ll'°l'\\lllllllllllllll lallllll J " 31293 00646 461? fir-{19v *6 .i L .a This is to certify that the dissertation entitled CHANGES IN THE POST-DIVORCE FAMILY SYSTEM AND CHILDREN'S ADJUSTMENT. presented by ‘, 1 Naomi S. Goldblum — 1' i has been accepted towards fulfillment of the requirements for Ph.D. Psychology degree in 142/” KM Major professor Date 1/6184 MS U is an Affirmative A (lion/Equal Opportunity Institution 0- 12771 PLACE IN RETURN BOX to roman this chookout from your record. TO AVOID FINES rotun on or baton duo duo. DATE DUE DATE DUE DATE DUE JAN-0 87932: O 2 0 l Jun 0 2 31; I QED—1'7; \v—‘q 34 a . ‘ "t .. to. ' . k. r .‘ : 7*7 .- o’ I ‘ _)‘ I ‘. ‘. . -." -._‘._- t - I .f' l , - - I“ ~ \ : r ' K . ,L MSU Is An Afflrmdivo ActioNEquol Opportunity Instituion i CHANGES IN THE POST-DIVORCE FAMILY SYSTEM AND CHILDREN'S ADJUSTMENT BY Naomi Sara Goldblum A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Psychology 1984 ABSTRACT CHANGES IN THE POST-DIVORCE FAMILY SYSTEM AND CHILDREN'S ADJUSTMENT BY Naomi Sara Goldblum The purpose of the current research was to investigate a theoretical model for the post-divorce family derived from family systems theory. The structure of the divorced family must permit the resolution of the dysfunctional marital dyad, while adapting and maintaining functional parental and parent-child dyads. The long-term adjustment of children from divorced families will primarily reflect whether the family has accomplished these changes in family organization. Subjects were 61 families 1 to 4 years post-divorce who were identified with the assistance of the Court. Data were collected from 24 female and 37 male children, aged 6 to 12, their custodial parents, their school teachers, and 38 of their noncustodial parents. Childrenls and parents' perceptions of the marital, parental and parent-child dyads were measured with a variety of self-report questionnaires. Childrenls adjustment was measured with the Child Behavior Checklist (Achenbach & Edelbrock, 1978) which each parent completed, and by the Teacher's Report Form (Achenbach & Edelbrock, 1980) which the childrenfs teachers completed. Naomi Sara Goldblum MANCOVAHS‘with sex of child and time since the divorce as factors, and age of child as a covariate, revealed no significant differences in children's adjustment. Signifi- cant interaction effects within the custodial parent data revealed that families of girls reported longer visitations at one year, more marital conflict at 3 years, and greater dependency behavior at 4 years since the divorce compared to the families of boys. Within the children's data, there was significant deterioration in their relationships with their fathers over time. Parents' ratings of childrenfs problem behavior were related to the quality of the parent-child interactions, and the quality of the parental dyad. Differences between the custodial and noncustodial parent's ratings of children's problem behavior and competence suggested qualitative dif- ferences in the parents' relationships with their children. The results provide moderate support for the proposed model of the post-divorce family and childrenis adjustment. Implications for clinical work and future research with divorced families are discussed. ACKNOWLEDGMENTS I would like to express my appreciation to the members of my Dissertation Committee. Dr. Andrew Newcomb, my chairperson, was wil ling to accept me as his student at a time when the department was going through a great deal of change. His support and encouragement at that time enabled me to complete the development of ideas which had interested me for a number of years. Dr. Gary Stollak and Dr. Elaine Donelson provided me with a sense of continuity throughout my time at Michigan State, serving on both my Master's and Dissertation Committees. I am particularly grateful to Dr. Lucy Ferguson, who first contributed to my professional development when I was an intern at the Psychological Clinic, and again when I was an intern at the Family Life Clinic. Dr. Ferguson intro- duced me to the family systems literature, and in many ways shaped my growth as a clinician. My research interests also have been indelibly marked by her influence. I would like to express my great thanks to Mr. James Pocock, Friend of the Court for Ingham County, and his entire staff. Mr. Pocock's assistance enabled me to obtain the subjects for this study, and his staff's enthusiasm bouyed me as I made my way through their offices. I would ii also like to thank the families who were willing to participate in this study, with the hope that their time and effort would help others in the future. My deepest thanks go to Rusty Burshell. He was willing to endure my preoccupation with this study for long periods at a time, and to fill my life with richness during our all our years together. iii TABLE OF CONTENTS LIST OF TABLES . . . . . INTRODUCTION . . . . . . REVIEW OF THE LITERATURE A Model for Post-divorce Family Adjustment. . . . . Longitudinal Studies . Cross-sectional and Correlational Studies. Studies Comparing Different Custody Arrangements . Father Absence