepchildren’s mean scores. On the SRCI, stepparents’ mean score was significantly lower than their stepchildren’s mean score.

284

FINE, COLEMAN, GANONG

the stepparent should not engage in discipline alone. In interpreting these results, readers should keep in m ind that these are stepfamilies that generally have been to­ gether for a reasonably long period o f time, which may m ean that they have d eter­ m ined what does and does not work for them . Taken as a whole, these responses reflect a stance that stepparents should help discipline stepchildren but their role should be com plim entary (or secondary) to that o f the parent. L eadership a n d Influence. Two codes from the behavioral observation task were relevant to leadership and influence. First, we coded leadership during the family interaction according to w hether an individual, dyad, or a subgroup within the family assumed prim ary leadership for helping the family reach its consensual response. Am ong the 16 stepfamilies, the parent was identified as the leader in 10 instances, the parent-stepparent dyad in 4 families, the stepchild in 1 family, and the stepparent in 1 family. These results are notew orthy because the wife was a parent in all 16 families, although in two o f these families, she was also a step­ m other o f residential stepchildren. Given findings suggesting that m en are m ore likely to take control o f conversations than are wom en (Anderson & Sabatelli, 1995), one m ight expect the stepfather to assume the prim ary leadership role m ore often. However, wives may have exhibited m ore leadership than husbands partly because the task was verbal and interpersonally oriented (Tannen, 1990). In addi­ tion, o ur findings are consistent with the notion that the parent (i.e., the m other) often m ediates and tries to negotiate harm ony between people she loves (i.e., her children and her new husband) but who often do not have strong positive feelings toward each other. Being invested in the greatest num ber of relationships in the family, she may feel the need to ease tensions between her spouse and children. Thus, particularly given the sex o f the stepparents and parents, these results are consistent with the inference that stepparents generally do no t assum e a leader­ ship or m anagerial role in the stepfamily and, when they do, it is typically in con­ junction with the parent. Second, as an index o f influence in the stepfamily, we exam ined which individ­ ual’s rankings m ost closely resembled the family’s consensus rating. O u r criterion for identifying an individual as having the greatest influence on the group decision was that he o r she had the individual responses that m ost resembled the family consensus on at least 2 of the 3 questions to which the family responded. Am ong our 16 families, the stepparent was the m ost influential individual in 7 families, the parent in 5, the stepparent and the parent in 3, and no clear influential individual in the rem aining family. Thus, in contrast to parents providing the m ost leadership in these families, the stepparents were slightly m ore influential than the parents in the process o f influencing a family decision. Perhaps males have m ore influence in affecting the group decision-m aking process than in leading the group interaction, or perhaps m others, in their role as m ediators, acquiesced to their husbands when they disagreed and influenced their children to go along as well. Clearly, m ore work needs to be done to explain these potentially discrepant findings pertaining to

12.

U N D E R S T A N D I N G T H E S T E P P A R E N T ROLE

285

leadership and influence, including an examination of the verbal communication that took place during the family interactions. O bligations o f Stepparents Toward Stepchildren. In the interviews, we asked stepparents, parents, and stepchildren about the obligations and responsibil­ ities they had to their stepchildren. Their responses were coded into four cate­ gories: the stepparent has the same obligations to stepchildren as a parent has to children; the stepparent should play a secondary role to the parent, such as friend, helper, or financial provider, but should not be a parent; the stepparent has obliga­ tions that are similar to a parent in most ways but not all; and the stepparent has no obligations to his or her stepchild(ren). Some quotes may shed light on the meaning assigned to these categories o f re­ sponses. A code o f 1 (i.e., the nuclear family model) was assigned to the response of a stepfather who stated, “. . . try to remember and realize that they shouldn’t be treated any different from your own child. I am obligated to treat them like one of my own.” A m other coded 1 said, “The same basic obligations as the biological par­ ent.” For a code o f 2 (i.e., secondary, supportive role), a stepfather said, “I am ob­ ligated to be honest with her (the stepdaughter), to be there for her, to be trust­ worthy. I owe them respect.” A m other stated, “Be there for emergencies. Child care, when I’m not available. Fix meals when I’m not around. He’s free to say that doesn’t work for me.” A stepdaughter indicated, "I look at him as just someone that’s there for my mom. He has some say i n . . . family activities and financial stuff . . . I guess stepparents are just friendship and support.” For a code o f 3 (i.e., almost a parent), a stepmother stated, “Pretty m uch to be the parent. Give them as much respect and love as possible.” A stepfather said, “They’re my kids. I’m not their fa­ ther, never will be, but as far as I’m concerned, all things given, I feel like I owe them.” A stepdaughter reported, “I think they should be just like the regular parent, but also keep in m ind that they’re not. They’re just not the biological parent.” For a code of 4 (i.e., the stepparent has no obligations), a m other said, “I don't think he has that obligation because they’re not his kids.” A stepson in the same stepfamily declared, “Nothing, because they’re not related, they’re not theirs. It’s nice if they do take care o f them, but they shouldn’t have to.” The results indicated that almost one half (47%) of the stepparents’ responses were coded as 1 (i.e., the nuclear family model), whereas only 27% of both parents’ and stepchildren’s responses were assigned this code. Only one fourth (23%) of stepparents’ responses were coded as 2 (i.e., secondary, supportive role), in con­ trast to 40% and 45% o f parents’ and stepchildren’s responses, respectively. There were no differences am ong stepfamily members in the two remaining codes; 23% to 29% of all three family members’ responses were coded as 3 (i.e., almost a par­ ent) and 0% to 7% were coded as 4 (i.e., no obligations to stepchildren). These results suggest that, even among a sample o f fairly long-term stepfami­ lies, the members o f the oldest dyad in a stepfamily— the parent and the child— have views o f the stepparent’s obligations that differ from those of the stepparent.

286

FINE, C O LE M A N , G AN O N G

In particular, stepparents were more likely to proclaim that their obligations were no different than they would be if the child were their “own” than were either par­ ents or stepchildren. Perhaps the parent and stepchild, because of their shared his­ tory together and experiences adapting to life in a single-parent family, feel that they do not need much from the stepparent. In addition, parents and children also know that another parent outside o f the immediate household may assume some obligations to meet the children’s needs. By contrast, stepparents, perhaps because they may have the unrealistic expectations that stepkin can and should instandy love one another and that the stepfamily is functionally equivalent to a first-marriage family (Visher 8c Visher, 1988), feel that it is im portant that they assume all obli­ gations for the stepchild that one has for a biological child. A stepparent, while knowing that the nonresidential parent is a potential support to the child, may still believe that it is his or her responsibility, as the residential adult, to meet the child’s needs. Gender differences in perceptions o f what it means to be a parent may also be a factor in explaining this difference between parents and stepparents. Men may be more likely than women to define being a parent in terms of providing financially for their stepchildren. Thus, although stepfathers and mothers were equally likely to label the ideal stepparent role as that o f a parent, stepfathers may have defined the term parent in a narrower, more financial way than mothers.

Clarity of the Stepparent Role As shown in Table 12.2, ratings of role clarity were generally in the middle range on the 1-7 scale of the SRCI, suggesting that family members were moderately clear about the role of the stepparent. In addition, two sets of analyses were conducted to compare family members’ role clarity scores. First, a one-factor (stepfamily m ember) repeated-measure ANOVA was conducted on the composite SRCI score. As shown in the table, parents and stepchildren reported having a clearer percep­ tion of the stepparent role than did stepparents. Stepchildren’s perceptions did not differ from those o f their parents. Second, correlations were computed between each pair o f stepfamily members’ SRCI scores. Whereas stepparents’ and parents’ scores were significantly correlated (r = .59), stepchildren’s scores were not corre­ lated with either their parents’ or their stepparents’. A quote from one o f the stepfathers nicely illustrates the lack o f role clarity that many stepparents felt. A stepfather in a long-term stepfamily, when asked if there was a change in his thinking regarding the best way to be a stepparent, stated: I probably never did know how to be one (stepparent). You know, no one ever told me that, “ Hey, you’re a stepparent, and it’s ABC now, and th at’s your responsibility, th at’s your role, these are the things you should do, these are the things you shouldn’t do, and these are the things th at you should look o u t for. So, nobody gave m e that ad ­ vice. As a m atter o f fact, 1 do n ’t rem em ber anybody giving m e any advice at all about that. But it’s like, “Wait a m inute. T here are som e things here th at som ebody should

12.

U N D E R S T A N D I N G T H E S T E P P A R E N T ROI-F.

287

have told me somewhere, that you have to look out for this” or “This is just a natural part o f being a stepparent.”

Perceived Stepparent Role Behavior Affinity-Seeking and Affinity-M aintaining Behaviors. Using o u r in ter­ view data, we explored the n atu re o f th e stepparents’ attem p ts to elicit liking from the stepchildren a n d th e success o f these efforts in developing an d m aintaining close stepparent-stepchild relationships (G anong et al., 1997). T here were three relatively distinct p atterns o f affinity-seeking and affinity-m aintaining strategies am ong the stepparents in o u r sample. First, nonseeking {NS) stepparents did no t do anything intentionally to elicit lik­ ing from their stepchildren an d seldom th o u g h t m uch ab o u t this issue. These step­ parents were n o t unfriendly to th eir stepchildren; in fact, usually in group outings and activities, they were often quite nice to the children w ho w ould later be their stepchildren. However, th e prim ary goal o f these friendly in teractio n s was to solid­ ify the ro m an tic relationship w ith th e child’s parent, n o t th e relationship w ith the child. These stepparents d id n o t seem to think th at eliciting liking from stepchil­ dren was an im p o rtan t goal for them . Second, early affinity-seeking (EAS) stepparents were m otivated and tried to elicit liking from the stepchildren early in the relationships (typically w hile they were also trying to achieve o th er goals, such as co u rtin g the children’s parents), b u t they stopped doing so relatively early in the relationships (usually after they began to share residences). These stepparents generally th o u g h t o f them selves as the chil­ d re n ’s new p arents and they identified th eir role as th a t o f being a p aren t (n o t a stepparent) to the child. O u r sense was th at these stepparents sto p p ed trying to elicit liking from th eir stepchildren prim arily because they did n o t believe th a t p a r­ ents should have to actively try to get th eir children to like them . T hird, continuous affinity-seeking (CAS) stepparents deliberately sought to elicit liking from th eir stepchildren-to-be, and they persisted in these efforts (although at a decreased frequency) after the stepfam ily was form ed. These stepparents th o ught a b o u t how they could get their stepchildren to like them , they did things w ith their stepchildren th at the stepchildren w anted to do (as opposed to w hat the stepparent w anted to do o r th o u g h t th e stepchildren should d o ), an d they seem ed genuinely interested in establishing close relationships w ith th e stepchildren. It is n o t su rp ris­ ing that these stepparents also were m ore likely to have stepchildren w ho themselves tried to elicit liking from th eir stepparents. Quality o f Stepparent-Stepchild Interaction. From th e behavioral obser­ vation tapes, we com pared o u r global ratings o f the quality o f step p arent-stepchild interaction (on th e d im ensions o f affect, a m o u n t o f interaction, co operation, and conflict) w ith those o f th e p a re n t-c h ild interaction. Across th e 16 families, we con­ sistently rated p a re n t-c h ild interaction m ore positively th an th at betw een step­

288

FINE, C O LE M A N , GANONG

parents and stepchildren on the dimensions of affect, am ount o f interaction, and cooperation, although there were no differences in conflict between the two types of dyads. Thus, the most salient difference was not that stepparent-stepchild inter­ action was more antagonistic than parent-child interaction but rather that it was characterized by less positive affect, interaction, and cooperation.

Adjustment Correlates of Role Constructions We examined the adjustm ent correlates of three aspects of stepfamily members’ role constructions by correlating the dyadic-level consistency scores with the adjustm ent scores, correlating scores from the SRCI with the adjustm ent scores, and comparing our independently assessed levels of closeness in the stepparentstepchild relationship (from the interviews) across the three categories o f step­ fathers derived from the affinity-seeking and affinity-maintaining data. A d ju stm e n t Correlates o f Consistency in C o n ten t o f the Stepparent Role. To examine whether there was a link between the extent to which stepfamily m em ­ bers had consistent role perceptions and adjustment, we correlated the ideal and actual warm th and control role discrepancy scores (stepparent -parent and step­ parent-stepchild) with the adjustment scores. Because o f the large num ber of cor­ relations, an alpha level of .01 was used. Tables presenting these correlations can be found in Fine et al. (1998). All significant correlations were in the expected direction (i.e., larger discrep­ ancy scores were related to poorer adjustm ent). Higher stepparent-parent ideal discrepancy scores were related to lower levels o f stepparents’ satisfaction with stepparenting (only on the warm th discrepancy score), parents’ reports o f less close stepparent-stepchild relationships, parents’ (only on warm th) and steppar­ ents’ perceptions of less success as a stepparent, and parents' and stepparents’ re­ ports of fewer family strengths (only on warm th). Few of the correlations between the stepparent-stepchild ideal role discrepancy score and the adjustm ent measures were significant. W ith respect to correlations between the actual warm th and control discrep­ ancy scores and the adjustm ent scores, the actual stepparent-parent discrepancy score was related in the hypothesized direction to many of the interpersonal and family-level adjustm ent measures. Smaller discrepancy scores were related to step­ children’s reports of fewer mental health symptoms (only on the control discrep­ ancy score), greater life satisfaction (for stepparents on control and for parents on warm th), stepparents’ greater satisfaction with stepparenting, closer stepparentstepchild relationships, perceived success as a stepparent, greater marital satisfac­ tion (only for stepparents on warm th), and more family strengths. The correlations between the actual stepparent-stepchild discrepancy scores and the adjustment measures were generally nonsignificant.

12.

U N D K R S T A N D I N G T H K S T K P P A R E N T R O I. K

28 9

A d ju stm e n t Correlates o f Role Clarity. Several patterns were found in the correlations between the role clarity and adjustm ent scores (see Fine et al., 1998). First, stepparents’ and parents’ role clarity scores, but not stepchildren’s, were re­ lated to the perceived closeness o f the stepparent-stepchild relationship and per­ ceived success as a stepparent. Second, stepparents’ role clarity scores were related to all three family members’ perceptions of greater family strengths. Finally, step­ parents’ and parents’ role clarity scores were related to greater satisfaction with stepparenting. In sum, there was m oderate support for the notion that stepparent role clarity is positively linked to the interpersonal dimensions of adjustm ent in stepfamilies. A d ju stm en t Correlates o f A ffin ity-S eekin g a n d A ffin ity -M a in ta in in g Stra­ tegies. The three groups o f stepparents defined by differences in affinity-seeking and affinity-maintaining behaviors differed substantially in the ratings we as­ signed to the closeness of the stepparent-stepchild relationship. In all but one in­ stance, continuous affinity-seeking (CAS) stepparents had warm and close bonds with their stepchildren. The early affinity-seeking (EAS) stepparents had relation­ ships with their stepchildren that varied in closeness. Those who had close bonds with their stepchildren— as did the CAS stepparents— had stepchildren who themselves recognized and reciprocated the stepparent’s affinity-seeking efforts. EAS stepparents who were not close to their stepchildren generally had stepchil­ dren who did not know that their stepparents had tried to get them to like them. In addition, the early disciplinary attempts of some EAS stepparents were often re­ sented by stepchildren, particularly older ones, and seemed to reduce closeness in the stepparent-stepchild relationship. It is not surprising that stepchildren of n on­ seeking (NS) stepparents were consistently not close to their stepparents, nor did they acknowledge or reciprocate the stepparent’s infrequent attempts to elicit lik­ ing from them.

DISCUSSION O ur inquiry into the way that the stepparent role is socially constructed by step­ family members was enhanced because we used three different sources of informa­ tion: quantitative surveys, qualitative semistructured interviews, and behavioral observation. Few studies of stepfamilies have used all three o f these approaches, and to our knowledge, previous researchers studying the stepparent role have not tried to integrate findings across these diverse methods.

Constructions of the Content of the Stepparent Role W hat have we learned about how the content of the stepparent role is constructed by stepfamily members? First, consistent with the notion that there is a lack of

290

FINE, C O LE M A N , G AN O N G

social consensus related to the role o f the stepparent (Cherlin, 1978; CrosbieBumett, 1989), we found considerable variation in perceptions o f the appropriate label for the actual and the ideal stepparent role. Although nearly half the parents and stepparents labeled the idea! and current role o f stepparents to be that of a par­ ent, the remaining half assigned varying labels— including stepparent, friend, and a variety o f other labels— to both the ideal and current stepparent role. Second, our results converged toward the conclusion that stepparents generally believe that they should play a more active role in parenting than do their step­ children and, on some dimensions, than their spouses. For example, stepparents (and parents) were more likely to report that the actual stepparent role or ideal step­ parent role was that o f parent or stepparent than were stepchildren, and stepparents reported that they should engage in warmth and control parenting behaviors more often than did stepchildren. Similar results were obtained from the interview data pertaining to perceptions o f stepparents’ obligations to their stepchildren. Step­ fathers were more likely than either parents or stepchildren to report that they had the same obligations as did fathers, which is partially consistent with Ganong et al.’s (1995) finding that stepparents are believed to have financial obligations to their residential stepchildren as long as they are married to the children’s parents. Collectively, our results suggest that the adults in a stepfamily often attem pt to recreate their families in the image of a first-marriage, nuclear family. Rather than constructing an image o f how the stepparent should function that is tailored to the unique aspects of the stepfamily situation, most adults believe that the stepparent should function as a parent or as a parent-like figure to the stepchild. O ur findings further indicate that the stepparent is the family member most likely to adopt the nuclear family model and that the stepchild is the least likely to do so. This conclu­ sion is strengthened by a finding not reported previously in this chapter; almost 40% of the stepparents reported that they did not consider themselves to be step­ parents, despite the fact that only one of the stepparents had legally adopted her stepchildren. Although a num ber of stepfamily scholars have speculated that this construction of the stepfamily as a nuclear family may be maladaptive (Visher 8c Visher, 1988), our results do not directly address this issue. In terms of leadership and influence, we found some tentative evidence that the parent assumes the major responsibility for leadership, serving as a manager of family activities, but that both the stepparent and parent have influence on the family decision-making process. In the behavioral observation task, we rated the m other as the leader in facilitating the group process the majority of the time but also found that the stepfather and the mother both had influence in the decision­ making process.

Constructions of the Clarity of the Stepparent Role Stepfamily members reported having moderately clear perceptions of the step­ parent role. Stepparents reported being'less clear about the stepparent role than did parents and stepchildren. Because stepparents live the role o f the stepparent on

12.

U N D ER STAN D IN G

T H E S TEPPA R EN T ROLE

291

a daily basis and m ust continually make decisions about how to behave towards their stepchild(ren), it makes sense that they would be m ore aware o f their uncer­ tainties concerning how they should behave than w ould o ther family mem bers. In addition, because the stepparent-parent dyad is m ost likely to discuss the steppar­ ent role and how it should be enacted, it is not surprising that their role clarity scores would be highly intercorrelated. By contrast, stepchildrens views appear to be unrelated to those o f their parents and stepparents, which may be developm entally characteristic o f children in a variety of family structures. W hen roles lack clarity, family m em bers can redefine the role in a way that a t­ tem pts to increase role clarity to them or they can creatively construct role defini­ tions out o f family discussions. About h alf the adults in our sample may have ch o ­ sen the form er m ethod to resolve a lack o f clarity surrounding the stepparent role. By defining the role o f stepparent as that o f a parent, family m em bers may hope to increase family consensus regarding the role, thereby reducing interpersonal diffi­ culties. The rem aining half o f the adults and the m ajority o f stepchildren, however, chose other labels to describe the stepparent’s current and ideal role.