METHOD 0 O O O O O O O 0 Subjects. . . . . . Design. . . . . . . Procedure . . . . . Instruments . . . . Hypotheses. . . . . RESULTS. . . . . . . . . Determination MANCOVA . . . . . . of Research Studies Organization. . Empirical Literature on Children's Post-divorce Variables Relationship Between Measures of Children's Adjustment and Family Relationship Variables. . Prediction of Children's Problem Behavior and Competence. . . . . DISCUSSION . . . . . . . The Impact of Time Since Divorce and Sex of Child on Children's Adjustment. Relationship Between Measures of Children's IO 10 15 20 23 28 28 32 34 36 4o 43 43 50 57 61 66 66 Adjustment and Measures of Family Relationships . . . . . . . . . . . . . . . . . Relative Contribution of Family Relationship Variables versus other Variables Towards Variance of Children's Adjustment . . . . . . . iv TABLE OF CONTENTS (Continued) Page A Reexamination of the Model for the Post-divorce Family System. . . . . . . . . . . 79 Considerations for Clinical Work. . . . . . . . 82 Considerations for Future Research. . . . . . . 83 APPENDICES A. Initial Contact Letter. . . . . . . . . . . 87 B.- Initial Letter to Teachers. . . . . . . . . 89 C. Child Questionnaire . . . . . . . . . . . . 90 D. Child Report of Parent Behavior Inventory (CRPBI) . . . . . . . . . . . . . 93 E. Parent Questionnaire. . . . . . . . . . . . 100 F. O'Leary-Porter Scale. . . . . . . . . . . . 102 G. Child-Rearing Practices Report (CRPR) . . . 104 H. Child Behavior with Parent Inventory (CBPI). . . . . . . . . . . . . . 112 1. Means and Standard Deviations of Research Variables. . . . . . . . . . . . . 121 J. Correlation Matrix. . . . . . . . . . . . . 123 REFERENCES 0 O O O O O O O O O O O O O O O O O O O O 130 10. ll. 12. l3. 14. 15. l6. 17. LIST OF TABLES Page Custodial Parent Socioeconomic Status . . . . . 29 Noncustodial Parent Socioeconomic Status. . . . 31 Age and Sex of Children . . . . . . . . . . . . 32 Distribution of Children by Time Since Divorce" and Sex of Child, Custodial Parent Sample . . . 33 Distribution of Children by Time Since Divorce and Sex of Child, Noncustodial Parent Sample. . 33 Research Measures . . . . . . . . . . . . . . . 35 Parental Research Variables . . . . . . . . . . 43 Factor Loadings CBPI Data . . . . . . . . . . . 45 Children's Research Variables . . . . . . . . . 47 CRPBI Factor Loadings . . . . . . . . . . . . . 48 Teacher's Research Variables. . . . . . . . . . 48 MANCOVA, Custodial Parent Data. . . . . . . . . 50 ANCOVA's for Interaction, Custodial Parent Data. 0 O O O O O O O O O O I O I O O O O O O O 50 Cell Means and Standard Deviations for Marital Conflict, Dependency, and Duration. . . . . . . 51 MANCOVA, Noncustodial Parent Data . . . . . . . 53 MANCOVA, Children's Data. . . . . . . . . . . . 53 ANCOVA of Children's Research Data for Main Effect of Time Since Divorce. . . . . . . . . . 54 vi TABLE l8. 19. 20. 21. 22. 23. LIST OF TABLES (Continued) Cell Means and Standard Deviations for Relationship with Father and Change in Relationship with Father. . . . . . . MANCOVA, Teacher's Data . . . . . . . Intercorrelations Between Measures of Children's Adjustment and Family Relationship Variables, Custodial Parent Data. . . Intercorrelations Between Measures of Children's Adjustment and Family Relationship Variables, Noncustodial Parent Data . Multiple Regression of Research Variables On Problem Behavior and Competence, Custodial Parent Data 0 O O O O O O I O O O O 0 Multiple Regression of Research Variables On Problem Behavior and Competence, Noncustodial Parent Data. . . . . . . vii Page 55 56 57 58 61 63 INTRODUCTION The number of divorces involving children under age 18 has been steadily rising. In 1969, 1.3 million children under 18 lived in divorced homes; by 1978 the number had increased 340 percent to 4.5 million. It is estimated that 45% of the children born in 1977 wil 1 live in a single- parent home before age 18 (Click, 1979). Historically, research on divorce has focused on its disruptive impact on childrenfis socialization and development (Burgess, 1970; Gasser & Taylor, 1976; Glasser & Navarre, 1965; Herzog & Sudia, 1968). However, recent research with non-clinical populations suggests that most children are able to return to their pre-divorce course of development following a period of crisis caused by the changes in their families (Hetherington, 1979; Hetherington, Cox & Cox, 1978; Luepnitz, 1979; Wallerstein and Kelly, 1980). The possibility that divorce's impact on children may not be detrimental in the long run has lead researchers to reexamine how divorce affects children. It now seems that at least two sets of factors modify children's response to parental divorce. One set relates to individual differences in the child, such as age, sex, and cognitive development (Biller, 1971; Biller & Weiss, 1971; Hetherington & Deur, 1971; Kurdek, Blisk, & Siesky, 1981). The other set of 2 factors relate to changes in family relationships. Variables such as parental conflict, parent-child relation- ships, and patterns of visitation have been shown to influence children's post-divorce adjustment. (Beal, 