Perceived Stepparent Role Behavior As noted earlier, m any stepparents reported that they believed that they should act as a parent to their stepchildren. However, we noted from the interviews that many did so in a m ore lim ited way than did the parent. This lesser involvement by some of the stepparents was reflected in a num ber o f ways: Some reported that they had not m ade m uch o f an effort to be involved in their stepchildren’s lives, with some regretting this choice later. A sizeable m inority m ade no deliberate attem pts to elicit liking from their stepchildren, and several others m ade initial attem pts but did no t continue these efforts after the remarriage. O ur behavioral observation data also showed that stepfathers and stepchildren had interactions that were char­ acterized by less positive affect, engagement, and cooperation than were inter­ actions between parents and children. The apparent discrepancy between constructions o f the content o f the steppar­ ent role and perceived role behavior may be due to variations in constructions o f what it m eans to be a parent. If stepparents thought o f being a parent prim arily in term s o f financially supporting their stepchildren, then their behavior was not in ­ consistent with their reported role construction. However, if they thought o f a p ar­ ent as som eone who provides em otional support and disciplines stepchildren, in addition to financially providing for them , then the discrepancy takes on greater salience.

Relations Between Role Content, Clarity, and Perceived Behavior and Adjustment in Stepfamilies O u r results provide som e initial, m oderate support for the notion that consistency am ong stepfamily m em bers in perceptions of the stepparent role is related to a

292

FINF., C O I .F . M A N , G A N O N G

num ber o f interpersonal dim ensions o f adjustm ent in stepfamilies. In particular, perceptions o f how the role is currently being enacted were m ore strongly related to adjustm ent in stepfamilies than was consistency in perceptions o f the ideal step parent role. Because the stepparents behavior affects family m em bers on a regular basis, family m em bers m ore likely are aware of, express, and discuss their per­ ceptions o f how the stepparent is currently behaving than they are aware o f their perceptions o f how the stepparent should behave, which increases the salience of consistency in these perceptions. As predicted, the m ore confident stepparents were about how to interact with their stepchildren and the m ore sure parents were about how their spouse should relate to their child, the m ore satisfied they were with a num ber o f interpersonal aspects o f stepfamily life. Stepchildren’s perceptions of the clarity o f the stepparent role, in contrast, were unrelated to their satisfaction, suggesting that they are less bothered by the lack o f clarity regarding the stepparent role than are adults. Clarity about the stepparent role may have less relevance for stepchildren than for adults because the stepchildren have less control over how the stepparent role is per­ form ed than do parents and stepparents. In aggregate, these studies suggest that constructions o f the stepparent role are related to the risk the family m em bers face for poor adjustm ent and to their resi­ lience in adapting to the challenges o f stepfamily life. In particular, our findings suggest that consistent views o f the stepparent role, role clarity, and active and con­ tinuous affinity-seeking and affinity-m aintaining efforts by the stepparent serve to reduce risk and enhance resiliency to the inevitable stresses of living in a stepfam ­ ily. In the next section, we present suggestions for future research that we believe will extend our knowledge o f factors that lessen risk and enhance resilience.

DIRECTIONS FOR FUTURE RESEARCH ON THE STEPPARENT ROLE O ur findings provide som e direction for future research into the m anner in which the stepparent role is constructed. First, investigators need to gather richer and m ore detailed inform ation about perceptions o f the stepparent role. For example, because one com m on them e that emerged from our project was that stepchildren have perceptions o f the stepparent role that are quite different from those o f their parents and stepparents, future researchers could profitably exam ine these differ­ ing perceptions in greater depth, perhaps with qualitative interviews. In addition, it could be fruitful to examine what m en m ean when they state that they should treat the stepchild as their own. W hat kinds of responsibilities and obligations does that entail for these men? Second, with longitudinal designs, we would be able to determ ine if and how perceptions o f various aspects o f the stepparent role change over the life course of the stepfamily. This is a critical issue to exam ine because som e ways of construing

12.

U N D ER S TA N D IN G

THE

STEPPARENT

ROI.E

293

the stepparent role m ay be grow th producing at one stage o f the fam ily life course, w hereas a different role construction m ay be necessary at a later stage or as cir­ cum stances change. T hird, although we touched on this tangentially in the sem istructured in ter­ views, future researchers w ould do well n o t only to include the nonresidential p a r­ ent in data collection efforts b u t also to explore how this p arent and his or her role perceptions a n d behavior affect functioning w ithin the stepfam ily household. O ur sense was that the nonresidential paren t often influenced the n ature and quality o f the stepparent-stepchild relationship and, thus, the m an n e r in which the role of the stepparent is constructed. Finally, and perhaps m ost im p o rta n t, if subsequent studies d o cu m en t b o th con­ cu rren t and prospective relations betw een role perceptions and adjustm ent in step­ families, researchers need to tu rn their attention to trying to identify the m echa­ nism s underlying these relations. To w hat extent do stepfam ily m em bers discuss their perceptions o f the stepparent role, and w hat are the outcom es and conse­ quences o f these discussions, if they occur? How do constructions o f the step­ p arent role and changes in these constructions affect later adjustm ent? In what ways do these constructions affect role behaviors?

REFERENCES Anderson, S. A., & SabateUi, R. M. (1995). Family interaction: A multigenerational developmental per­ spective. Needham Heights, MA: Allyn 8c Bacon. Beavers, W. R., Hampson, R. B., 8c Hulgus, Y. F. (1985). The Beavers systems approach to family assess­ ment. Family Process, 2 4 ,398-405. Borduin, C. M. (1990). Observational coding system: University o f Missouri- Columbia Delinquency Pro­ ject. Unpublished manuscript. Bray, J. H., Bergen S. H., & Boethel, C. L. (1994). Role integration and marital adjustment in stepfather families. In K. Pasley 8c M. Ihinger-Tallman (Eds.), Stepparenting: Issues in theory, research, and prac­ tice (pp. 69-86). Westport, CT: Greenwood. Buehlman, K. T., Gottman, J. M., 8c Katz, J. F. (1992). How a couple views their past predicts their future: Predicting divorce from an oral history interview. Journal o f Family Psychology, 5, 295-318. Cherlin, A. (1978). Remarriage as an incomplete institution. American Journal o f Sociology, 84,634-650. Crosbie-Burnett, M. (1989). Application o f family stress theory to remarriage: A model for assessing and helping stepfamilies. Family Relations, 3 8,323-331. DeRogatts, L. R. (1993). BriefSymptom Inventory: Administration, scoring, and procedures manual. M in­ neapolis, MN: National Computer Systems. Diener, E., Emmons, R. A., Larsen, R. J., 8c Griffin, S. (1985). The satisfaction with life scale. Journal of Personality Assessment, 49, 71-75. Erera-Weatherly, P. I. (1996). On becoming a stepparent: Factors associated with the adoption of alter­ native stepparenting styles. Journal o f Divorce and Remarriage, 2 5 ,155-174. Fine, M. A., Coleman, M.> 8c Ganong, L H .(1 998). Consistency in the perceptions of the stepparent role among stepparents, biological parents, and stepchildren. Journal o f Social and Personal Relation­ ships, 15,811-829. Fine, M. A., Ganong, L. H., 8c Coleman, M. (1997). The relation between role constructions and adjust­ ment among stepfathers. Journal o f Family Issues, 18,503-525.

294

FI NE, COLEMAN, GANONG

Fine, M. A., 8c Kurdek, L. A. {1994a). A multidimensional cognitive-devclopmental model of stepfamily adjustment. In K. Pasley 8c M. Ihinger-Talfman (Eds.), Stepparenting: Issues in theory, research, and practice (pp. 15-32). Westport, CT: Greenwood. Fine, M. A., & Kurdek, L. A. ( 1994b). Parenting cognitions in stepfamilies: Differences between parents and stepparents and relations to parenting satisfaction, journal of Social and Personal Relationships, It, 95-112. Fine, M. A., Kurdek, L. A, 8c Hennigen, L. (1992). Perceived self-competence and its relations to step­ family myths and (step)parent role ambiguity in adolescents from stepfather and stepmother fami­ lies. Journal o f Family Psychology, 6 ,69-76. Ganong, L. H., 8c Coleman, M. (1989). Stepfamily Unrevealed Differences Questionnaire. Unpublished instrument, University of Missouri, Columbia. Ganong, L. H., & Coleman, M. (1994). Adolescent stepchild-stcpparent relationships: Changes over time. In K. Pasley & M. Ihinger-Taliman (Eds.), Stepparenting: Issues in theory, research, and practice (pp. 87-104). Westport, CT: Greenwood. Ganong, L H., Coleman, M., Fine, M. A., 8c Martin, P. (in press). Stepparents’ affinity-seeking and affinity-maintaining strategies with stepchildren. Journal of Family Issues. Ganong, L. H.> Coleman, M., 8c Mistina, D. (1995). Normative beliefs about parents' and stepparents’ financial obligations to children following divorce and remarriage. Family Relations, 44,306-315. Kahn, R. L, Wolfe, D. M., Quinn, R. P., Snoek, J. D., 8c Rosenthal, R. A. (1964). Occupational stress: Stud­ ies in role conflict and ambiguity. New York: Wiley. Kurdek, L. A., 8c Fine, M. A. (1991). Cognitive correlates of adjustment for mothers and fathers in step­ father families. Journal of Marriage and the Family, 53,565-572. Marsiglio, W. (1992). Stepfathers with minor children living at home: Parenting perceptions and rela­ tionship quality. Journal of Family Issues, 13,195-214. Olson, D. H., Larsen, A. S., 8c McCubbin, H. I. (1982). Family strengths. In D. H. Olson, H. I. McCubbin, H. Barnes, A. Larsen., M. Muxen, 8c M. Wilson (Eds.), Family inventories: Inventories used in a na­ tional survey offamilies across the family life cycle (pp. 121-134). St. Paul, MN: University of Min­ nesota, Department of Family Social Science. Patton, M. Q. (1990). Qualitative evaluation and research methods. Newbmy Park, CA: Sage. Schumm, W. R., Paff-Bergen, L A., Hatch, R. C., Obiorah, F. C., Copeland, J. M., Meens, L. D., 8c Bugaighis, M. A. (1986). Concurrent and discriminant validity of the Kansas Marital Satisfaction Scale. Journal of Marriage and the Family, 48,381-387. Schwebel, A. I., Fine, M. A., & Renner, M. A. (1991). An empirical investigation of perceptions of the stepparent role. Journal of Family Issues, 12,43-57. Sweet, J. A., Bumpass, L. L., & Call, V. R. A. (1988). The design and content of the National Sun'ey of Fam­ ilies and Households (Working Paper NSFH-1). Madison, WI: University of Wisconsin, Center for Demography and Ecology. Tannen, D. (1990). You just don't understand: Women and men in conversation. New York: Ballentine. Thomas, V., 8c Olson, D. H. (1993). Problem families and the Circumplex Model: Observational assess­ ment using the Clinical Rating Scale (CRS). Journal of Marital and Family Therapy, 19,159-175. Visher, E., 8c Visher, J. S. (1988). Old loyalties, new ties. New York: Brunner/Mazel. Visher, E. B., 8c Visher, J. S. (1996). Therapy with stepfamilies. New York: Brunner/Mazel.

13

The Dynamics of Parental Remarriage Adolescent, Parent, and Sibling Influences

E dw ard R. A n derson S h a n n o n M . G reene Arizona State University E. M avis H e th erin g to n University o f Virginia W. G lenn C lingem peel Francis Marion University

c h a p t e r is concerned with the transition o f parental remarriage and its implications for children’s well-being. We start by examining adolescent adjust­ m ent in remarried families, focusing on externalizing, which includes acting out, antisocial behavior, noncompliance, and aggression. Consistent with many other studies of remarriage, we dem onstrate that adolescents living in stepfamilies are at higher risk for these problematic behaviors than those living in nondivorced households.1We then examine possible reasons for why the higher levels o f exter­ nalizing occur in stepchildren, including the role of stressful life events for adoles­ cents and the quality of family relationships. A major theme from these analyses is the reciprocal relation between adolescent adjustm ent and family process mea­ sures. Said another way, adolescents in stepfamilies are im portant products o f and contributors to the stepfamily experience. T h is

1We use the term nondivorced families to refer to families where the two residential parents are in an original marriage. 295

296

A N D E R S O N E T AL.

THE LONGI TUDI NAL STUDY OF REMARRI AGE The data presented here com e from the H etherington and Clingem peel lo n g itu d i­ nal Study o f Rem arriage (LSR; H etherington et al., 1992), an intensive study of 202 families w ith a child betw een the ages o f 9 and 13. A pproxim ately one th ird o f the families were those in w hich the m o th e r had divorced, retained custody o f her children, and was in the first few m o n th s o f a rem arriage. A nother th ird o f the fam ­ ilies consisted o f divorced m others w ho had n o t rem arried. They had been d i­ vorced for approxim ately the sam e length o f tim e as the rem arried m others. The rem aining group o f families contained couples w ho were n o t divorced. All families w ere Caucasian, and parents had at least a high school education. Fam ilies were as­ sessed three tim es over the course o f a 2-year period, corresponding to 4 ,1 7 , and 26 m onths after their rem arriages. T he LSR used an intensive battery o f instrum ents designed to yield com parable inform ation across m ultiple inform ants, derived from stru ctu red interview s and standardized assessments. For exam ple, perceived quality o f the stepfather-stepchild relationship was obtained from each adolescent, stepfather, and m other. In addition, family m em bers were videotaped in dyads and triads, and these tapes were rated by trained observers across various aspects o f global family functioning. Family functioning was derived from com posites form ed from observational m easures and self-report m easures from m ultiple inform ants. Adolescent exter­ nalizing behavior was derived from m others, (step) fathers, teachers, adolescents, a nd videotaped observations o f adolescents interacting w ith family m em bers. A com posite m easure o f externalizing was created by aggregating these m easures in standard score form .

ADOLESCENT ADJUSTMENT IN R E MARRI ED FAMILIES Do adolescents in stepfam ilies fare less well than adolescents living in continuously m arried families? An estim ate o f the total effect o f rem arriage on adolescents’ ex­ ternalizing behavior is m ade by using the m u ltim ethod, m u lti-in fo rm an t m easure o f externalizing and creating an effect code (Aiken & West, 1991) representing the classification o f rem arried versus nondivorced. Using the com posite m easure as­ sessed at 4 m onths after rem arriage, the effect code represents the total average dif­ ference in initial level o f externalizing across rem arried and nondivorced groups. In addition, because the study was longitudinal, we can exam ine the influence of rem arriage on change in externalizing. T his latter question is addressed th rough the use o f grow th curve analysis (e.g., M cArdle & A nderson, 1990). For each su b ­ ject, a linear grow th curve is calculated based on available data. T his procedure provides a slope score for each adolescent, which indicates the rate o f change in ex­ ternalizing across the 2 years o f the study. An effect code for rem arriage that signi­

13.

PARENTAL REMARRIAGE, ADOLESCENT ADJ UST ME NT

297

ficantly predicts this slope score indicates that adolescents in rem arried families increase in externalizing at significantly faster rates than counterparts in n ondi­ vorced families. The correlation between the effect code and outcom e was .36, m eaning that the experience of remarriage explains 13% of the variance in adolescent externalizing behavior 4 m onths into the remarriage. The longitudinal relation o f P = .23 indi­ cates that adolescents in rem arried families grow at faster rates in externalizing than adolescents in nondivorced families, even after controlling for the initial level of behavior. Thus, differences in externalizing between adolescents in stepfamilies and nondivorced families are apparent in the first m onths after remarriage, and on aver­ age, externalizing increases at a faster rate for adolescents in rem arried families.

WHAT ACCOUNTS FOR ADOLESCENT ADJUSTMENT IN STEPFAMILIES? The prim ary theoretical question, then, is to identify those factors that account for the differences in adolescent externalizing behavior between rem arried and n o n ­ divorced families. In this section, we address how co-occurring stressful life events and distressed stepfamily relationships may account for these differences in adoles­ cent adjustm ent. First, we examine whether life events and relationship quality are associated with adolescents’ externalizing behavior w ithin rem arried families. The focus is to identify those factors that explain externalizing behavior for adolescents across rem arried families. Next, we examine w hether life events and relationship quality account for differences in externalizing between rem arried and nondi­ vorced families. These factors are exam ined in an attem pt to remove the betweengroup difference identified by the effect code described previously. It is possible, however, to identify a factor that leads to improved functioning for stepfamilies but w ithout enough o f a boost to close the gap between rem arried and nondi­ vorced families completely. The role of co-occurring stressful life experiences and distressed relationships (i.e., parent-child, current m arital, form er m arital, and sibling) in explaining children’s externalizing behavior each are considered in turn.

Stressful Life Events and Adolescent Adjustment In addition to establishing relationships with stepparents, adolescents whose cus­ todial parent remarries often encounter a variety of significant other transitions. In order to assess these transitions, we asked adolescents to identify which o f 57 possible events or relationship changes had occurred since the rem arriage (or since the last interview in later waves) and to rate their corresponding reactions on a 7point scale from - 3 (extremely negative) to +3 (extremely positive). Perhaps the most telling finding lies in adolescent responses to a single item on this inventory, “parent rem arried.” The mean stress rating was a positive 1.55,

298

ANDERSON t r A L .

meaning (hat, on average, an adolescent experiences the custodul parent^ reraajtiage as moderaiely positive, Individ ual response's tanged from - 3 1o + j, indicating a wide diversity in adolescent reactions 10 tlm transttion. Overall, though, 65% of adoicscraU scored remarriage at or above +2, whereas only 12% gave a negsliwc rating < ■l t -2, or i) to this event. Because this question was asled only once. we can not tell it ral ings would have changed over time. It is not surprising lhat the event most commonly associated with parental re ­ marriage involved relocation, with abnyi half (52%J moving lo a new residence, tor many, the move required an additional transition: surtmj; a new icJiool. Yet only about 20% o f adolescents in remarried families pcrceived these experiences as negative, with (he average ratin g for b o th moving and starting h new school a< +1. 5 (i.e., a moderately positive evenl). Additionally, important differences between adolescents in nundiwxreed and reniaiTied families emerged As opposed to counterparts in noridivotted families, adolescente in remarried families reported more arguments with iheir parents 4 months after the rcmamaflc. 'links difference may reflect in part increases in stresses and challenges that the parents themselves encounter early in remarriage. In contrast, one year later the adoiesceti ts in the n o n d i v o K c d families reported an increase in sixdi arguments. figure 13.1 shows ihj.t d u n ge in frequency of family arguments. Perhaps adolescents in remarried farm lies were merely experiencing the nor n a ­ tive increase in difficult parent-adolescent relationships earlier than other adoles­ cents because o f the reorganization required hylhe remarriage. An analysis of this

Wave of assessment FIG. IV I

C fia n p in frc q u rw y p f aijjurr^nl* lirrw rrr pcircnlj i r d a d n lr v m l) .

13.