1979; Kaslow, 1980; Lamb, 1977; Prince-Bonham & Balswick, 1980). The research that has demonstrated that family rela- tionships modify children's response to divorce has been primarily exploratory. Research on children's post-divorce adjustment has not utilized a theory of family organization to identify which aspects of the family structure were important for children's post-divorce adjustment. As a consequence, empirical data on children's adjustment to divorce has rarely been integrated with theoretical concep- tualizations of family structure and organization. This has made it difficult to determine whether children's adjustment and ongoing development is continuous in both divorced and intact families. Research on the impact of divorce has tended to stress the differences in development attributable to divorce, rather than uncovering a set of family relation- ship variables which could account for childrenwsdevelop- ment regardless of parental marital status. Family systems theory provides a theoretical framework which seems able to explain how family relationships affect the ongoing development of children. The theory specifies functions within the family system which are necessary for childrenfls ongoing development. This emphasis on specific child-rearing functions enables the theory to embrace the 3 pattern of family relationships which are critical for children's ongoing development in both divorced and intact families. Family systems theory is able to account for the differences between divorced and intact family systems, and how these differences make children of divorce more vulnerable to problems in their development. The same concepts of family structure and organization which explain children's development in intact families can be applied to identify how the divorced family can continue to meet children's ongoing needs. Divorce has been conceptualized.as a developmental phase a family may enter during the course of its life cycle (Carter & McGoldrick, 1980). This paper will focus on the particular tasks which a family must accomplish during this phase of development in order to ensure the onoing develop- ment of children. The model for the post-divorce family will specify the structures of the family system which are essential for childrenfscmavelopment in both divorce and intact families. The model will also specify the structures of the family system which will cease to be useful following parental divorce. This model for the process of family reorganization will be presented in detail in the first section of the literature reviewu The second section of the literature review will examine empirical literature on childrenfls adjustment to divorce in light of the proposed model. The model's capacity to account for existing 4 findings on childrends post-divorce adjustment will be evaluated in this second section of the literature review. The current research represents an attempt to study empirically the relationship between childrewfls post-divorce adjustment and variables selected on the basis of the pro- posed model for post-divorce family reorganization. A cross-sectional quasi—exprimental design was used to explore specific structures in the post-divorce family system which the the proposed model identifies as important for children's continued development. Data was collected using a variety of self-report questionnaires and checklists with families one to four years following the date of divorce. The study examined the relationship between parents' and teachers perceptions of childrenfs adjustment and the parents' and childrens perceptions of specific parts of the family system. The research attempted to study the rela- tionship between children's post-divorce adjustment and family realtionship variables within a theoretical frame- work which accounts for the impact of family organization on chi ldren's adjustment. REVIEW OF THE LITERATURE A Model for Post-Divorce Family Organization Family systems theorists have argued that certain aspects of family structure are crucial for children's deve- lopment. The most elemental of these is the boundary between the generations (Glick & Kessler, 1974; Haley, 1976). This boundary differentiates the roles and responsi- bilities of the different generations, establishing a power hierarchy within the family. The parental subsystem, which the marital couple develops with the birth of the first child, maintains this boundary. The parental subsystem must nurture, guide, and control children, while protecting the privacy of the marital relationship (Minuchin, 1974). Psychodynamic family theorists have also argued that two competent opposite-sex adults are necessary for children's normal sex-role development (Lidz, Fleck, & Cornelison, 1965; Skynner, 1976). A parent of each sex is important for the successful development of gender identity and resolution of the Oedipal crisis. Structural and strategic family theorists argue on the basis of social learning theory that the presence of two parents is important for children's development (Minuchin, 1967; Satir, 1967). The child learns appropriate