PARENTAL REMARRIAGE, ADOLESCENT ADJUSTMENT

299

idea o f accelerated developm ent has been described previously (A nderson, H eth ­ erington, & Clingem peel, 1989). In that paper, we dem onstrated that associations com m only found in nondivorced families betw een the child’s p ubertal status and p a ren t-ch ild relationship quality did n o t occur for stepfam ilies in the first m o n th s after rem arriage. It is possible th at the reorganization required by the rem arriage tem porarily d isrupted this norm ative developm ental process. Alternatively, the de­ crease in argum ents could be a sign o f increasing disengagem ent from the step­ family, as has been found in o th er stepfam ily studies (e.g., H etherington, Cox, & Cox, 1985). As opposed to cou n terp arts in nondivorced families, by the end o f the first year, adolescents in rem arried families were m ore com m only reporting fewer m oney problem s for th eir families, an undoubtedly welcom e experience. Several o f the adolescents in rem arried families reported the b irth o f a sibling to the new family, an experience less likely to have occurred in nondivorced families. Interestingly, the new sibling arrival was universally positive; none o f the adolescents in the sam ­ ple w ho had this experience rated it negatively. By 2 years into the rem arriage, m any adolescents also reported that their parents started school, again an experi­ ence less likely to occur in nondivorced families. O nly a sm all p ro p o rtio n (15%) viewed this change as negative. Finally, as com pared to nondivorced families, 2 years into the rem arriage, adolescents reported m ore frequent dates, a positively rated experience offset by a negative— m ore frequent fights w ith boyfriends or girlfriends. F urther analyses o f the LSR provide co ntinued su p p o rt for adolescents viewing their custodial p aren t’s rem arriage positively. As p a rt o f the LSR, adolescents could identify from a list o f adjectives any that described their reactions to rem arriage, such as “happy,” “angry w ith m other,” “hurt,” or “pleased.” The m ost frequent were all positive; “happy,” checked by 82% o f adolescents, “satisfied” (75% ), and “pleased” (66% ). The m ost frequently checked negative item was “jealous,” experi­ enced by only 18% o f adolescents. M others provide additional evidence for the generally positive adolescent re­ actions to rem arriage. Using the sam e list o f adjectives, m others described their children’s responses to the rem arriage as well as to the earlier divorce. A lthough m others generally rated adolescent reactions to the divorce as negative (i.e., “sad,” “hurt,” and “fearful”), reactions to rem arriage corroborated views expressed by adolescents— “happy,’’“satisfied,” a n d “pleased.” In sum m ary, the vast m ajority o f adolescents view th eir custodial parents’ re­ m arriages as positive events. O nly a sm all m inority (15-20% ) expressed negative attitudes. H ow m ight these relatively positive attitudes tow ard rem arriage be rec­ onciled with the significantly higher levels o f behavior problem s show n by adoles­ cents in stepfamilies? Adolescents entering stepfam ilies have experienced m any transitions before the legal rem arriage: conflict in the previous m arriage, parental separation and d i­ vorce, cohabitation o f the biological m o th er and stepfather o r o th er partn ers

A'aDfcRSOM F.T Al..

before remarriage (experienced hy nKist of the *lolescents in the I.SR steplamilies).and transition into a stepfamily. Additionally, many custodial nunhers in ihe I.SR had dated extensively prior to remarriage, and earlier analyses of (hi* dataset ■showed that a history o f frequent dating related 10 less positive adolescent adjust­ ment and more difficult parcnt-adnlcscenl relationships after the remarriage (Montgomery, Anderson, HettwjinptOu, Ct Quigemptel, ]9 0 2 ). Moreover, jiiin u adolescents in the study had also experienced the remarriage o f their noncustodial lather. The cumulative impact ot these transition s m t evident The lowest levels of adolescent social coropcintw were observed when aU three marital transitions had occurred— divorce, ihe custodial parent's reinamap*, and the noncustodial par­ ent's remarriage (Fig. 13.2). To examine how the experience o f (ra nsitionsaccompanying u-marriage related to an adolescent's adjustment m the l-SH, we developed indicators of the general level o f stress expetienced by adolestenis. Wc distinguished between life events typically experienced as negative b> the majority o f childten, such as gelling sick, damage to home, and death o f a family member, and events that related to adjustii»g to new situation* and family members (i.e., mure commonly experienced by children in stepfainilies}. These latter events included moving to a new home, starting at a new school, birth of a new sibling, and the stepparent moving in, To creaie an indicator o f stress, wesimplv counted the number o f events thai occurred within these two categories. Adolescents in remarried and nondivorced families

C -5 -i

04-

None

One Two Nurrber cf family transitions

Three

I Kj J J ?■ A c H e s rtn ts 'w c ijI co m p v tn K f as a functiun o f nuenfcn o f m > ritil transitions.

13.

PARENTAL REMARRIAGE, ADOLESCENT ADJUSTMENT

301

T A B L E 13.1 Effect of Stressful l ife Events on Initial 1-evel and Change in Adolescents’ Externalizing Regression Coefficients Initial Level

Predictors Negative stresses Transition events Parental remarriage Initial level of externalizing R2 d.f. F

Change

Without Remarriage

With Remarriage

Without Remarriage

With Remarriage

.41*** .14’ — — .19 2.124 14.1***

.38*** .01 .31*** — .26 3,123 14.4***

.05 .12 — -.45’ ** .19 3,103 7.9***

.06 .02 .22* -.51*** .22 4,102 7.3***

Note. ’ p< .10; *p< .05; ’ *p< .01; ""'/>< .001.

did n o t differ in the total n u m b er o f negative stresses experienced, b u t adolescents in rem arried families had significantly higher levels o f transition events. As show n in Table 13.1, the n u m b er o f negative stresses was significantly related to adolescent externalizing behavior. The n u m b er o f transitions experienced, how ­ ever, was only m arginally related to externalizing. In longitudinal analyses, neither transition events n o r negative stresses predicted change in externalizing beyond w hat was accounted for by initial level. W hen the effect code for parental rem arriage was entered, negative stresses con­ tinued to relate significantly to externalizing problem s. M ore im p o rta n t, the n u m ­ b e r o f negative stressful events did n o t dim inish the effect o f parental rem arriage. In other words, the negative im pact o f life stress on adolescent externalizing p ro b ­ lem s was independent o f any effect o f parental rem arriage. In longitudinal analy­ ses, parental rem arriage continued to p redict increases in adolescent externalizing after initial levels o f externalizing and m easures o f stressful events were entered. H ow can this finding be explained? First, the n u m b er o f stressful events clearly m atters for adolescent adjustm ent. Figure 13.3 shows the level o f externalizing problem s for adolescents w ith low (0 ,1 , o r 2) and high (3 or m ore) num bers o f stresses. In stepfam ilies, those w ith m ore stresses fared less well than those w ith fewer stresses. This finding was also tru e for adolescents in nondivorced families. Im portantly, though, the difference betw een family context rem ains. Adolescents in rem arried families w ith low stress were approxim ately equivalent to adolescents in nondivorced families w ith high stress. Stress m attered, b u t it did n o t close the gap in functioning. In sum m ary, adolescents experiencing th eir custodial parents’ rem arriages e n ­ c ounter a n u m b er o f co-occurring life transitions, such as m oving to a new hom e,

302

A N D E R S O N EX' AL.

Low stress

High stress

FIG. 13.3. Adolescents’externalizing behavior in r.ondivorced and remarried fam­ ilies as a function o f low and high numbers of stressful life events.

starting a new school, o r adjusting to living with new siblings. M ost adolescents, however, do not perceive these events as negative, although a m inority, perhaps 15%, do. Im portantly, it is n o t the cumulative experience o f transitions that ex­ plains the elevated levels o f conduct problem s lor adolescents in rem arried fam i­ lies com pared to nondivorced families. Adolescents who experience m ore negative events do show m ore problem s, but the difference in adjustm ent between types of households remains. In the next section, we examine aspects o f parent-adolescent relationships that m ight account for differences in adolescent adjustm ent between stepfamily and nondivorced households.

Parent-Adolescent Relationships G eneral P arenting Style. Prior research provides am ple support for the ben­ efits to children o f authoritative parenting (i.e., high levels o f w arm th and respon­ siveness, effective m onitoring and control o f children’s behavior, enforcing and following through on rules, low levels o f coercion and conflict). A com m on expla­ nation for why children show elevated problem s in rem arried families is that such parenting-related processes become disrupted, particularly during the period of adjustm ent to remarriage. M any studies, in fact, find tem porary increases in con­ flict and negativity during the first year o f adjustm ent to rem arriage (e.g., Bray, 1988; Hetherington, 1988,1989). We have reported elsewhere (H etherington et al., 1992) that the m ost difficult interpersonal challenge for rem arried families is the

13.

PARENTAL R EM ARRI AG E, A D O L E S C E N T A D J U S T M E N T

303

developm ent of constructive relationships between parents an d their children. Over tim e, custodial m others in rem arried stepfather families seemed to negotiate this task effectively w ith their adolescent children. Initial high conflict and low m onitoring and control gave way to levels sim ilar to those in nondivorced families by 26 m onths after remarriage. Relationships between stepfathers and their nonbiological stepchildren follow an opposite pattern, with difficulty increasing over time. Early after the remarriage, stepfathers themselves felt a lack of affection and rapport with their stepchildren but were acting toward their stepchildren in ways that appeared overfly positive or friendly, attem pting to engage them in discussion and interaction— a pattern we have described previously as “polite strangers” (H etherington et al., 1992). The re­ sponses of adolescents to these initiations, however, were frequently negative. As a consequence, stepfathers showed a relatively disengaged style o f parenting, report­ ing little control or awareness of their stepchildren’s activities and attem pting to exert little discipline. Previous studies o f rem arriage with younger children (H etherington et al., 1985) have found that positive, supportive relationships with stepfathers are re­ lated to decreases in externalizing for boys. In prior analyses o f the LSR, when the relation between parenting style and adolescent behavior was exam ined in step­ families, a clear, positive effect o f authoritative parenting was found relative to other parenting styles, a finding that held across both custodial m others and resi­ dential stepfathers (H etherington et al., 1992). Figure 13.4 shows this relation. Ado­ lescents in both family types benefited from the presence o f an authoritative par­ ent. As was found in the analyses for life stress, however, adolescents in stepfamilies reared by authoritative parents still showed higher levels o f externalizing than those in nondivorced families who were reared by authoritative parents. In other words, offspring in stepfamilies under favorable parenting conditions fared about as well as those in nondivorced families under unfavorable parenting conditions. Why do these differences in externalizing remain even after accounting for p ar­ enting style? It may be that particular aspects o f parenting, such as coercive disci­ pline or ineffective m onitoring, are m ore salient in affecting adolescent behavior in stepfamilies than in nondivorced families. For example, the relative reluctance of stepfathers to engage in discipline m ay provide adolescents with opportunities to exploit any inconsistency in discipline between residential parents. This situation m ay provide the adolescent m ore frequent opportunities to engage in antisocial activities or fewer consequences for doing so. Thus, the next sections address pa­ rental contributions in m ore detail by exam ining first the role o f m onitoring and control and then the role o f conflict, coercion, and punitiveness. Although a u th o r­ itative parenting combines several broad dim ensions (i.e., effective m onitoring and control, low coercion, conflict, and punitiveness, together w ith involved and n u rtu ran t parenting), we have found in other analyses (Anderson, Lindner, & Bennion, 1992; H etherington et al., 1985) that indices o f parental w arm th and involve­ m ent relate m ost strongly to positive outcom es o f social and scholastic compe-

A N D E R S O W f.T AL.

Moilier'3 parenting style

FKX I i ^ W*:bt ion beiwwn n lt rn j liiin n Iw bfinor and paroiti nri oanflfa, cocrcion. u d nrfl* liviiy n ith Kdatimlvunenl m l h !uw tfsilKil and ntunil:'iri:l|*. Da '.j frnirt Aruli'jMiii rf *2.11W2 • tried wilh prtvTiiv^icm

lencc. Instead, this chapter focuses on adolescent cxt«naJizins problems, so the two dimensions of parenting most closely related to externalizing iri adok-sconcc, taw parental monitoring and inept, coerdw, conElictual discipline (toebeT & StOUllumer Loeber. ]9*Wi; PjlttTsnri. I9H2), art1foeused (in. jVft>niUH-ing cjn ti C o n t r o l . When a stepfather is tntroduced into the tamily, he may have limited knowledge and awareness aboiil the i:Uj|H'*r liM; .j|id activities, a:, compared to the custodial mo?hit. The child,, then, is likely to be exposed to incon­ sistent parental monitoring and control attempts between the residential mother and Ihe stepfalher.as well ax between Ihe resident! jl and non residential parents. In tact the adolescent may be keenly serisi live tn whit each pa rent knows and eip eili. In the JSR, adolescents reported or* itiorthof Lng and control reccived from as many as three parents (i.e., custodial mothers, (step] fathers, twnresidential fathers), making possible a direct companion of percrived consistency between different authorities in ait adulesLcnt’s life. 'lb assess parental monitoring and control, adolescents used * 5-point scale To rate the extent o f each parent s awareness and control over behavior in seven areas, iu d i as choice of friends, school imeresls, and adivilKs away from home, Cton-

13.

P A R E N T A L KR>J A R R T A C E , A D O L E S C E N T A D J U S T M E N T

305

bath alpha ootffisclejut for a stale created from avenging these seven areas were above .79 for each parent at all three waves. Mean level diflcreiKes in parental control and m o n ito rin g are presented In Fig. [3.5 It is n o t surprising that adolescents view m others « b e tte r m o n ito rs lliao fathers. T his difference is sm allest in nondivorced families a n d largest when cnmp a n n g custodial niother* and n o n residential biological fathers. Parental difierenccs i r e m o re m arked for m o n ito rin g than fox c o n tro l W ithin stepfam ilies, we can com pare m ean level drlierences across th ree differ­ ent parents, as in l'ig, 13.6. Ajp in, m others are rated highest, ncinresidential fathers are rated lowest, and these parental differences are m ore m arked for m o n ito rin g than lo r control, Parental differrnccs are m o re apparen t for girls than hoys, p rim a ­ rily because o f llie grcaicT degreec o f involvem ent o f nonresidential fathers in the lives o f hoys. Level o f control is equivalent across parents for boys, whereas jjirls rate nonrcsidcntial fathers as le s sslrid than residential parents. T he im p o rta n t question becom es w hether variation in parental consistency re ­ lates to adolescent adjustm ent. Here, we focus nn parental m o n ilo rin p because ihe decree o f parental differences was greater in this dim ension. Parents were classified as either good m onitors (i.e., usually o r always know ing ab o u t th e ir adolescent's behavior) o r p*x>r m o n ito rs (i.e., som etim es. rarely, o r never know ing about (heir adolescent's behavior), based o n the average score-of ihe seven m oiiiiurUi); areas. M others w ho w ere classiticd as good m o n ito rs a t 4 m onths into Ihe rem arriage )uh3 adolescents w h o were time![Sternly low er in exierikalixing throu|^H>ut the 2 years o f ih e study. Im portantly, th ere were n o interactions by family type for m o th ­ ers, b u t there was an interaction n f falher s m onrtoring with family lype, show n in Fig. 13.?. Although adolescents in stepJam ilies evidenced m o re externalising than adolescents in n o jid iw rc ed families, those w hose stepfathers slatted oul as po o r

Mean levels af nvanitonng and corwgl

PIG. J3.S.

Mean d rfftrm c ti between pitE nta in m o n ito rin g andcontnot in different lim ily iir u o u r t s .

Paranfrf diffarunefrS in r«rnarr*d fu n lies

►T (he effect o f remarriage disappeared from (he cross-sectional and Umgitudinal analyses. This finding indicates that adolescents in eithrr contort involring di­ vorce differ from adolescents in notidiviaTced households, but (hose traversing (he pathway nf a prior parental divorce (i.e.. remarried and remained-iingle families) do not differ. Implied in these analyses is that all initial differences between ado­ lescent in stepfamflics and nondivorced families arc attributed to the experience of parental divorce or some «1 of factors associated with the divorce process, Some o f this effed was captured by the variables o f marital instability (i.e., number o f separations} and current arguments, but other aspects o f the divorce must be considered.

A N D E R S O N ET A L .

316

TA BLE 13.6 Effect of Parental Divorce and Remarriage on Initial Level and Change in Adolescents' Externalizing Regression Coefficients Initial Level Predictors Parental divorce Parental remarriage Initial level o f externalizing R1 if F

.34'** -.01 — .11 2,199 12.2*”

Change

.09 .10 -.47* .21 3,160 13.9*

Note. ***p< .001.

It can be speculated th at the experience o f parental divorce and its concom itant disru p tio n s in family process (e.g., living w ith dim inished econom ic resources; experiencing changes in residence, neighborhood, and peer groups; exposure to m arital conflict; involvem ent o f the child in m arital disputes; escalating cycles o f coercion betw een parents and children) changes children in fu ndam ental ways that m ake the transition to parental rem arriage m ore difficult. W itnessing parents dem ean and criticize one an o th e r in an environm ent o f hostility m ay dim inish the capacity o f parents in two ways, as effective role m odels for a n d as socializers of their children. F urtherm ore, in a clim ate w here a p arent uses the child to gain a d ­ vantage over the oth er spouse, a child m ay learn readily the skills necessary to be an effective m anipulator o f his o r her parents. It should be kept in m in d th a t the chil­ dren in the LSR were in early adolescence at the tim e o f the rem arriage and thus were socially advanced enough to take advantage o f the lessons unw ittingly offered by their parents. O th er studies w ith younger children find m ore positive effects o f stepfathers on boys’adju stm en t (H etherington et al., 1985). T hus, for early adoles­ cents, the experience o f divorce m ay provide an arena to hone skills in aversive control. Later, w hen en countering parental rem arriage, these skills serve to shape the behavior o f the custodial and residential (step)parent, and thus, the rem arried family system m ay be driven m ore strongly by characteristics o f the adolescent rath er th an the parent. T here is a great deal o f su p p o rt for the notion that adolescents becom e involved in fam ily conflict u n d e r conditions o f m arital distress (see Davies & C um m ings, 1994, for a review). A recent m icroanalytic study o f family negativity finds that boys w ith m aritally distressed parents— w hether form ed from first or second m a r­ riages— were draw n into family conflict (G reene & A nderson, 1998). W hen be­ havioral contingencies were exam ined (i.e., the likelihood th at negativity by one person w ould be followed by negativity from a nother), this group was at a clear disadvantage. Elder sons tended to escalate the negativity they received from their

13.

PARENTAL REM A R R IAG E, ADO LE SCE NT A D JU ST M E N T

317

parents, and (step)fathers escalated negativity from their sons. These contingent ways o f responding to negativity were not found in families w ith girls, nor in fam ­ ilies with boys when m arriages were not distressed. Thus, boys m ay be particularly vulnerable to effects o f m arital conflict. We speculate that the processes that put children m ost at risk in rem arried fam ­ ilies are likely form ed during the m arital dissolution. Active involvement in esca­ lating family conflict may provide extended trials for children to learn aversive m echanism s for dealing with conflict. T he echoes o f these trials appear later in re­ m arried families when we see the continuing role o f parent-adolescent, sibling, and m arital negativity in m aintaining conduct problem s. Especially under times o f stress, overlearned use o f negativity may m aintain and exacerbate conduct pro b ­ lems in adolescents. If m arital dissolutions and its associated processes are in fact largely responsible for externalizing problem s in rem arried families, we expect that stepfamilies form ed with a previously single nondivorced m other would not be as­ sociated with the sam e pattern o f externalizing as that with divorced mothers. Stress and family factors are clearly im portant for adolescents in rem arried fam i­ lies, but these factors alone cannot explain why adolescents in stepfamilies start the remarriages at a disadvantage. Just as relationships between divorced family m em ­ bers continue long after the legal decree, many o f the processes at w'ork in step­ families begin before the decision to remarry. Potential stepparents evaluate the likelihood o f establishing and m aintaining satisfactory relationships with their stepchildren, and the circumstances under which those relationships begin may determ ine to a large extent whether o r not a rem arriage takes place at all.