social behavior through direct interaction with each parent, and through the observation of the parents 6 interacting with each other. Consequently, according to both psychodynamic, structural, and strategic family theorists, an effective parental subsystem, and access to each parent are important aspects of the family system for childrenis development. When a couple divorces, the entire family system must undergo a massive reorganization. This process of reorgani- zation has been described as a dislocation within the normal family life cycle (Carter & McGoldrick, 1980). The family must address certain emotional issues in order to continue with its ongoing development, as it would during any deve- lopmental phase. As emotional issues are addressed, the family organization and structure changes and develops. Families which accomplish the emotional work of divorce are able to develop new family structures which reflect the emotional work of the family. According to this model, the emotional issues of divorce include the acceptance of the inability to resolve the marital problems, the resolution of the marital attachment, and the willingness to maintain a cooperative parental relationship. Carter and McGoldrick conceptualize the family reorga- nization which accompanies divorce as one of many develop- mental phases a family may face during its life cycle. This is an important idea, permitting comparisons between the emotional problems which accompany divorce and emotional problems which may accompany other transition periods in the family life cycle. However, their model does not 7 differentiate between the emotional issues which parents and children face during the divorce process. In addition, their model does not address how the post-divorce family can continue to meet the ongoing developmental needs of children. When children are involved, the structure of the new family system must accommodate certain childrearing tasks which existed prior to the divorce. The needs of the children for nurturance, guidance and control remain the same. The parental dyad, which maintained the generational boundary and defined the power hierarchy within the family must continue to face the complex needs of growing children. Children also continue to benefit from the unique contribu- tions of their relationships with each parent. The parental and parent-child subsystems must adapt to provide the same functions in a new context. The divorcing couple must differentiate their roles as marital partners and parents, ending one role while continuing the other. This process of change and reorganization is complex, and failure to achieve any step in the process may have serious consequences for children's long-term adjustment. For example, failure to resolve the marital relationship could lead to continued conflict between the couple, providing little relief from the hostility of the marital relationship. Failure to maintain access to each parent could lead to the loss of an important adult model for the children, and potential problems in the development of 8 normal sex-role identity and interpersonal skills. Failure to maintain the parental dyad could lead to one parent being overwhelmed by the burden of single—parenting, leading to decreased quality in parenting for the children. The most serious consequence of failure in the process of family reorganization would be for the child to be fixed in an alliance with one parent against the other--a process which has been called triangulation (Minuchin, 1974). An alliance which joins a member of one generation with a member of another generation against a peer distorts the power hierarchy in the family. When this occurs, children become enmeshed in a conflict which is not their own, and their own developmental needs are neglected (Boszormenyi- Nagy & Sparks, 1973). The consequences of this type of an alliance for the child can be very severe particularly if the alliance becomes a fixed part of the family system (Haley, 1976; Madanes,1981L Research with intact families has shown that the proba- bilities of these unhealthy alliances are greater when the parents are unable to maintain a supportive alliance as parents (Lidz, et al., 1965). In single-parent households, one might expect a greater risk for the development of such alliances, especially if the family system has been unable to resolve the marital conflict (Beal, 1979; Tooley, 1976). Overburdened single—parents, without the emotional support of another adult, may find it very easy to turn to their children to meet their own emotional needs. Both the 9 continuation of the conflicted marital subsystem and the absence of a functioning parental subsystem would seem to increase the risk of such maladaptive alliances. When the divorced family is able to accomplish the process of reorganization, the prognosis for the children should be quite good. Ending the dysfunctional marital relationship should provide relief for the children by decreasing the hostility which had been present in their home. Continued access to each parent should decrease the sense of loss associated