SUMMARY By the m ajority o f adolescent accounts, parental rem arriage and the accom pany­ ing transitions o f m oving to a new hom e, starting at a new school, and dealing with new siblings are n ot inherently negative circumstances. Nevertheless, when the av­ erage level o f adolescent externalizing behavior in rem arried families is com pared to that in nondivorced families, significant differences are found, and the experi­ ence of parental rem arriage predicts increasing problem s over time. In this chap ­ ter, a variety of potential explanations for this effect were examined. The frequency o f negative stressful events was significantly related to adolescent problem behavior, in accord with previous research. Moreover, negative relation­ ships within a family, w hether between adolescents and (step)parents, the new m arital partners, or siblings, were associated with adolescents’ conduct problems. In addition, qualities o f the relationship between form er spouses, such as num ber o f separations prior to divorce and frequency o f current child-related argum ents, predicted the level o f externalizing in the adolescent. None o f these variables, taken together or individually, accounted for the differences in externalizing behavior between nondivorced and rem arried families. Only the experience o f parental di­

A N D E R S O N E T Al..

318

vorce itself appeared to account for these initial differences. Additionally, charac­ teristics o f the divorced couple relationship accounted for the effect o f rem arriage on increasing externalizing problem s over tim e. W hatever initial differences were found betw een adolescents in rem arried and nondivorced families, at least w ith re­ spect to conduct problem s, could be a ttributed to factors associated w ith the expe­ rience o f parental divorce. It is likely th at involvem ent in parental hostility p ro ­ vides adolescents w ith o p p o rtu n ity to engage in coercive interchanges and to learn aversive ways o f handling conflict a n d disagreem ent. These patterns m ay n o t only increase adolescent conduct problem s b u t m ake it m ore difficult for parents to exert effective socialization. O th er analyses suggest that adolescent problem behav­ ior n o t only leads to negativity from parents (H etherington et al., 1992) b u t m ay m ake the transition to rem arriage less likely to occur (M ontgom ery et al., 1992). It should be kept in m ind that a great deal o f variation in adolescent adjustm ent to rem arriage rem ains. M any stepfam ilies were able to establish w arm , involved, supportive relationships across all household m em bers. Adolescents in such situa­ tions fared m uch better than w hen fam ily relationships were characterized by m ore coercive, punitive, and conflictual exchanges. For the typical early adolescent, liv­ ing in a rem arried family appears equivalent to living in a m oderately stressful n o n ­ divorced family. T his evidence suggests lasting influence from the cum ulative ef­ fect o f experiencing m ultiple family transitions (i.e., divorce, life in a single-parent household, cohabitation, and en try into a rem arriage). Future research and theory on adolescent adjustm ent to rem arriage needs to be guided by fu rth er u n d e rstan d ­ ing the critical processes that are set into m o tio n by a dissolving parental m arriage.

ACKNOW LEDGM ENTS This research was supported by a gran t from the John D. an d C atherine T. M acA rth u r F oundation to E. Mavis H etherington and W. G lenn Clingem peel. Sup­ p o rt for the w riting o f this chapter was provided by N IM H 5P30M H39246 and 2T32M H18387.

REFERENCES Aiken, L. S., 8c West, S. G. (1991). Multiple regression: Testing and interpreting interactions. Newbury Park, CA: Sage. Anderson, E. R. (1998). Sibling, half-sibling, and stepsibling relationships in remarried families. M anu­ script under review. Anderson, E. R„ Hetherington, E. M., & Clingempeel, W.G. (1989). Transformations in family relations at puberty: Effects o f family context. Journal o f Early Adolescence, 9, 310-334. Anderson, E. R.> Lindner, M. S., & Bennion, L. D. (1992). The effect of family relationships on adoles­ cent development during family reorganization. Monographs o f the Society for Research in Child De­ velopment, 5 7 (2-3, Serial No. 227).

13.

p a r e n t a l r e m a r r ia g e , a d o le s c e n t a d ju s t m e n t

319

Bray, J. H. (1988). Children’s development during early remarriage. In E. M. Hctherington & J. D. Arasteh (Eds.)» Impact o f divorce, singleparenting and stepparenting on children (pp. 279-298). Hills­ dale, NJ: Lawrence Erlbaum Associates. Davies, P. T., & Cummings, E. M. (1994). Marital conflict and child adjustment: An emotional security hypothesis. Psychological Bulletin, 116,387—411. Deal, J. E., Anderson, E. R., Ratliff, D., Hctherington, E. M., & Clingempcel, W. G. (1990, November). A social relations model analysis offamily behavior in remarried families. Paper presented at the meet­ ings of the National Council on Family Relations, Seattle, WA. Greene, S. M., & Anderson, E. R. (1998). Behavioral contingency in tetradic family interactions. Manu­ script under review. Hetherington, E. M. (1988). Parents, children, and siblings six years after divorce. In R. Hinde 8c J. Stevenson-Hinde (Eds.), Relationships within families (pp. 311-331). Cambridge, England: Cambridge University Press. Hetherington, E. M. (1989). Coping with family transitions: Winners, losers, and survivors. Child De­ velopment, 60, 1-14. Hetherington, E. M.» Clingempeel, W. G., Anderson, E. R., Deal, J. E., Stanley Hagan, M.> Hollier, E. A., 8c Lindner, M. S. (1992). Coping with marital transitions: A family systems perspective. Monographs o f the Society for Research in Child Development, 57 (2-3, Serial No. 227). Hetherington, E. M., Cox, M. J., 8c Cox, R. (1985). Long-term effects of divorce and remarriage on the adjustment o f children. Journal o f the American Academy of Child and Adolescent Psychiatry, 24, 518-530. Kenny, D. A., 8c LaVoie, L. (1984). The social relations model. In L. Berkowitz (Ed.), Advances in experi­ mental social psychology (Vol. 18, pp. 141-182). New York: Academic. Loeber, R., & Stouthamer-Loeber, M. (1998). Development of juvenile aggression and violence: Some common misconceptions and controversies. American Psychologist, 53,242-259. McArdle, J. J., & Anderson, E. R. (1990). Latent variable growth models for research on aging. In J. E. Birren and K. W. Schaie (Eds.), Handbook of the psychology o f aging (3rd ed., pp. 21-44). San Diego: Academic Press. Montgomery, M. J., Anderson, E. R., Hetherington, E. M., 8c Clingempeel, W. G. (1992). Patterns of courtship for remarriage: Implications for child adjustment and parcnt-child relationships. Journal o f Marriage and the Family, 54,686-698. Patterson, G. R. (1982). Coercive family process. Eugene, OR: Castalia. T ze lg o v ,8 c Henik, A. (1991). Suppression situations in psychological research: Definitions, implica­ tions, and applications. Psychological Bulletin, 109,524-536.

This page intentionally left blank

PART

V

intervention

This page intentionally left blank

14

Psychological Interventions for Separated and Divorced Families R o b e rt E. E m e ry University o f Virginia K a th e rin e M . K itz m a n n University o f Memphis M a ry W ald ro n University o f Virginia

o b s e r v a t i o n s serve as points o f departure for the present overview of interventions with divorcing families. First, dem ographic data clearly indicate that both children and adults from divorced families are two to three times as likely to receive psychological treatm ent as are m em bers o f m arried families (How ard et a l, 1996; Zill, M orrison, & Coiro, 1993). Divorced family m em bers are overrepre­ sented in the treatm ent population in part because psychological disorders are m ore com m on am ong children and parents from divorced families. At the same tim e, resilience is the norm ative psychological outcom e following divorce, and the proportion o f children and parents who receive treatm ent exceeds the proportion with dem onstrable psychological disorders (for a review, see Emery, 1998). Many m em bers o f divorced families likely seek treatm ent, not for a psychological dis­ order but for help in dealing with the subclinical distress caused by the family disruption (Laumann-Billings 8c Emery, 1998). Treatm ent also may be initiated for other reasons, such as a custody dispute. Regardless o f how one interprets the data, however, epidemiogical evidence underscores the im portance o f research on psy­ chological intervention in divorce from the perspective of service utilization if for no other reason. The second point concerns the wide range o f potential targets and tim ings of intervention in divorce. Divorce involves social, legal, and econom ic consequences as well as psychological ones, and various societal controls have been im plem ented

T hree

323

324

EMERY, K I T Z M A N N , W A LD R ON

and co ntinue to be considered as m eans o f regulating divorce (Eekelaar, 1991). M oreover, longitudinal research has clearly dem o n strated that divorce is best con­ ceptualized as a process o f adaptation and change over a period o f several years (H etherington, 1989, 1991, 1993), and interventions m ay take place at various p oints in tim e before the divorce (e.g., laws designed to p ro m o te m arriage), d uring the divorce (e.g., m ediation or family therapy), o r long afterw ards (e.g., individual or group psychotherapy). All these interventions are o f potential interest not only for their in tended im pact b u t also because o f their potential indirect psychological consequences in preventing, treating, o r possibly exacerbating psychological p ro b ­ lem s am ong family m em bers. T he th ird observation concerns the trem endous disparity betw een how fre­ quently intervention in divorce is discussed in the applied literature (in psychol­ ogy, law, and o th er fields) and how infrequently divorce interventions are su b ­ jected to m ethodologically sophisticated studies, particularly random ized trials. For exam ple, a Psychlit search conducted for this chapter revealed over 500 articles and chapters published w ith the keywords divorce and therapy o r treatment. As reviewed here and elsewhere (Emery, 1998; Grych 8c Fincham , 1992; Lee, Picard, 8c Blain, 1994), few o f these published accounts are em pirical reports; only a handful are random ized trials. T he present review focuses on research on divorce intervention w ith an em pha­ sis on the psychological consequences for children’s w ell-being, including both their psychological disorders and their subclinical distress. The review also includes interventions designed to affect family relationships o r individual parents’ m ental health because these factors are critically im p o rta n t to children’s m ental health fol­ low ing divorce. T he review pays special attention to those few studies that have used experim ental o r quasi-experim ental designs, although selected em pirical studies or clinical reports are also included when they suggest creative directions for future research on assessm ent o r intervention. Finally, the review is divided into three m ajor sections based on the target o f intervention: the individual child, parents a nd parenting, and the co-parenting relationship. It should be noted at the outset that system atic research on treatm ents for individual children is lim ited exclusively to school-based interventions. C ontrolled trials o f therapy for postdivorce p a ren t­ ing are lim ited to group therapy, and m ethodologically adequate research on co­ p arenting interventions is lim ited to divorce m ediation. O th e r interventions are frequently used for individual children (e.g., play therapy), parents and parenting (e.g., individual psychotherapy), and coparenting (e.g., family therapy), b u t we discuss these treatm ents only briefly here due to the absence o f research.

C HILD-FO CUSED TREATMENTS: SCHO OL-BASED DIVORCE IN TERVEN TIO N S As noted, children from divorced families are overrepresented in the m ental health treatm ent po p u latio n , b u t there are no system atic data on w hat types o f treatm ent

14.

PSYCHOLOGICAL INTERVENTIONS

325

these children typically receive. M any children likely are in individual psychother­ apy, including b u t n o t lim ited to play therapy (see Hodges, 1991, for a clinical overview), and others m ay be involved in group therapy in the com m unity. In a d ­ d ition, a n u m b e r o f children from divorced families probably are treated in fam ily therapy, w hich is discussed in the next section on p arenting interventions. It is su r­ prising, however, there have been no system atic attem p ts to evaluate the effective­ ness o f individual psychotherapy w ith children from divorced families, a treatm ent th at is likely to be recom m ended w hen lack o f co m m itm en t o r intractable p ro b ­ lem s prevent family therapy (H odges, 1991). Rather, research on child-focused intervention in divorce is lim ited to group therapy and m ore specifically, schoolbased treatm ent groups. G roup therapy is the m ost com m on form at for various school-based divorce interventions, as m ost schools d o n o t have adequate resources to address divorcerelated difficulties in a child-specific m an n e r (G rych & Fincham , 1992; Kalter 8c Schreier, 1993; Pedro-C arroll, 1993). The prim ary benefit o f a group form at is co n ­ tact w ith o th er children w ho have had sim ilar experiences (Pedro-C arroll, 1997; Stolberg & Cullen, 1983). Supportive peer groups m ay help to lessen feelings o f isolation a n d loneliness, foster feelings o f su p p o rt and trust, and offer o p p o rtu n i­ ties to clarify divorce-related m isconceptions, such as feelings o f responsibility or blam e for their parents’ break-up (Pedro-C arroll, 1997; Stolberg & Cullen, 1983). In contrast to individual therapy w ith an adult clinician, groups m ade up o f peers m ay also offer a m ore open, natural, and com fortable environm ent in which to dis­ close sensitive experiences (G rych & Fincham , 1992; Pedro-C arroll, 1997; Stolberg & C ullen, 1983). A lthough techniques and m aterials vary across specific program s, school-based divorce interventions share m any goals a n d strategies (G rych 8c Fincham , 1992). M ost incorporate both educational and therapeutic activities in the atte m p t to fos­ ter feelings o f support, clarify divorce-related m isconceptions, and develop effec­ tive coping skills to deal w ith the challenges and stressors associated w ith parental divorce. In addition, school-based divorce interventions are typically tim e-lim ited, ranging from 6-week sessions (B ornstein, B ornstein, & Walters, 1988) to 16-week sessions (Alpert-Gillis, Pedro-C arroll, 8c Cow en, 1989). The shorter d u ra tio n of the interventions is advantageous given lim ited resources as well as concerns ex­ pressed by som e parents th at children n o t m iss o u t o n o th er school activities for extended periods o f tim e (Kalter 8c Schreier, 1993). There are m any advantages o f school-based interventions for children o f d i­ vorce. T he n u m b er o f children served in school-based groups is m uch larger than is possible w ithin individual therapy; indeed, m any children o f lower incom e could n o t afford such services on an individual basis (G rych 8c Fincham , 1992; Pedro-C arroll, 1997). Schools are also a natu ral context for providing support given the n u m b er o f bo th peers and teachers present (Cowen, H ightower, PedroCarroll, 8c W ork, 1989; Grych 8c Fincham , 1992). In addition, children o f divorce are m ore likely than children from m arried families to experience behavior p ro b ­ lem s in the classroom (Emery, 1982; G uidubaldi, Perry, 8c Clem inshaw, 1984) re­

326

EMKRY, K I T Z M A N N , W A L D R O N

suiting in difficulties in concentration and com prom ised academ ic perform ance (Kalter & Schreier, 1993). C onsistent w ith the educational goals o f the school (Hodges, 1991), school-based program s m ay help to reduce b o th behavioral and academ ic problem s.

Evaluation of School-Based Interventions A lthough school-based intervention program s for children o f divorce are fre­ quently im plem ented, em pirical evaluation o f such program s is lim ited (see Grych & Fincham , 1992). T he two exceptions include Pedro-C arroll and colleagues’ C hil­ dren o f Divorce Intervention Project and Stolberg and colleagues’ Divorce Ad­ ju stm e n t Project, b o th o f w hich have been subjected to one or m ore replication studies. In the follow ing sections, we review research on these two program s as well as several sm aller scale studies evaluating various school-based divorce program s. C h ild ren o f D iv o rc e In te r v e n tio n P roject ( C O D IP ). C O D IP is a schoolbased, child-focused, supportive gro u p intervention designed to be a preventive intervention. Five specific objectives include: providing a supportive group envi­ ronm ent, helping children identify and express appropriate feelings, clarifying divorce-related m isconceptions, enhancing coping skills, and enhancing children’s perceptions o f self and family (Cow en et al., 1996). Various C O D IP program s ta r­ geting children o f different ages and backgrounds have been evaluated since the program was first im plem ented in the early 1980s. In the initial evaluation (Pedro-C arroll & Cowen, 1985), 75 fourth- through sixth-grade children from four su b u rb an schools were random ly assigned to ei­ ther an im m ediate 12-week intervention o r a delayed intervention control group m atched for sex, grade, and length o f tim e since their parents’ separation. Accord­ ing to teacher report, children in the intervention group exhibited significantly greater reductions in shy-anxious behavior, learning problem s, and overall schoolrelated problem s, and they im proved significantly m ore on m easures o f total com petence and m ore specific com petencies including peer sociability, frustration tol­ erance, rule following, and adaptive assertiveness as com pared to children in the control group. Similarly, the reports o f parents and group leaders indicated signifi­ cant decreases in feelings o f self-blam e and increases in personal problem -solving abilities. C hildren in the intervention reported feeling less anxious and having less negative self-attitudes and perceptions o f divorce than d id controls; however, no differences were found betw een children in the intervention and control groups on perceived self-com petence. A replication o f this initial study w ith different group leaders and different schools, b u t w ithout random assignm ent to conditions, re­ vealed sim ilar findings {Pedro-Carroll, Cowen, Hightower, 8c G uare, 1986). C O D IP was subsequently adapted for u rb a n children w ith diverse ages, e th ­ nicities, an d socioeconom ic backgrounds (Pedro-C arroll & Cow en, 1985; PedroCarroll et al., 1986). For exam ple, in the revised program the concept o f “d i­

14.

PSYCHOLOGICAL INTERVENTIONS

327

vorce” was extended to include n o t only legal m arriages b u t also long-term live-in partnerships. In an evaluation o f a 16-session ad aptation, intervention and com ­ parison groups o f second- an d third-grade urban children were recruited sepa­ rately through school announcem ents, resulting in 52 p rogram participants, 52 com parison children from divorced families, and 81 dem ographically m atched com parison children from intact families (Alpert-Gillis et al., 1989). Pre- to post­ intervention statistics indicated that C O D IP children reported greater im prove­ m en ts on m easures assessing coping abilities in problem situations and feelings a b o u t self and family than did children from either divorce o r in ta rt com parison groups. Parents reported m ore gains in children’s ability to deal w ith feelings, be­ have appropriately, and solve problem s. In addition, teachers reported m ore im ­ provem ent for C O D IP children in school com petencies including peer social skills, assertiveness, and frustration tolerance b u t n o t school-related problem behaviors. An evaluation o f C O D IP for fo u rth - th ro u g h sixth-grade u rb a n children revealed sim ilar results (Pedro-C arroll, Alpert-Gillis, & Cowen, 1992). A dditional evidence o f C O D IP ’s effectiveness across ages can be found in evaluations o f seventh- and eighth-grade children (Pedro-C arroll, Sutton, 8c Black, 1993, cited in Cowen et al., 1996). Finally, in a recent study o f 105 kindergarten and first-grade children (PedroC arroll 8c Alpert-Gillis, 1997), teacher ratings indicated that C O D IP children dis­ played significantly m ore gains in school-related com petencies, as well as exhibited less anxious, w ithdraw n, and disruptive behavior than dem ographically m atched control children from divorced families. In contrast to children in the com parison group, children in the program reported feeling less w orried a b o u t changes that had taken place in their family and m ore positive ab o u t them selves, th eir families, and their experience in the group. T hey also reported talking m ore w ith their p a r­ ents and in general, said they enjoyed talking ab o u t their feelings. Parents reported im proved p a ren t-ch ild com m unication; increased ability to share feelings, display a ppropriate behavior, and cope w ith problem s; and decreased m oody and anxious behavior (Pedro-C arroll, Sutton, 8c W ym an, 1996, cited in Pedro-C arroll 8c AlpertGillis, 1997). D ivorce A d ju s tm e n t P roject (D A P ). DAP is a stru ctu red prim ary interven­ tion program consisting o f two com ponents; a tw o-part school-based C hildren’s S upport G roup (CSG) and a com m unity-based Single Parents’ S upport G roup (SPSG; Stolberg 8c G arrison, 1985). Because the present focus is treatm ents for children, SPSG is discussed in a later section on parenting. CSG is a 12- to 14-week psychoeducational su p p o rt group designed to provide em otional su p p o rt for 7th rough 13-year-old children o f divorce and to teach cognitive behavioral skills including problem solving, anger control, im pulse control, com m unication, and relaxation. In the first o f the DAP evaluations (Stolberg 8c G arrison, 1985), 82 children aged 7 through 13 years old and their custodial m others were assigned to either a

328

EMERY, K T T Z M A N N , W A L D R O N

n o -trea tm e n t control group o r to one o f three 12-week treatm ent groups, includ­ ing CSG alone, SPSG concurrent w ith CSG, o r SPSG alone. In com parison to both the control and com bined intervention groups, children in the CSG alone group im proved in self-concept as m easured im m ediately after the p rogram and a t a 5m o n th follow -up. At follow -up, children in the CSG alone group also exhibited greater increases in adaptive social skills. Surprisingly, the com bined CSG and SPSG intervention group did n o t result in expected a d ju stm en t gains, b u t this finding m ay reflect n o n ran d o m assignm ent and preexisting dem ographic differ­ ences betw een the groups. M others in the com bined intervention were separated longer, h a d lower em ploym ent status, and reported less frequent visitation with fathers, whereas children in the com bined group were b e tte r adjusted p rio r to the intervention. In a m ore recent evaluation, Stolberg and M ahler (1994) supplem ented the ex­ isting DAP by adding adjustm ent m easures tapping clinical ranges, teacher reports o f child adjustm ent, and m ore gam e-like activities to engage the interest o f the families. Participating schools were random ly assigned to a n o -treatm en t divorce control group o r one o f three 14-week treatm ent groups, including su p p o rt only, su p p o rt w ith skill building, o r sup p o rt, skill building, skill transfer (i.e., th erap eu ­ tic h om e w orkbooks for children and parents designed to facilitate transfer o f inclass adjustm ent gains), an d parenting training. T h ird - th ro u g h fifth-grade chil­ dren o f separated and divorced parents were recruited, resulting in a sam ple o f 103 children from divorced families. In contrast to the controls w ith divorced parents, children in b o th skill-building groups displayed significant im provem ent im m ed i­ ately follow ing intervention and at a one-year follow -up. G reater reductions in both internalizing and externalizing behavior, total pathology, and clinical sym p­ tom atology were observed for children in the skills and su p p o rt group im m edi­ ately follow ing intervention. C hildren in the skill-building condition w ith skill transfer and parenting training exhibited less trait anxiety p ost-intervention; how ­ ever, behavioral gains were not observed until follow -up. C hildren in the su pportonly condition exhibited greater reductions in clinical sym ptom atology at followup, w ith the im pact greatest for those children entering the gro u p w ith significant preexisting clinical problem s. A d d itio n a l E va lu a tio n S tu d ies. In addition to these tw o program s, several sm aller scale group interventions with children have been studied empirically. In one o f the earliest reports, A nderson, Kinney, and G erler (1984) evaluated a school-based group intervention targeting children’s attitu d es tow ard divorce, classroom behavior, and academ ic perform ance in m athem atics and language arts. T hird to 6th grade children were recruited through school announcem ents: 52 were random ly assigned to one o f seven 8-week divorce groups, and 32 children served as a control group. G roups included role playing new behaviors, discussing feelings a b o u t divorce, pu p p et activities to aid in u n derstan ding and acceptance of divorce, and learning new relationships skills. P re -p o st com parisons indicated ad­

14.

PSYCHOLOGICAL INTERVENTIONS

329

ju stm e n t gains for those children in the treatm ent group significantly exceeded those o f control children on m easures o f attitudes tow ard divorce and classroom conduct. No significant im provem ent in language arts or m athem atics was found for either the treatm ent o r the control group. Roseby and D eutsch (1985) evaluated a school-based program to determ ine w hether skill train in g in social role-taking and assertive co m m unication result in im proved divorce adjustm ent. In this study, 57 children 9 th rough 11 years old p a r­ ticipated in one o f two 10-week intervention groups: experim ental or placebo con­ trol. The experim ental condition provided training in cognitive social role taking and assertive com m unication skills, whereas the placebo control group provided n o skill training, focusing instead on identification and discussion o f divorcerelated feelings. Following the intervention, children in the experim ental group w ere less likely to blam e th eir m others, fathers, o r them selves for the divorce, to fear peer ridicule or parental aban d o n m en t, to hold unrealistic hopes for reconcil­ iation, or to deny feelings o f pain as com pared to those children in the placebo condition. In contrast, no differences betw een children in the experim ental and placebo conditions w ere observed on m easures o f classroom behavior, specifically aggressive behavior, o r on m easures o f depression follow ing the intervention; that is, b o th the experim ental and placebo groups exhibited im proved classroom be­ havior and decreases in depression. Gw ynn and B rantley (1987) investigated the effectiveness o f an 8-week educa­ tional su p p o rt group designed to educate children o f divorce about divorce-related concepts and issues, encourage expression o f feelings, a n d prom ote adaptive prob lem -solving skills. P articipants included 30 children aged 9 th ro u g h 11 years whose parents had been separated for at least 1 year. T hirty children m atched for gender and tim e since parental separation served as a control group. P re -p o st co m p ar­ isons indicated adjustm ent gains for those children in th e trea tm e n t gro u p sig­ nificantly exceeded those o f control children on m easures o f depression, anxiety, negative feelings ab o u t divorce, and know ledge about divorce. O m izo and O m izo (1987) evaluated the efficacy o f a 12-week school-based group on children’s self-concept and locus o f control. Sixty fo u rth - through sixthgrade children were random ly assigned to either treatm en t o r control groups. T reatm ent activities involved group discussion of divorce-related feelings, role playing, and relaxation and assertiveness training. Post intervention, children in the treatm en t group scored higher on m easures o f locus o f control and two o f four indices o f self-concept. B ornstein, Bornstein, and W alters (1988) exam ined a 6 -week school-based su p ­ p o rt group for 7- through 14-year-old children o f divorce. T hirty-one children were random ly assigned to either an experim ental group, em phasizing identifica­ tio n and clarification of divorce-related feelings, anger m anagem ent, training in com m unication skills, problem solving, and group su p p o rt, o r a delayed treatm ent control group. A positive effect was found for only one o f seven outcom es m ea­ sured. C hildren in the experim ental group exhibited greater decreases in school-

330

EM ER Y , K IT Z M A N N , W A L D R O N

related behavior problem s m easured at follow -up, b u t no p re -p o st differences were found in children’s attitudes about divorce, p a ren t-ch ild conflict, and a d d i­ tional m easures o f child adjustm ent, including child anxiety, self-esteem , or oth er behavior problem s. It is also im p o rta n t to n ote that this group was shorter than m any o th er such groups, and the assessm ent tools focused solely on adjustm ent and n o t on children’s com m unication o r problem -solving skills. Kalter, Schaefer, Lesowitz, Alpern, and Pickar (1988) found no differences in internalizing and externalizing behaviors for girls w ho participated in a group de­ signed to norm alize the divorce experience, clarify divorce-related issues, provide a safe place to express and discuss feelings, develop coping strategies, and teach p a r­ ents about children’s concerns. However, parents reported significant reductions in boys’ externalizing and overall aggression com pared to controls. Furtherm ore, b o th girls and boys reported m ore positive perceptions o f divorce 6 m onths fol­ lowing program com pletion. Burke and Van de Streek (1989) evaluated an 8-week school-based divorce in ­ tervention targeting children’s self-concept. In this study, 39 children in fourth through sixth grades were random ly assigned to either trea tm e n t or delayed trea t­ m ent control groups. T reatm ent exercises were designed to facilitate group discus­ sion and included films, stories, and gam es w ith divorce-related them es. P re-post com parisons in self-concept indicated significant gains for only those children in the experim ental treatm ent group. N o outcom e o th er th an self-concept was reported. C rosbie-B urnett and N ew com er (1990) evaluated the effectiveness o f a schoolbased group intervention on children’s beliefs ab o u t divorce, depression, an d self­ esteem . Eleven sixth-grade students were random ly assigned to one o f two con­ ditions, experim ental or w ait-list control groups. Following intervention, the experim ental group exhibited greater im provem ent in post-divorce adjustm ent. C hildren in the experim ental group reported less problem atic beliefs about their parents’ divorce, increased scholastic com petence, and lower scores for depression as com pared to controls.

Summary and Evaluation T he findings reviewed are generally consistent and su p p o rt the utility o f schoolbased divorce interventions, b u t there nevertheless is m uch ro o m for im prove­ m ent (see also Grych 8c Fincham , 1992; Lee et al., 1994). M ost studies are w rought w ith m ethodological pro b lem s— few em ployed ran d o m assignm ent; raters o f chil­ dren’s adju stm en t were aware o f the treatm ent, and a placebo treatm en t was used in only one study; and assessm ents were com pleted soon after the intervention, an obvious lim itation given the preventive rationale for the treatm ents. Exceptions include a 2-year follow -up conducted by Pedro-C arroll, Sutton, and Black (1993) and a 1-year follow -up o f the DAP (Stolberg 8c Mahler, 1994). In addition, few a t­ tem pts have been m ade to isolate the active com ponents o f change, and potential

14.

P S Y C H O L O G IC A L IN T E R V E N T IO N S

331

m oderating variables such as age, gender, ethnicity, socioeconom ic status, tim e since divorce, rem arriage, and child tem peram ent have rarely been considered. In addition to these m ethodological issues, an im p o rta n t substantive concern is the n ature and the m agnitude o f change p roduced by school-based groups. In a review o f child and parent-focused interventions, Lee, Picard, and Blain (1994) n oted that child interventions, including m ost o f the studies reviewed here, gener­ ally produced m odest results as m easured by an average effect size o f .27 standard deviation units. T his effect is considerably sm aller th an the com m only reported effect sizes for psychotherapy in general (e.g., Sm ith & Glass, 1977). It appears, m oreover, th at child-focused interventions m ay be m ore successful in alleviating children’s distress o r altering their beliefs ab o u t divorce than in elim inating d is­ turbances such as co n d u ct disorders in school. In short, school-based groups appear to be a helpful b u t lim ited intervention for children coping w ith parental divorce.

PARENT AND PARENTING IN TERVEN TIO N S A lthough parent-focused program s are less com m on th an are child- and schoolbased program s, they appear to be m ore effective in helping parents adjust to d i­ vorce. Lee et al. (1994) fo u n d that evaluations o f p arenting program s yield an aver­ age effect size o f .80, com parable to the average effect reported for psychotherapy in general. However, the greater effectiveness appears to be lim ited to parents. As sum m arized in the next section, im proved parent adjustm ent typically is n o t re­ lated to im proved child adjustm ent in these intervention studies.

Goals, Format, and Content Divorced adults are overrepresented in individual psychotherapy patien t po p u la­ tions, but m ost parent-focused interventions designed specifically for divorce and all research evaluations o f parent-focused divorce interventions involve group therapy. T hese groups have three im plicit goals: im proved psychological adjust­ m ent o f the parent, im proved parenting, and im proved family relationships. In practice, m ost place a p rim a ry em phasis on im proving adult adjustm ent, w ith less em phasis on parenting skills an d family relationship issues. M ore specific topics m ay include coping w ith the lack o f a su p p o rt system, having to ru n a household independently, task overload, em otional overload, and perhaps o th er issues in ­ cluding co m m unication skills, financial planning, social su p p o rt, dating, feelings o f isolation and rejection, parenting and coparenting, self-esteem, depression, relaxation skills, interpersonal skills, legal issues, and career planning. M em bers typically are custodial parents so m any groups have a single-parent focus. P ro ­ gram s vary in the n u m b er o f contact hours p articipants have with the treatm ent and typically m eet in schools (often in coordination w ith child-focused groups),

332

EM ER Y , K IT Z M A N N , W A L D R O N

co m m u n ity m ental health centers, o r churches o r synagogues (C antor & Drake, 1983; Lee e ta l., 1994).

Review of Empirical Studies A handful o f studies o f parent-based program s have been conducted using a com ­ parison o r control group. Four o f the m ore sophisticated studies include the C ol­ orado Separation and Divorce Project, the Divorce A djustm ent Project, the C hil­ dren o f Divorce Parenting Intervention, and the Parenting T h ro u g h Change p ro ­ gram . A few sm aller studies also used com parison groups and are included in the follow ing review. C olorado S e p a ra tio n a n d D ivorce P roject. T his 6-m o n th intervention p ro ­ gram addressed newly separated (but n o t divorced) adults, 43% o f w hom were parents. By ran d o m assignm ent, 100 adults participated in treatm ent, and 50 were p art o f a n o -trea tm e n t control group. T he program provided general su p p o rt as well as specific skills-related training in o rd e r to reduce o r control stress in several key areas o f relevance to the newly separated. Paraprofessionals contacted partici­ pan ts on a regular basis to prom ote social interaction and to m ake referrals to other parts o f the program . O ptional study groups were also available; participants could m eet w ith an expert w ho provided info rm atio n and guidance related to one o f the stress areas targeted by the program . O f the 100 participants, 59 chose to m ake use o f a study group. T he groups m ost utilized by the p articipants were groups related to socialization, childrearing a n d single parenting, and financial planning (Bloom , Hodges, & Caldwell, 1982; Bloom, Hodges, Kern & M cFaddin, 1985; Hodges 8c Bloom, 1986). C om parisons were m ade bo th o f p re -p o st scores w ithin each group and o f d if­ ferences betw een the intervention and control groups (using pretest scores as covariates). C om pared to initial scores, scores at 6 m o n th s indicated th at all partic­ ipants (experim ental and control) had m ore legal an d financial stresses. However, the intervention group show ed a significant d ro p in overall distress and m aladjust­ m ent, in psychological distress, and in anxiety, whereas the control group show ed a significant increase in psychological sym ptom s. C om parisons o f the two groups using pretests as covariates show ed th at the control group also had significantly m ore problem s w ith housing th an did the intervention group. Interestingly, im provem ent in the intervention group was n o t related to extent o f use o f specific study groups. In fact, p articipants w ho m ade use o f the study groups were those w ith m ore problem s initially, and these p articipants continued to have m ore problem s at the end o f the program . Bloom and colleagues con­ cluded that the im m ediate benefits o f the intervention program were general ra th e r than specific and that know ing th at services were available was enough to prom ote better functioning in the intervention group 6 m o n th s later (Bloom et al., 1982).

14.

PSYCHOLOGICAL IN T E R V E N T IO N S

333

The sample was reevaluated at 18, 30, and 48 m onths. Over time, virtually all participants (intervention and control) showed improved functioning, but the in­ tervention group m aintained its better adjustm ent. At 30 m onths, the benefits o f intervention were m ore apparent than they were at 6 or 18 m onths. By 4 years, m ore than half o f the previously dem onstrated significant effects were no longer apparent, but the intervention group showed new im provem ents not seen in the control group, related to psychological functioning, quality o f life, and job satisfac­ tion (Bloom et al., 1985). T he D ivorce A d ju s tm e n t P roject— Single P arent S u p p o rt G roup (SPSG ) (Stolberg & G arrison, 1985). Participants in the .SPSG were 82 pairs o f single, custodial m others (separated less than 33 m onths) and their 7- to 13-year-old chil­ dren. Each m other and child was assigned (although not random ly) to one o f four treatm ent conditions: the Children’s Support G roup (CSG) alone; the CSG and the Single Parents’ Support G roup (SPSG); the SPSG alone; or a nontreatm ent control group. As described earlier, the CSG alone condition was associated with benefits that were not apparent when the CSG m et in parallel with the SPSG (com bined condition). Like the Children’s S upport Group, the SPSG m et for 12 one-hour sessions. Based on the assum ption that improved adjustm ent o f the m others would lead to better parenting, the parent group was designed as a psychoeducational support group and did not attem pt to modify parenting behaviors directly. Some parents participated in the SPSG alone condition, and their children were p art of the con­ trol group; other parents participated in the SPSG in parallel with their children’s participation in the CSG. In the latter case, parent-child pairs assigned to the com ­ bined treatm ent condition were not w orking together in treatm ent. The SPSG did not produce any significant changes in parent or child adjust­ ment relative to the no-treatm ent control group. On the other hand, involvement in the single-parent support group may have prevented deterioration in parent ad­ justm ent that was seen in control-group parents. This was only the case, however, for m others in the SPSG-alone condition. M others involved in the combined SPSG/CSG condition not only did not show any evidence of prevention but showed drastic reductions in adjustm ent. As Stolberg and G arrison noted (1985), this finding could reflect preexisting dem ographic differences that were not con­ trolled for during o r after assignm ent to groups. C hildren o f D ivorce P arenting In terve n tio n (W o lch ik e t al., 1993). Wolchik and colleagues developed an intervention program that targeted five p ar­ enting practices hypothesized to mediate the relation between divorce and child adjustm ent: parent-child relationships, interparental conflict, discipline, contact with and support for nonparental adults, and contact w ith the noncustodial par­ ent. Participants were m others who had divorced w ithin the past 2 years and who had at least one child between 8 and 15. Because the prim ary purpose o f this inter­

334

EMERY, K I T Z M A N N , W A LD R O N

vention program was prevention, the researchers excluded b o th low- and highfunctioning families from their sam ple, hoping to target those families w ho w ould m ost benefit from the program . T he final sam ple consisted o f 70 families. By ra n ­ d o m assignm ent, 36 m o th e r-ch ild pairs were enrolled in treatm ent, and 34 were p u t on a w aiting list, w ith treatm ent to follow' 6 m onths later. As p a rt o f the intervention program , m others participated in 10 weekly sessions w ith a group o f about 6 to 8 m others and in 2 individual sessions (after the third an d sixth groups). G roups were led by a m ale-fem ale team an d strongly e m p h a­ sized skills training related to discipline, listening skills, anger m anagem ent, and the quality o f the m o th e r-ch ild relationship. The researchers were careful to m ake ongoing “quality c o n tro l” assessm ents, ensuring full participation and consistent application o f the program plans by group leaders. M others and children com pleted parallel versions o f the sam e questionnaires, assessing the m o th e r-ch ild relationship, m others’ parenting, and o th er aspects o f family functioning m ore specific to the divorce. Evidence for positive change in family functioning was m ore notable based on m others’ reports th an on children’s reports. C hildren w hose m others participated in the program reported slightly fewer divorce-related negative events and significantly m ore acceptance o f their m others. However, these children d id n o t differ from c o n trol-group children in their descriptions o f oth er areas o f family life, such as the openness o f co m m u n i­ cation, positive routines, the consistency o f m others’ discipline, or extent o f con­ tact w ith fathers. In addition, co n trary to expectations, children in the intervention group reported receiving less su p p o rt from outside adults. By contrast, pre- and posttest scores indicated several program -related changes in family functioning according to m others’ reports. Intervention m others reported significantly m ore open com m unication and m ore positive routines after treatm ent and described them selves as slightly m ore accepting o f th eir children than they were before the p rogram . M others w ho had described themselves as show ing very inconsistent dis­ cipline before the p rogram showed gains in this area. Finally, m others in the p ro ­ gram reported fewer negative events related to the divorce and a m ore positive a t­ titu d e ab o u t changing fathers’ visitation schedules. Program involvem ent by m others was associated inconsistently w ith benefits in children’s m ental health outcom es, according to m others’ ratings o f child behavior o n the Achenbach Child Behavior Checklist. Overall, children whose m others p a r­ ticipated in the program show ed drops in aggression b u t were no different than control children in their anxiety, behavior problem s, o r depression. C hildren w ith the m ost preprogram problem s showed significant gains, however, in their m o th ­ ers’ Achenbach ratings o f total behavior problem s. P a re n tin g T h ro u g h C h a n g e (F orgatch & D eG a rm o , 1998). This interven­ tion program grew o u t o f the O regon Divorce Study, a longitudinal study o f re­ cently separated families. Based on Patterson’s (1982) m odel o f coercive family processes, this p rogram is geared specifically tow ard im proving children’s o u t­

14.

PSYCHOLOGICAL INTERVENTIONS

335

com es at hom e and in school by altering m others’ discipline practices in the early stages o f divorce. The researchers noted that it m ay be m ore effective to intervene w ith m others rath er than w ithin the school setting because parents are m ore co n ­ sta n t in the lives o f their children th an are teachers a n d school program s. T he re­ searchers also hypothesized th at changes in m aternal parenting practices would effectively im prove children’s behavior w ithout intervening with children directly. Participants were m others w ho had been separated w ithin the last 24 m onths and their sons, w ho were in first to third grade. By ran d o m assignm ent, 153 fam i­ lies were enrolled in the intervention program , an d 85 families served as experi­ m ental controls. Parents m et in a series o f 14 p arent gro u p sessions, m ade up o f 6336to 16 m em bers, held at the O regon Social Learning Center. T he groups were designed to provide training in parenting practices (e.g., discipline, contingent encouragem ent, m on itoring, and problem solving) and in o th er issues relevant to divorcing w om en (e.g., regulating negative em otions, m anaging interpersonal conflict). Staff (w ho were all w om en, b o th professionals and paraprofessionals) also m ade m id-w eek telephone calls to m others, encouraging use o f the proce­ dures and troubleshooting problem s, and provided individual sessions as needed. A m anualized treatm ent approach (Forgatch, 1994) and ongoing quality control assessm ents were used to ensure the integrity o f the intervention program . The effectiveness o f the Parenting T hrough C hange program was assessed w ith m ultiple m easures and by m ultiple inform ants, w ith assessm ents m ade every 6 m o n th s (i.e., baseline, treatm en t term ination, a n d follow up). M othrs’ ratings in ­ dicated a high degree o f consum er satisfaction. M ore im p o rtan t, the program was associated with significant im provem ents in m aternal discipline (i.e., reductions in negative reinforcem ent, observed in stru ctu red m o th e r-ch ild problem -solving tasks) and significant im provem ents in teachers’ ratings o f children’s behavior at school. Im provem ents in school behavior were in tu rn related to gains in reading achievem ent scores on the W oodcock-Johnson Psychoeducational Battery, al­ th o u g h p ath analysis showed that the intervention did not have a direct effect on achievem ent scores. S m a lle r C o n tro l-G ro u p S tu d ie s o f P a ren t-F o cu sed T rea tm e n ts. In this section, several sm aller studies are reviewed because they m ade use o f com parison groups to assess effectiveness. In general, the interventions studied are m ore brief and focused, and research show s m ixed evidence o f effectiveness. T he C hildren First p rogram was designed w ith the narrow focus o f alerting p a r­ ents to the adverse consequences th at poorly m anaged conflict has for children and was lim ited two 90-m inute group sessions. Parents heard a local judge discuss the negative im pact o f divorcing parents’ behavior on children, they viewed six video­ taped vignettes depicting m aladaptive interactions am ong divorced family m em ­ bers, and a group leader then m oderated discussion (K ram er & Washo, 1993). In a study o f the program , outcom e was assessed by com paring 168 parents w ho were m andated to participate in Children First to a group o f 43 divorcing parents from

336

EM F.RY, K I T Z M A N N , W A L D R O N

a nearby area. The results showed high consum er satisfaction but no evidence of im provem ents in parenting, parent-child relationships, or the coparental relation­ ship am ong the parents involved in the Child First program . Over time, partici­ pants reported m ore involvement in other divorce-related program s. No evidence o f effectiveness was found in another study o f a series of weekend workshops designed to assist family m em bers to cope em otionally with divorce and to improve contact and com m unication between ex-spouses and between n o n ­ custodial parents and children (Woody, Colley, Schlegelmilch, & M aginn, 1985). By contrast, som e effectiveness was dem onstrated in a psychoeducational program for divorced fathers (Devlin, Brown, Beebe, 8c Parulis, 1992). The experim ental group com prised the first 15 fathers to enroll in the com m unity-based program , and the waiting list com prised the next 15 fathers who called. Fathers m et as a group for 90 m inutes in 6 weekly sessions, typically led by two male facilitators, with a female leader present for sessions focused on coparenting. Fathers in the treatm ent group showed m ore satisfaction with parenting and reported greater effectiveness in talking with and listening to their children. However, there were no program -associated changes in child support payments, contact with children, or num ber o f child-related problem situations reported by the fathers. Lee and H ett (1990) dem onstrated increases in general adult adjustm ent associ­ ated with a program designed to teach coping skills to recently separated and d i­ vorced adults. T heir sample o f 24 adults (mostly women) were random ly assigned to a treatm ent group or to a waiting list control group. The eight group sessions fo­ cused on helping group m em bers cope with divorce-related traum a through stress m anagem ent, com m unication skills, and discussion of divorce-related problems. The adults in the program showed program -related drops in depression and anxi­ ety and gains in general adjustm ent and ability to form m eaningful relationships. Several other studies have used control-group com parisons to assess the effec­ tiveness o f divorce-related intervention program s (Kessler, 1978; Malouff, Lanyon, ScSchutte, 1988; Salts 8c Zongker, 1983; Thiessen, Avery, 8c foanning, 1980). H ow ­ ever, because these program s targeted general adult adjustm ent w ithout a focus on parenting, they are not described here (for a review, see Lee et al., 1994).

Summary and Evaluation Parent-based intervention program s for families o f divorce are com m only used by the com m unity and even are court-m andated in some areas. A lthough parents in these program s consistently report high consum er satisfaction, the quality o f these program s in term s o f objective gains is quite variable. Some program s show gains in parent adjustm ent, but these gains generally do not have the beneficial impact on child adjustm ent that researchers m ight hope. Studies that reported im prove­ m ents in child adjustm ent showed that the quality o f the m other-child relation­ ship and parenting skills were im portant m ediators o f change (Wolchik et al., 1993). C om m unity enthusiasm for parenting program s offers investigators an o p ­

14.

PSYCHOLOGICAL INTERVENTIONS

337

p o rtu n ity to develop m ore effective m eans o f service delivery, to use a stronger the­ oretical and em pirical foundation, to include a stronger developm ental focus, and to develop alternative form ats for delivering different interventions.

COPARENTING: DIVORCE M ED IATIO N Troubles in the ongoing coparenting relationship betw een separated and divorced parents are the focus o f treatm ent in family therapy a n d increasingly in courtm andated or voluntary parent education program s such as C hildren First. U nfor­ tunately, there are no em pirical evaluations o f family therapy for im proving c opar­ enting follow ing divorce (except as a p a rt o f ongoing m ediation and a rbitration), and studies o f paren t education to date have been lim ited to poorly controlled evaluation research. T hus, research on intervention in coparenting follow ing di­ vorce is lim ited to the investigation o f divorce m ediation. M ediation is intended prim arily to be a dispute resolution technique, b u t the process also can be viewed as a preventive intervention that m ay im prove the coparenting relationship and thereby affect children’s well-being follow ing divorce (Em ery 8c Wyer, 1987b).

Characteristics and Goals Divorce m ediation offers divorcing parents a forum for dispute settlem ent at a tim e w hen com m unication betw een them obviously is very difficult. M ediation is based on the assum ption that cooperative negotiation will b o th facilitate sh o rt­ term dispute settlem ent a n d p rom ote contin u ed cooperation in the ongoing rela­ tionship betw een form er spouses w ho rem ain parents. T he process m ay take place in courts, social service agencies, o r private practices, and it typically involves a lim ited n u m b er o f sessions ranging from 1 up to 15 or perhaps 20 hours total. However, m ost m ediation, and m ost research on m ediation, has focused on struc­ tured, sh o rt-term intervention offered in a co u rt setting. The clients seen in the court setting are an unrepresentative sam ple o f the divorcing population, namely, those people w ho are contesting various aspects o f their settlem ents and are re­ questing court involvem ent. In short, m ediators and m ediation research deals w ith the m ost acrim onious divorces. Divorce m ediators com e from m any different professions, b u t they prim arily are attorneys, psychologists, and social workers. T he m ed iato r’s professional back­ g ro u n d m ay influence how m uch the process addresses relationship issues, b u t in al! cases, the overriding focus o f m ediation is on negotiating a divorce settlem ent, n o t on im proving fam ily relationships (Emery, 1994).

Research on Mediation M ediation has been com pared w ith litigation in a handful o f controlled studies. As such, m ediation has been evaluated in term s o f its success in resolving disputes, the

338

EMERY, K I T Z M A N N , W A L D R O N

durability o f m ediated agreements, and client satisfaction with the process. The fo­ cus o f the present review is on m ediation as a preventive intervention, although the success o f the procedure in m eeting other goals is also noted. D isp u te R esolution. In an initial study and in a replication, where families were random ly assigned to m ediate or litigate their custody disputes, m ediation clearly prom oted parental agreem ent and reduced the need for court hearings (Emery, Matthews, & Wyer, 1991; Emery & Wyer, 1987a). Only 4 o f the 35 fam i­ lies assigned at random to m ediation proceeded to court, in contrast to 26 of 36 families in the litigation control group. Disputes also were resolved m ore quickly in m ediation than in adversary settlem ent (Emery et al., 1991). Two other studies that used random assignm ent also found that m ediation reduced court hearings (Irving, Benjamin, Bohm, 8c M acDonald, 1981; Pearson 8c Thoennes, 1984). In one study of private m ediation, 59% of the partners reached a com plete divorce agreement, and an additional 15% reached a partial agreem ent (Kelly 8: Duryee, 1992). Beyond encouraging initial settlements, compliance w ith agreements may be som ewhat higher following m ediation than litigation. In random ized trials, compliance with child support was higher following m ediation than litigation, and evidence indicated that fathers otherwise stayed m ore involved in their chil­ dren’s lives (Dillon 8c Emery, 1996; Emery, Matthews, 8c Kitzm ann, 1994). O ther controlled studies also have found lower rates o f relitigation following m ediation, at least under certain circumstances (Kelly, 1990; Margolin, 1973; Pearson 8c T hoennes, 1989). The assessment of parental satisfaction following m ediation and litigation has been another m ajor dispute resolution focus of m ediation research. Evidence is consistent in finding m ore satisfaction with m ediation than with adversary settle­ m ent in studies using random assignm ent or com parison group designs (Emery 8c Wyer, 1987a; Emery et al., 1991; Emery et al., 1994; Irving et al., 1981; Kelly, 1990; Pearson, 8c Thoennes, 1984). Some controversy has been generated about w hether fathers are m ore satisfied with m ediation than are m others, but gender differences appear to be m ore attributable to the changing background o f litigation (which is m oving toward gender neutrality) than of m ediation (Emery, 1994). C o p a ren tin g a n d P sychological A d ju s tm e n t. In the m ost extensive re­ search on the topic, no differences were found for m ost m easures o f parents’ and children’s m ental health following m ediation or litigation (Em ery 8c Wyer, 1987a; Emery et al., 1991; Emery et al., 1994). On a m ore positive note, fathers who p a r­ ticipated in m ediation reported less coparenting conflict one year after dispute set­ tlem ent, but the difference fell just short o f statistical significance (Emery et al., 1994). Furtherm ore, a m ore positive coparenting relationship and m ore father in ­ volvement in childrearing was reported in a 9-year follow-up pilot study (Dillon 8c Emery, 1996). Finally, a decline in parental conflict over tim e was associated with

14.

PSYCHOLOGICAL IN TERVENTIO NS

339

im proved child m ental health when the m ediation and litigation groups from this study were com bined (K itzm ann & Emery, 1994). T he data available from o th er studies also have failed to indicate differences in m ental health betw een family m em bers using m ediation and adversary settlem ent. N o differences in child outcom es were found in a b rie f assessm ent in one study {Pearson & T hoennes, 1984), and detailed assessm ents o f the form er partn ers in an o th er study sim ilarly found no differences in ad u lts’ psychological adjustm ent {Kelly, 1990; Kelly, Gigy, & H ausm an, 1988). O ne interesting direction for intervention and research on the effect o f m edia­ tion on m ental health and family relationships is increasing the d u ratio n o f m edi­ ation, particularly by including in term itten t contact follow ing dispute settlem ent or even ongoing supervision or arbitration. Johnston a n d Roseby (1997) found that “therapeutic m ediation,” w hich included rath er extensive assessm ent and in ­ dividual prenegotiating counseling, produced considerable agreem ent am ong families w ho had repeatedly litigated divorce disputes. O f particular interest, m ore cooperative and less conflictual coparenting relationships were produced by group intervention m ethods (i.e., seeing several disputing families in a g roup), whereas b e tte r p a ren t-ch ild relationships cam e from family- rath er than group-focused intervention.

Summary and Evaluation The failure to dem onstrate differences in ad u lts’ and children’s m ental health fol­ low ing m ediation m ight be due to any nu m b er o f substantive and m ethodological problem s. M ediation typically is very brief, and the parents seen in m ediation are highly acrim onious. In addition, only a few studies o f possible psychological o u t­ com es have been conducted, and m any possible targets o f change have not been m easured in research conducted to date. F urtherm ore, the d u ra tio n o f follow -up has been sh o rt in alm ost all research, and m ore differences m ay em erge in long­ term follow -up studies (D illon & Emery, 1996). At the sam e tim e, som e d o c u ­ m ented effects, such as the success in resolving disputes, client satisfaction, and the increased involvem ent o f nonresidential fathers follow ing m ediation rath er than litigation, are im p o rta n t outcom es regardless o f th eir relation w ith m easures of children’s or adults’ m ental health. Clearly, m ore research is needed on the psycho­ logical and relationship consequences o f m ediation and adversary settlem ent, and such research obviously is o f considerable social im portance.

SUMM ARY AND DIRECTIO N S FOR FUTURE RESEARCH Taken as a whole, the em pirical evidence on the success o f intervention w ith c h il­ dren, parents, and coparents d u rin g or following separation and divorce is far from com pelling. D ata from controlled studies indicate that school-based groups for

340

E ME R Y , K I T Z M A N N , W A L D R O N

children have beneficial effects, bu t the positive outcom es are generally small in m agnitude and focused on subclinical distress rather than disordered em otions, cognitions, o r behaviors. Parenting groups are viewed positively by adults and aid in facilitating their adjustm ent to divorce, but the eased parental transition pro­ duces few dem onstrable benefits for children. Moreover, divorce m ediation settles m any disputes, produces higher levels o f consum er satisfaction than adversary set­ tlem ent, leads to som ewhat better coparenting relationships, an d helps to encour­ age nonresidential parents’ continued econom ic and parenting responsibilities, but m ediation has not been shown to im prove parents’ or children’s psychological functioning. The limited benefits dem onstrated in intervention research to date m ust be ac­ knowledged squarely. At the same tim e, there are several im portant reasons for tem pering the conclusion that the program s produce only m odest benefits. First, m ost research conducted to date was based on brief, focused intervention that is often offered in a group form at w ith n o follow-up sessions. Given the num ber and range o f disruptions that predate and follow separation and divorce, it would be surprising if eight sessions in a children’s group, six weeks in a parenting class, o r four sessions in m ediaton produce substantial psychological im provem ents in family m em bers in the short or the long run. The extensive body o f descriptive re­ search on divorce indicates that the transition involves substantial psychological, social, and econom ic stressors over a prolonged period o f time. A lthough various brief interventions are helpful, it would be unrealistic to expect any o f them to offer a quick fix for separated and divorced families. Clearly, there is a need for m ore research on divorce interventions that offer a range of services to children, parents, and divorced families and include or at least offer occasional follow-up or “booster” sessions over a period o f perhaps a year or m ore. Second and m ore broadly, there sim ply is a need for m ore research on divorce interventions. Given the limited research conducted to date, it w ould be prem ature to conclude that intervention in divorce produces lim ited benefits, that is, to accept the null hypothesis. This is especially true because only a few styles and form s of intervention have been studied systematically. A plethora o f program s have been developed to help divorcing families. Too often, these interventions are devel­ op ed — and often m andated by c ourts— w ithout em pirical su p p o rt or evaluation, especially evaluation involving random assignm ent or adequate com parison groups. Both for clinical and research purposes, moreover, intervention program s would benefit if they systematically incorporated b oth the empirical findings and the theoretical perspectives that can be gleaned from the rich body o f descriptive research on divorce (Barber, 1995; Grych & Fincham, 1992; Lee et al., 1994). In addition to research on divorce program s, epidemiological data make it clear that children and adults from divorced families are significantly overrepresented in the outpatient m ental health treatm ent population. IJnfortunately, there is no re­ search on the outcom e of outpatient psychotherapy directed at divorce, and little research has been conducted on w hether traditional therapies (e.g., parent train­

14.

PSYCHOLOGICAL

in ter v e n tio n s

341

ing) are m ore or less effective w ith divorced parents (for a review, see Emery, 1998). Both o f these psychotherapy issues m erit detailed investigation, even in an era w here psychotherapy research is dom inated by studies o f single m ental disorders (which often are c o-m orbid w ith life events w ith divorce as well as w ith o th er em o­ tional difficulties). T hird and even m ore broadly, it m ust be recognized that the social and legal culture o f divorce has changed rapidly in recent decades, an d o u r societal views a nd laws continue to evolve. Interventions in divorce m ust be viewed against this evolving backdrop because the social, legal, and cultural context m akes it m ore or less easy (or appropriate) to norm alize divorce for children, su p p o rt single parents, o r encourage cooperative coparenting. For exam ple, differences found betw een w om en’s relative satisfaction w ith m ediation o r litigation in two studies, one in Virginia (Emery, M atthews, 8c Wyer, 1991; Em ery 8c Wyer, 1987a) and an o th e r in C alifornia (Kelly, 1989,1990), appear to be due to the con trastin g social and legal background of litigation, n o t m ediation, in the two states (Emery, 1994). Indeed, w ith the exception o f child su p p o rt enforcem ent, the law appears to have been a weak intervention in divorce, as changes in divorce law in recent decades appear to have followed rath er than preceded changes in the public’s attitude and behavior (Emery, 1998). For exam ple, the passage o f no-fault divorce laws seem s to have been a consequence, n o t a cause, o f increasing rates o f divorce in the U nited States (G lenn, 1997). Such observations draw attention to the im p o rtan ce o f culture in affecting the behavior o f individuals, and the lim ited effectiveness o f psychological interventions in divorce also seems m ore significant w hen viewed against the back­ dro p o f the lim ited influence o f the law.

REFERENCES Alpert-Gillis, L., Pedro-Carroll, J., & Cowen, E. (1989). The children o f divorce intervention program: Development, implementation, and evaluation o f a program for young urban children. Journal o f Consulting and Clinical Psychology, 5 7,583-589. Anderson, R„ Kinney, J., & Gerler, E. (1984). The effects of divorce groups on children's classroom be­ havior and attitudes toward divorce. Elementary School Guidance and Counseling, 19, 70-76. Barber, B. (1995). Preventive intervention with adolescents and divorced mothers: A conceptual frame­ work for program design and evaluation. Journal o f Applied Developmental Psychology, 16,481-503. Bloom, B. L., Hodges, W. F., 8c Caldwell, R. A. (1982). A preventive program for the newly separated: Ini­ tial evaluation. American Journal o f Community Psychology, 10,251-264. Bloom, B. L., Hodges, W. F., Kern, M. B., 8c McFaddin, S. C. (1985). A prevention program for the newly separated: Final evaluations. American Journal o f Orthopsychiatry, 55,9-26. Bornstein, M., Bornstein, P., & Walters, H. (1988). Children of divorce: Empirical evaluation of a grouptreatment program. Journal o f Clinical Child Psychology, 17,248-254. Burke, D., & Van de Streek, L. (1989). Children o f divorce: An application of Hammond's group coun­ seling for children. Elementary School Guidance and Counseling, 2 4 ,112-118. Cantor, D., 8c Drake, C. (1983). Divorced parents and their children: A guide for mental health practition­ ers. New York: Springer. Cowen, E., Hightower, A. D., Pedro-Carroll, J., 8c Work, W. (1989). School-based models for primary

342

EMERY, K I T Z M A N N , W A LD RO N

prevention programming with children. In R. Lorion (Ed.), Prevention in human services (Vol. 7, pp. 133-160). Binghamton, NY: Haworth Press. Cowen, E. Hightower, A. D., Pedro-Carroll, J., Work,W.,Wyman, P., 8c Haffey, W. (1996). School based pre­ vention for children at risk: The primary mental health project. Manuscript submitted for publication. Crosbie-Burnett, M., 8c Newcomer, L. (1990). Group counseling children of divorce: The effects of a multimodel intervention. Journal of Divorce, 13,69-78. Devlin, A. S., Brown, E. H., Beebe, J., 8c Parulis, E. (1992). Parent education for divorced fathers. Family Relations, 41, 290-296. Dillon, P., 8c Emery, R. (1996). Divorce mediation and resolution of child custody disputes: Long-term effects. American Journal o f Orthopsychiatry, 66, 131-140. Eekelaar, I. (1991). Regulating divorce. New York: Oxford University Press. Emery, R. (1982). Interparental conflict and the children of discord and divorce. Psychological Bulletin, 92,310-330. Emery, R. E. (1994). Renegotiating family relationslrips: Divorce, child custody, and mediation. New York: Guilford. Emery, R. (1998). Marriage, divorce, and children’s adjustment (2nd ed.). Thousand Oaks, CA: Sage. Emery, R. E., Matthews, S., 8c Kitzmann, K. (1994). Child custody mediation and litigation: Parents’ sat­ isfaction and functioning a year after settlement. Journal of Consulting and Clinical Psychology, 62, 124-129. Emery, R. E., Matthews, S., 8c Wyer, M. M. (1991). Child custody mediation and litigation: Further evi­ dence on the differing views of mothers and fathers. Journal o f Consulting and Clinical Psychology, 59,410-418. Emery, R. F.., & Wyer, M. M. (1987a). Child custody mediation and litigation: An experimental evalu­ ation of the experience of parents. Journal o f Consulting and Clinical Psychology, 55, 179-186. Emery, R. E., 8c Wyer, M. M. (1987b). Divorce mediation. American Psychologist, 42,472—480. Forgatch, M. S. (1994). Parenting through change: A training manual. Eugene, OR: Oregon Social Learn­ ing Center. Forgatch, M. S., 8c DeGarmo, D. S. (1998). Promoting boys* academic outcomes: An effective parent trainitig program for divorcing mothers. Unpublished manuscript, Oregon Social Learning Center, Eu­ gene, OR. Glenn, N. D. (1997). A reconsideration of the effects of no-fault divorce on divorce rates. Journal of Marriage and the Family, 5 9 ,1023-1025. Grych, J., 8c Fincham, F. (1992). Intervention for children of divorce: Toward greater integration of re­ search and action. Psychological Bulletin, 111, 434-454. Guidubaldi, J., Perry, J., 8c Cleminshaw, H. (1984). The legacy of parental divorce. In B. Lahey 8c A. Kazdin (Eds.), Advances in child clinical psychology. {Vo\. 7, pp. 109-151). New York: Plenum. Gwynn, C., 8c Brantley, H. (1987). Effects of a divorce group intervention for elementary school chil­ dren. Psychology in the Schools, 24A 61-164. Hethcrington, E. M. (1989). Coping with family transitions: Winners, losers, and survivors. Child Developmen t, 6 0 ,1-14. Hetherington, E. M. (1991). Presidential address: Families, lies, and videotapes. Journal o f Research on Adolescence, 1, 323-348. Hctherington, E. M. (1993). An overview of the Virginia Longitudinal Study of Divorce and Remarriage with a focus on early adolescence. Journal o f Family Psychology, 7, 39-56. Hodges, W. (1991). Interventions for children o f divorce: Custody, access, and psychotherapy. New York: Wiley. Hodges, W., & Bloom, B. (1986). Preventive intervention program for newly separated adults: One year later. Journal o f Preventive Psychiatry, 3( 1), 35-49. Howard, K. I., Cornillc, T. A., Lyons, J. S., Vessey, J. T., Lueger, R. J., 8c Saunders, S. M. (1996). Patterns of service utilization. Archives of General Psychiatry, 53,696-703. Irving, H., Benjamin, M., Bohm, P., 8c MacDonald, G. (1981). A study of conciliation counseling in the

14.

PSY CH O LO G ICA L IN T E R V E N T IO N S

343

family court of Toronto: Implications for socio-legal practice. Toronto: Department ofNational Health and Welfare and the Ontario Ministry of the Attorney General. Johnston, J. R., 8c Roscby, V. (1997). In the name of the child: A developmental approach to understanding and helping children o f conflicted and violent divorce. New York: Free Press. Kalter, N., Schaefer, M., Lesowitz, M., Alpcrn, D., & Pickar, J. (1988). School-based support groups for children of divorce. In B. Gottlieb (Ed.), Marshaling social support: Formats, processes, and effects (pp. 165-185). Newbury Park,CA: Sage. Kalter, N„ 8c Schreier, S. (1993). School-based support groups for children of divorce. In J. Zins 8c M. Elias (Eds.), Promoting student success through group interventions (pp. 39-66). Binghamton, NY: Haworth. Kelly, J. B. (1989). Mediated and adversarial divorce: Respondents’ perceptions of their processes and outcomes. Mediation Quarterly, 24,71-88. Kelly, J. B. (1990). Final report. Mediated and adversarial divorce resolution processes: An analysis of post­ divorce outcomes. (Available from the author, Northern California Mediation Center, 100 Tamal Plaza, Suite 175, Corte Madera, CA 94925.) Kelly, J. B.,8r Duryee, M. A. (1992). W omens and men’s views of mediation in voluntary and mandatory mediation settings. Family and Conciliation Courts Review, 30,34-^9. Kelly, J. B., Gigy, L., 8c Hausman, S. (1988). Mediated and adversarial divorce: Initial findings from a longitudinal study. In J. Folbcrg&A. Milne (Eds.), Divorce mediation: Theory and practices (pp. 453474). New York: Guilford. Kessler, S. (1978). Building skills in divorce adjustment groups. Journal o f Divorce, 2 ,209-216. Kitzmann, K. M., 8c Emery, R. E. (1994). Child and family coping one year after mediated and litigated child custody disputes. Journal o f Family Psychology, 8 , 150-157. Kramer, L., 8c Washo, C. A. (1993). Evaluation of a court-mandated prevention program for divorcing parents. Family Relations, 4 2 ,179-186. Laumann-Billings, L., 8c Emery, R. li. (1998). Young adults* painful feelings about parental divorce. Un­ published manuscript, University of Virginia. Lee, J. M., 8c Hett, G. G. (1990). Post-divorce adjustment: An assessment of a group intervention. Cana­ dian Journal of Counseling, 2 4 ,199-209. Lee, C. M., Picard, M., 8c Blain, M. D. (1994). A methodological and substantive review of intervention outcome studies for families undergoing divorce. Journal of Family Psychology, 8, 3-15. Malouff, J. M., Lanyon, R. I., &Schuttc, N. S. (1988). Effectiveness of a brief group RET treatment for di­ vorce-related dysphoria. Journal o f Rational-Emotive and Cognitive-Behavior Therapy, 6 ,162-171. Margolin, F. M. (1973). An approach to resolution o f visitation disputes post-divorce: Short-term counsel­ ing. Unpublished doctoral dissertation, United States International Uni%ferity, San Diego, CA. Omizo, M., 8c Omizo, S. (1987). Group counseling with children of divorce: New findings. Elementary School Guidance and Counseling, 22,46-53. Patterson, G. (1982). Coercive family process. Eugene, OR: Castalia. Pedro-Carroll, J., Sutton, J., 8c Black, A. (1993). The Children o f Divorce Intervention Program: Preventive outreach to early adolescence (Final Report). Rochester, NY: Rochester Mental Health Association. Pearson, J., 8c Thoennes, N. (1984). Final report of the divorce mediation research project. (Available from authors, 1720 Emerson St., Denver, CO 80218.) Pearson, J., 8c Thoennes, N. (1989). Divorce mediation: Reflections on a decade of research. In K. Kressel 8c D. Pruitt (Eds.), Mediation research (pp. 9-30). San Francisco: Jossey*Bass. Pedro-Carroll, J. (1993). The children of divorce intervention program: A prevention program designed to foster resilience and enhance children s capacity to cope. Unpublished manuscript. Pedro-Carroll, J. (1997). The children of divorce intervention program: Fostering resilient outcomes for school-aged children. In G. Albee 8c Gullotta (Eds.), Primary prevention works, Vol 6: Issues in chil­ dren’s and families' lives (pp. 213-238). London: Sage. Pedro-Carroll, J., 8c Alpert-Gillis, L. (1997). Preventative interventions for children of divorce: A devel­ opmental model for 5 and 6 year-old children. Journal o f Primary Prevention, 18, 5-23.

344

EMERY, K I T Z M A N N , W A L D R O N

Pedro-Carroll, J., Alpert-Gillis, L., & Cowen, 11. (1992). An evaluation o f the efficacy of a preventive in tervention for 4 th—6th grade urban children of divorce. Journal o f Primary Prevention, 13,115-130. Pedro-Carroll, J., & Cowen, E. (1985). The children cope with divorce program: An investigation of the efficacy of a school-based prevention program. Journal o f Consulting and Clinical Psychology, 53, 603-611. PedrO'Carroll, J., Cowen, E., Hightower, A., 8c Guare, J. (1986). Preventive intervention with latencyaged children of divorce: A replication study. American Journal of Community Psychology, 14, 277290. Roseby, V., & Deutsch, R. (1985). Children of separation and divorce: Effects of social role-taking group intervention on fourth and fifth graders. Journal o f Clinical Child Psychology\ 14,55-60. Salts, C., 8c Zongker, C. (1983). Effects of divorce counseling groups on adjustment and self-concept. Journal of Divorce, 6, 55-67. Smith, M., 8c Glass, G. (1977). Meta-analysis of psychology outcome studies. American Psychologist, 32, 752-760. Stolberg, A., 8c Cullen, P. (1983). Preventative interventions for families of divorce: The Divorce Adjust­ ment Project. In L. A. Kurdek (Ed.), New directions in child development, Vol. 19: Children and di­ vorce (pp. 71-81). San Francisco: Jossey-Bass. Stolberg, A., 8c Garrison, K. (1985). Evaluating a primary prevention program for children of divorce. American Journal o f Community Psychology, 13,111-124. Stolberg, A., 8c Mahler, J. (1994). Enhancing treatment gains in a school-based intervention for children o f divorce through skill training, parental involvement, and transfer procedures. Journal of Consult­ ing and Clinical Psychology, 6 2 ,147-156. Thiessen, J. D., Avery, A. W., 8c Joanning, H. (1980). Facilitating postdivorce adjustment among women: A communication skills training approach. Journal o f Divorce, 4 ,35-44. Wolchik, S. A., West, S. G., Westover, S., Sandler, L N., Martin, A., Lustig, J„ Tein, J., 8c Fisher, J. (1993). The Children of Divorce parenting intervention: Outcome evaluation of an empirically based pro­ gram. American Journal of Community Psychology, 21,293-331. Woody, J. D., Colley, P., Schlcgelm ilch,8c Maginn, P. (1985). Relationship between ex-spouses: Impact of a weekend educational workshop. International Journal o f Family Psychiatry, 6 ,99-115. Zill, N., Morrison, D. R., 8c Coiro, M. J. (1993). Long-term effects of parental divorce on parent-child relationships, adjustment, and achievement in young adulthood. Journal o f Family Psychology, 7, 91-103.

Author Index

B Abidin, R. R.. 49, 62 Ablon, S., 48, 4 9 ,6 2 A chenbach, T. M .f 49, 62, 97, 1 0 3 ,115, 185, 187, 242, 24 9 , 2 5 9 ,2 6 9 Acitelli, L. K., 34, 3 5 , 4 1 ,4 3 Acock, A., 24, 40, 41, 2 5 9 , 2 7 0 Adams, R. G.. 1 6 9 ,189 A dclm ann, P K., 25, 27, 41 A hrons, C , 2 4 9 ,2 4 9 Aiken, L. S., 296, 318 Alexander, K. L., 197,2 1 9 A lpern. D., 330, 343 Alpert-Gillis, L., 325, 327, 3 41, 343 Amato, P. R., 23, 24. 4 1 , 47, 62, 66, 88, 9 3 ,9 4 , 95, 1 1 4 , 115, 147, 148, 150. 151, 153, 154, 161, 161, 162, 259, 26 9 A nderson, E. R., 68, 85, 89, 227, 229, 2 4 9 ,2 5 1 , 296, 299, 300, 302, 303. 304, 307. 308, 310, 31 2 , 3 16, 3 18. 3 1 8 ,3 1 9 A n d e rso n ,). Z., 261, 262, 267. 2 6 9 A nderson, R., 328, 341 A nderson, S. A., 284, 293 Aquilino, W. S., 1 5 5 ,162 Arbuckle, J. L., 158, 162 Arthur, J., 108, 116 Aseltine, R. H., 66, 88, 237, 2 4 9 A stone, N ., 119, 144, 167,1 8 9 Avery, A. W., 3 3 6 ,3 4 4

Baerger, D. R., 25, 2 7 , 41 Baker, D., 51, 62 Baldwin, A ., 48, 49, 64, 67, 68, 88 Baldwin, C ., 48, 49, 64, 67, 68, 8 8 Baldwin, W., 196, 21 9 Bank, L., 229, 2 4 9 Barber, B., 340, 341 Barnard, K. E., 209, 2 1 9 B arnett, R. C., 158, 162 Barrera, M . J., 2 3 7 ,2 4 9 Barry, W. A., 1 1 ,2 2 Baruch, G. K., 158,1 6 2 Basham, R. B , 2 3 8 ,251 Bates, J. E.. 48, 49, 63, 64, 68, 88, 201, 202 220 222 B aum rind,’D., 67, 68, 72, 88, 2 0 2 ,219, 2 6 3 , 2 6 6 ,2 6 9 Baydar. N ., 174,1 8 7 Beam an, J., 229, 237, 238, 2 5 2 Beardsall, L., 49, 63 Beavers. W R ., 2 8 1 ,2 9 3 Beebe, J., 336, 342 Belle, D. H., 2 3 8 ,2 4 9 Belsky, J., 18, 20, 23, 41, 148, 162, 177, 187, 237, 238, 239, 2 4 9 Benjam in, M ., 338, 342 B ennett. T. L., 239. 2 5 0 Bennion, L. D., 227, 249, 303, 304, 308, 318 Bensley, L., 180. 190 345

346

AUTHOR INDEX

Benson, P. R., 17, 2 1 Berger, S. H., 227, 2 5 0 , 253, 254, 256, 257, 259, 260, 261, 262, 263, 264, 2 6 5 ,2 6 6 , 267, 2 6 9 , 276, 2 9 3 Bcrgner, R. M.. 108, 1 1 6 Bianchi, S. M., 24 , 4 3 , 161 , 1 6 3 Biederman, J., 48, 49. 6 2 Bigian.A., 108, 1 1 6 Billingsley, A., 24, 25, 4 1 , 166, 1 8 7 Black, A., 327, 330, 3 4 3 Black, M. M., 180, 1 8 7 Blain, M. D., 324, 330, 331, 332, 336, 340, 3 4 3 Blankenhorn, D-, 1 1 7 , 1 4 4 , 147, 1 6 2 Blechman, E. A., 150, 1 6 3 , 237, 2 5 2 Block, J., 93, 9 9 , 1 1 5 , 205, 2 2 1 , 259, 2 6 9 Block, J. H., 93, 99, 1 1 5 , 2 0 5 , 2 2 1 , 259, 269

Bloom, B .,3 3 2 , 3 4 2 Bloom, B. L., 332, 3 4 1 Boethel, C., 257, 262, 264, 2 6 6 ,2 6 9 , 276, 293

Bohm, P., 338, 3 4 2 Bolger, N., 238, 2 5 0 Booth, A., 47, 6 2 , 93. 94, 95, 1 1 5 , 148, 151, 154, 161, 1 6 1 , 1 6 2 , 2 5 9 , 2 6 9 Booth, C. L., 2 0 9 , 2 1 9 Booth, K., 119, 1 4 5 Borduin, C M., 280, 2 8 1 , 2 9 3 Bornstein, M., 325, 329, 3 4 1 Bornstein, P., 325, 329, 3 4 1 Borquez, J., 2 1 7 , 2 2 1 Borrine, M. L., 259, 264, 265, 2 6 9 B os.H ., 1 9 4 ,2 2 2 Bos, J. M., 194, 2 0 5 , 2 2 2 Bowman, P., 32, 4 1 Bradbury, T. N ., 3, 2 1 Bradley, R. H., 180, 1 9 0 , 198, 202, 207, 2 1 1 ,2 1 9 , 2 2 0

Brantley, H., 329, 3 4 2 Braun-Curtin, R., 119, 137, 1 4 5 Brav, J. H., 227, 2 5 0 , 253, 254, 256, 257, 259, 260, 261, 262, 263, 264, 265, 266, 267, 268, 2 6 9 , 2 7 0 , 276, 2 9 3 , 302, 3 1 9 Bridges, M., 259, 2 7 0 Brillon, L., 86 , 8 9 Brody, G., 65, 8 8 Brody, G. H., 227, 2 5 0 Broman, C. L., 25, 26, 37, 4 1 Bronars, S., 126, 1 4 4 Bronfenbrenner, U., 67, 8 8 , 183, 1 8 7 , 197,

220 Brooks-Gunn , )., 48, 49, 6 3 , 166, 167, 171, 172, 174, 177, 178, 179, 180, 181. 182, 183, 185, 186, 1 8 7 , 1 8 8 , 1 8 9 , 1 9 0 , 1 9 1 , 195, 196, 2 2 0 , 2 2 1

Brown, A. C., 267, 2 7 0 Brown, B., 194, 2 2 1 Brown, B. B., 68, 9 0 Brown, C. H., 169, 1 8 9 Brown, E. H., 336, 3 4 2 Brown, N. Y., 259, 26 4 . 265, 2 6 9 Brubaker, T H., 23, 4 1 Brunelli, S. A., 180, 1 9 1 Buchanan, C. M., 47, 62, 98, 1 1 5 Buehlman, K. T., 5, 6, 16, 20, 2 0 ,

280,

293

Bugaighis, M. A., 280, Bumpass, L., 119,129,

294 144, 145,

166,

187

Bumpass, L. L., 147, 1 5 3 , 1 6 2 , 253, 2 7 0 , 278, 279, 2 9 4 Burke, D., 330, 3 4 1 Buriel, R., 66, 68 8 9 Burleson, B. R., 238, 2 5 0 Burton, L. M., 49, 6 3 , 166, 169, 180, 187, 1 8 8

Bus, A. G., 2 0 2 , 2 2 0 Bushwall, S. J., 65, 66, 72, Bryk, A. S., 232, 2 5 0

88,

108,

115

C Cain, V, 196 , 2 1 9 Caldwell, B. M., 198, 202, 207, 2 1 1 ,2 7 9 ,

220 Caldwell, R .A .,3 3 2 , 3 4 1 Call, V. R. A., 278, 279, 2 9 4 Campbell, L. E., 98, 1 1 6 Cancio, J., 24, 40, 4 3 Candell, K., 2 3 7 , 2 5 1 Cantor, D., 332, 3 4 1 Capaldi, D. M., 227, 229, 2 5 0 Carey, W. B., 53, 6 3 Carlsmith, J. M., 65, 66, 72, 8 8 , 108, 1 1 5 Carlson, 197, 2 2 0 Carrere, S., 13, 14, 15, 16, 17, 19, 2 0 , 2 0 Carstensen, L. L., 23, 4 1 Carter, B., 256, 2 7 0 Carter, E. A., 265, 2 5 7 , 2 7 1 Casey, P. H., 180, 1 9 0 Caspi, A., 2 5 , 4 1 , 148, 154, 1 6 2 Castro-Mariin, T., 253, 2 7 0 Cave, G., 1 9 4 ,2 2 2 Ceballo, R., 217, 2 2 1 Chadwick, K., 25, 27, 4 1 Chao, R. K., 68, 8 8 Chao, W., 229, 237, 238, 252 Chase-Lansdale, P. K., 93, 1 1 5 , 1 5 7 , 1 6 2 , 166, 167, 171, 172, 176, 177, 178, 179, 180, 181, 182, 183, 184, 185, 1 8 6 , 1 8 7 , 1 8 8 , 1 8 9 , 1 9 0 , 1 9 1 , 196, 2 2 1 , 2 2 7 , 2 5 0 ,2 5 9 , 2 7 0

AUTHOR INDEX

C hatters, L. M ., 32, 37, 43, 173, 189 Cherlin. A. ]., 93, 11 5 , 147, 157, 162, 178, 188, 2 2 7 ,25 0 , 2 5 9 , 27 0 , 273, 2 9 0 ,2 9 3 C horost, A. F. 23, 42 C hristensen, A., 1 4 ,2 7 C icchetti, D., 49, 63 C lem ents, M ., 23, 42 Cleminshaw, H. K., 68, 89, 325, 342 Clingem peel, W. G., 2 4 , 42 , 61, 63, 94, 96, 116, 178, 189, 2 2 7 ,2 5 1 ,2 5 5 ,2 5 8 , 259, 260, 261, 262, 263, 264. 270, 269, 299, 300, 302. 303, 307, 308, 3 1 0 ,3 1 2 ,3 1 8 ,3 1 8 ,3 1 9 Coan, J., 13, 14, 15, 16, 17, 19, 20, 2 0 Cohen, S., 7 , 2 1 ,7 8 , 88, 237. 2 5 0 Coiro, M. J., 195, 199, 221, 223, 227, 25 2 , 3 2 3 ,3 4 4 Cole, R. E.f 67, 68. 88 Colem an, J., 139, 144 Colem an, M ., 2 5 4 , 27 0 , 260, 27 1 , 274, 275, 276, 277, 278, 280, 287, 289, 290, 293, 294 Coley, R. L., 166, 175, 176, 185, 188 Coll. C. G., 49, 63 Colley, P, 336, 344 Collins, P. H., 27, 41 Com stock, G. W., 199, 2 2 0 Conger, R., 160, 162, Conger. R. D., 229, 237, 238, 25 0 , 252 Conger, K. J., 160, 162 C onners, C. K., 49. 62 Cook, J., 12, 1 9 ,2 / Cooley, M ., 180, 190 C oon, H. M ., 28, 41 Cooney, T. M., 155,1 6 2 C oontz, S., 148,1 6 2 Cooper, H., 260. 271 C opeland, J. M ., 280, 294 Corbin, S. B.„ 87. 90 Cornille, T. A., 323, 342 C asta, F. M., 48, 63 Cow an, C .,2 3 ,4 1 , 101, 103, 115 Cow an, P., 23, 41 Cow an, P. A., 101, 103,1 1 5 Cow en, E., 325, 326, 327, 341, 342, 343 Cox, J. L., 52, 63 Cox, M.. 68, 85, 89, 254, 258, 259, 265, 270 Cox, M. I., 238, 25 0 , 299, 303, 316, 319 Cox, R., 68. 85. 89, 254, 258, 259, 265, 2 7 0 , 299, 3 0 3 ,3 1 6 ,3 1 9 Coyne, J. C., 49, 63, 237, 2 3 9 ,2 5 0 Crawford, A. G., 167, 188 Creighton, L. L., 166, 188 Crockenberg, S., 179, 188, 200, 2 2 0 C rohan, S. 35, 36, 41

347

C rosbie-B urnctt, M., 273, 275, 2 9 0 ,29 3 , 326, 330, 342 Crouter, A. C., 67, 88 Cum m ings, M . £., 95, 96, 97, 103, 115 Cum m ings, E. M ., 154, 162, 3 1 6 , 3 1 9 Cullen, P., 325, 344 Custer, L., 23, 37, 42 C utrona, C. E., 238, 239, 2 5 0

D Darling, C , 27, 42 Davies, P T., 95, 96, 97, 1 0 3 ,115, 154, 162, 3 1 6 , 3 / 9 Davis, J. E., 2 7 , 3 1 , 4 1 ,4 2 Davis, T. J., 31, 32, 41 Deal, J. E., 296, 302, 303, 307, 308, 310, 312 318 3 1 9 Deater-Deckard', K., 48, 49, 50, 52, 53, 6 1 ,6 3 , 68, 88 DeG arm o, D. S., 227, 229, 238, 241, 242, 244, 246, 247, 334, 342 D em o, D. H ., 24, 40, 41, 148, 162, 259, 270 DeRogatis, L. R., 279, 293 De Temple, J.. 204, 205. 2 2 0 D eutsch, R., 329, 344 Devlin, A. S.. 336, 342 Dickerson, B. J., 166, 169, 188 Diener, E., 2 7 9 ,2 9 3 Dillon, P., 338, 3 3 9 ,3 4 2 D ion, M. R., 1 9 9 ,2 1 7 ,2 2 0 D ishion, T., 231, 251 Dodge, K. A., 48, 49, 63, 68, 88, 201,

202 , 220 , 222 D ornbusch, S. M ., 47, 62, 65, 66, 67, 68, 69, 72, 74, 8 5 ,9 0 , 98, 108, 115 D ouvan, E., 23, 25, 28, 33, 37, 41, 42, 43 D ouvan, L., 34, 35, 43 Downey. C ., 49, 63, 237, 2 3 9 ,2 5 0 Drake, C ., 332, 341 D ru ck m an , )., 267, 2 7 0 Dubow, E., 196, 1 9 7 ,2 2 0 Duffy, M. E., 2 3 8 ,2 5 0 Dukes, L. J., 201, 221 D uncan, G., 195, 2 2 0 D unkel-Schetter, C., 239, 2 5 0 D unn, J., 47, 49, 50, 51, 52, 61, 62, 63, 64 D unn, L. M., 185,1 8 8 D uryee, M. A., 338, 343

E Eddy, M. J., 2 3 0 ,2 5 1 Edelbrock, C. S., 9 7, 1 0 3 ,115, 185, 187, 242, 24 9 , 2 5 9 ,2 6 9

348

AUTHOR INDEX

E dm onston, B., 194,2 2 1 G E dw ards,J. N ., 23, 4 3, 148, 151, 1 6 2 Eekelaar, J., 324, 342 Galco, F L., 202, 223 Egeland, B., 197, 202, 203, 204, 205, 217, G anong, L. H ., 254, 2 60, 27 0 , 2 7 1 , 274, 220 , 221,222 275, 276, 277, 278, 280, 287, 289, Ekm an, P., 4, 21 290, 2 9 3 ,2 9 4 Elardo, R. D., 202, 2 2 0 G arbarino, J., 261, 2 7 0 Elder, G. H ., 25. 41, 148, 154, 160, 1 6 2 Garcia Coll, C .f 86, 8 9 Eldred, C. A., 204, 2 0 5 ,2 2 0 ,2 2 3 Garfinkel, 1., 130, 143, 144 Emery, R. E., 48, 49, 64, 66, 88, 93, 116, Garmezy, N ., 48, 63, 1 4 9 ,163, 197, 198, 136, 144, 148, 163, 181, 1 8 8 , 261, 221 2 6 5 ,2 7 0 , 323, 324, 325, 337, 338, G arrison, K., 327, 333, 344 339, 3 4 1 ,3 4 2 , 343 Ge, X., 229, 2 5 0 E m m ons, R. A., 279, 2 9 3 George, S. M ., 166, 169, 188 Ensminger, M. E., 169, 180, 189 Gerler, E., 328, 341 Entw istle, d. R., 197, 2 1 9 Geronim us, A., 126, 144 Epps, E., 32, 41 GiGy, L., 339, 343 Erera-Weatherly, P. I., 275, 2 93 Gjerde. P. F., 93, 99, 115, 25 9 , 26 9 Erickson, M. E, 2 0 2 ,2 2 1 Glass. G., 3 3 1 ,3 4 4 Estes, A., 2 3 1 ,2 5 1 Glenn, N. D ., 153, 1 5 7 , 163, 3 4 1 ,3 4 2 Eyster, S. L., 23, 26, 27, 37, 38, 39, 4 2 Glick, P. C., 61. 63, 65, 89, 25 3 , 27 0 Gold, M., 74, 8 9 Golding, J., 48, 50, 51, 52, 61, 63, 64 F G onzalez, R., 9 8 , 116, 169, 188 Good, E. S., 2 3 7 ,2 5 1 Faraone, S. V., 48, 49, 62 G oodm an, R., 51, 63 Farres, A. M ., 9 8 ,1 1 6 Gordon, L. U., 7 ,2 1 Featherm an, D., 72, 88 G ordon, R. A., 166, 171, 172, 177, 179, Feldman, S., 71, 90 183, 184, 188, 1 8 9 ,1 9 1 , Fetrow, R. A., 229, 230, 249, 251 G ottm an, J., 4, 12, 19 ,2 1 Fincham , F., 324, 325, 326, 330, 340, 342 G ottm an. J. M ., 5, 6, 7, 9, 10, 11, 12, 13. Fincham , F. D., 97, 9 8 ,1 1 6 14, 15, 16, 17. 18, 19, 20, 2 0 ,2 1 , Fine, M. A., 27, 42, 2 6 6 ,2 7 1 , 2 7 4 , 276, 23, 2 8 , 38, 4 1 . 42, 242, 250, 265. 277, 278, 280, 287, 289, 2 9 3 ,2 9 4 270, 280, 293 Fink, B. I ., 194, 2 22 Green, G. J., 267, 2 7 0 Finlay, B., 151,1 6 3 G reen, S., 179, 191 Fischel. J. E., 2 0 2 ,2 2 3 Green, S. M ., 316, 3 1 9 Fisher,J., 333, 336, 344 Greenherger, E., 74, 8 9 Floyd, F. J., 23, 42 Greene. A. D., 185, 190, 196. 20 0 , 217, Ford, M ., 7 1 ,9 0 221 Forehand, R., 65, 88, 22 7 , 2 5 0 Griffin, S., 2 7 9 ,2 9 3 Forgatch, M. S., 65, 66, 88, 227, 229, 238, Grogger, J., 126, 144 241, 242, 244, 245, 246, 2 4 7 , 249, Gross, R. T., 65. 66, 72. 88, 108 ,1 1 5 250, 25 1 , 263, 2 6 5 ,2 7 0 , 334, 335, Grych, J., 324, 325, 32 6 , 330, 340, 342 342 G rynch, J. H., 97, 9 8 ,1 1 6 Foster, M ., 126, 145, 231, 2 3 8 ,2 5 0 ,2 5 1 Guarc, J., 3 2 6 ,3 4 3 Fraleigh, M., 68, 69, 74, 8 8 Guem ey, B. G., 14,2 1 Frate, D. A., 166, 169,1 9 0 G uidubaldi, J., 68, 89, 325, 342 Friedm an, L., 108, 116 G uite, J., 48, 49, 62 Friesen, W .V , 4 ,2 1 Gw ynn, C., 329, 342 Fii, V. R., 68, 89 Furakawa, S., 166, 188 Furstenberg, F. F., 65, 89, 93, 115, 126, H 145, 156, 163, 166, 167, 175, 178, 180, 185, 1 8 7 ,1 8 8 , 196, 220, 2 21, Haas, S. D . .5 ,2 1 227, 250, 2 5 9 ,2 7 0 Hacker, T. A.. 23 8 , 251

AUTHOR INDEX

Haffey, W , 3 2 7 ,3 4 2 Halverson, C. F., 183, 191 Ham pson, R. B., 281, 293 Handal. P. J., 259. 264, 2 6 5 .2 6 9 Hanson, T., 119, 130, 131, 136, 144, 145 Hao. L„ 173, 190 Hardy, J., 167, 189 Harrington, G., 205, 221 Harris, B. F., 166,1 8 9 Harris, K. M „ 175, 188 Harvey, D. M., 261, 270 Haskey, J., 48. 63 Hastorf, A. H„ 65, 66, 72, 88 H atch, R. C , 2 8 0 .2 9 4 H atchett, S., 23. 25, 28, 33, 3 8 ,42, 43 Hausm an, S., 339, 343 Havem an, R., 121, 145 Hawkins, M. W , 17,21 Hawkins, N., 51, 64 Hcavey, C. L., 14, 21 Helsing, K. J., 1 9 9 ,2 2 0 Henderson, V. K„ 178, 184, 190, 227, 229, 238, 250, 251 Henderson-King, D. I I , 3 5 ,4 2 Hendrix, H „ 10, 13, 14, 19, 21 Henick, A., 3 1 3 ,3 1 9 Hennigen, L., 2 7 7 ,2 9 4 Hennighausen, K. H „ 2 0 4 ,2 2 2 H ernandez, D. J., 65, 89, 160, 163, 166, 189 Hertel, R. K., 1 1 ,2 2 H etherington. E. M. 24, 40. 42, 50, 61, 63, 64, 68, 85, 87, 89. 94, 95. 96. 97, 108, 113, 116, 148, 150, 163, 178, 189, 254, 255, 257, 258, 259, 260, 261, 262, 263, 264, 2 6 5 .2 7 0 , 296, 299, 300, 302, 303, 307, 308, 310, 3 1 2 ,3 1 6 , 318, 318. 319, 324, 342 H ett, G. G„ 336, 343 Hightower, A., 326, 343 Hightower, D„ 325, 327, 341, 342 Hill, R .B ., 166,1 8 9 Hinde, R„ 177, 189 Hobfoll, S. E., 238,2 5 1 Hodges, W„ 325, 326, 332, 342 Hodges, W. F„ 332, 341 HoffG insberg, E., 86, 89 Hoffm an. S., 126, 145, 156,1 6 3 Hogan, D. P., 173,1 9 0 Holden, J. M „ 52, 63 llolm berg, D., 35, 43 Hollier, A., 256, 260, 2 6 1,2 6 9 Hollier, E. A., 296, 302, 303, 307, 310, 312 318 319 Hollingshead, A. B„ 231, 247,2 5 1

349

Holloway, S. D „ 237, 251 Hops, H., 108, 116 Hornyak, L. A., 17, 21 H orton, C. B., 108, 116 Horwitz, s M ., 180, 185, 189 Hossain, Z., 3 7 ,4 2 House. J., 24, 43 House, J. S., 2 3 ,4 2 , 153, 163 Howard. IC I., 323, 342 Howe, G., 94, 116 Howell, C. T., 49, 62, 185, 187 Hsdorf, H „ 108,1 1 5 Huck, S., 1 6 0 ,7 6 2 Hughes, M. E., 185,1 8 8 Hughes, R„ 237,2 5 1 Hulgus, Y. F„ 2 8 1 ,2 9 3 Hunter, A. G . , 2 7 ,3 1 ,3 7 , 42 H uston, A. C., 49, 63 H uston, T. L., 23, 42 I Insabella. G. M „ 259, 270 Irving, H., 338, 342

J Jackson, J. S., 32, 43 jayakody, R., 173, 189 Jayaratne.T. E., 217, 221 Jekel, J. F„ 180, 185, 189 Jcndrek, M. P., 166, 189 Jenkins, C. L... 49. 62 Jessor, R., 48, 63 joanning, H., 336, 344 Jodi, K. M., 61, 63, 2 5 9.2 7 0 Johnson, C., 237, 238, 239, 252 Johnson, D. R., 151, J 62 Johnson, L. B„ 2 5 ,3 2 ,4 2 , 43 Johnson, P L., 93, 116 Johnston, J. 108, 116 Johnston, J. H ., 242, 251 Johnston, J. R. 98, 116, 2 6 5 ,2 7 0 . 339, 343 Jouriles, E. N „ 98, 116 Jouriles, E., 261, 268. 270

K Kahan, J. R, 261,2 7 1 Kahn, R. L„ 274, 294 Kalter. N.. 325, 326 Kane, E. W„ 27, 42 Kaniasty, K., 2 3 8 ,2 5 1

350

Karney, B. R., 3, 27 K atz, L. F., 6, 16, 20, 2 0 , 2 8 0 , 2 9 3 Kenney, D. A.. 3 0 9 , 3 1 9 K eith, B., 2 4 , 4 1 , 47, 6 2, 6 6 , 8 8 , 93, 95, 114, 115, 147, 148, 150, 1 62, 2 5 9 , 269 K eith, V .M ., 1 5 1 ,1 6 3 ICellam, S. G., 169, 1 8 0 ,1 8 9 Kelfcher, IC. J., 180, 1 9 0 Kelly, J. B., 3 3 9 , 3 4 1 ,3 4 5 K ent, R. N ., 2 3 1 ,2 5 1 K cphart, G ., 2 5 , 4 2 K ern, M . B., 3 3 2 , 341 Kessler, R. C., 7 , 2 1 , 2 3 7 , 2 4 9 Kessler, S., 3 3 6 , 3 4 3 K ibria, N ., 158, 16 2 Kicly, K., 48. 48, 6 2 K iernan, K. E., 93, 115, 15 7 , 1 6 2 , 227, 2 5 0 . 259, 2 7 0 King, V., 167, 175, 189 K inney, J., 3 2 8 , 341 K itson, G. C., 2 3 7 , 2 5 / K itzm a n n , K., 3 3 8 , 3 4 2 K itzm an n , K. M ., 3 3 9 , 3 4 3 Klebanov, P. K., 1 9 5 .2 2 0 Klepinger, D., 126, 145 H e rm a n , L. V., 180, 185, 189, 1 9 4.2 2 1 K line, M ., 2 6 5 , 2 7 0 K ochm an. T., 3 5 , 3 6 , 4 2 K orenm an, S., 126, 144 Kou, H . S., 180. 1 8 5 ,1 8 9 Kram er, K. B., 153, 157, 163 Kram er, L., 3 3 5 , 3 4 3 K rein, S .F ., 151, 163 Krokoff, L. J.. 5, 1 3 ,2 7 K urdek,L. A., 2 5 9 , 2 6 5 , 2 7 1 , 2 7 6 , 27 7 , 294

L Ladner, J. A ., 1 6 6 , 1 8 9 , 2 0 5 ,2 2 2 La G aipa, J. L., 2 3 8 , 251 Lahey, B., 179, 191 L am born, S. D., 67, 68, 69, 72 , 85, 9 0 Lane, C . 2 3 8 , 251 L anyon, R. I., 3 3 6 , 3 4 3 L aP oint, V , 1 6 6, 1 8 9 Larkin, I