with divorce, as well as continue to provide the child with multiple role models. Continua- tion of the parental dyad should prevent either parent from feeling overwhelmed with the demands of parenting, and help to maintain better quality parenting for the children. The accomplishment of these tasks should decrease the amount of stress in the total family system and decrease the risk of inappropriate cross-generational alliances. Children in families which can accomplish these tasks should show few behavioral problems following the initial crisis of the divorce. These families should continue to provide an ade- quate environment for their childrenfls development despite the ending of the marital relationship. Childrends long-term adjustment following parental divorce will depend upon the ability of their families to end the marital subsystem while adapting and maintaining the parental and parent-child subsystems. While research has not explicitly tested this model for post-divorce 10 organization and its impact on children's adjustment, there is a growing body of research on children's post-divorce experience and adjustment. This literature will be reviewed in the following sections, with an emphasis on research which has considered the relationship between children's adjustment and their family's post-divorce organization. Empirical Literature on Children's Post-Divorce Adjustment Longitudinal Studies. Two longitudinal studies of children from divorced families provide most of what is known about the impact of divorce on children and the factors that facilitate their adjustment. The work of Wallerstein and Kelly (1975,1976,1980) and Hetherington, Cox and Cox (1978,1979) is particularly important for two reasons. First, their subjects represented normal popula- tions, not clinical samples, or populations identified as having problems other than those related to the process of divorce. The other strength of these two studies is the fact that they were longitudinal and prospective. They were able to fOIIOW'the course of development in the post-divorce family and identify factors which mediated the impact of the divorce on children. Wallerstein and Kelly (1975,1976,1980) studied 131 children and adolescents from 60 primarily white, middle class families. The children and families were volunteers, and children with histories of psychological problems were eliminated from the sample. The initial data were collected through 4-6 individual interviews with each child and parent 11 over a six-week period. Detailed notes of all sessions were taped, transcribed, and finally coded to permit correla- tional analyses. The subjects were invited to return for follow-up sessions one and five years later, and 58 of the original families continued to participate. At the five year follow-up, the authors determined that 34% of the children showed excellent adjustment, 29% showed adequate but uneven functioning, and 27% showed some problems such as loneliness, depression, or disappointment in a relationship with a parent. The authors concluded that the outcome for the children five years after the divorce was primarily linked to components of family relationships. These in- cluded the extent to which parents had resolved or set aside their old conflicts, the resumption or improvement of parenting since the divorce, and the extent to which the children had maintained a regular relationship with the non- custodial parent. School performance, which was assessed through contact with the children's teachers, seemed parti- cularly sensitive to availability of the father, regardless of the sex of the child. These familial factors interacted with the individual strengths and weaknesses of the parti- cular children, as wel l as their age and sex. However, at the five year follow-up, individual differences contributed less to childrenfsibehavior than family relationship variables. The work of Hetherington and her colleagues (1978, 1979) differs from the work of Wallerstein and Kelly in the 12 use of objective measures and the inclusion of a control group of intact families. There were also a number of differences in the composition of the populations they studied. Hetherington limited her sample to white, mother- custody families. Wallerstein and Kelly included in their sample a number of black (3%) and interracial (9%) families. They also included families with shared or father-custody arrangements, although they do not report an exact number for their sample. The other major difference is that Hetherington's sample was limited to preschoolers, with a mean age of 4 years at the start of her study; Wallerstein & Kelly”s sample included children ranging from age 2 to 18. The subjects in the Hetherington study were 48 divorced families identified through the courts with a first or second child in nursery school, and a control group of intact families matched for age and sex of child, as well as age, education, and length of parent's marriage. The fami- lies were studied at two months, one year, and two years following divorce» Data‘were collected through interviews and structured diaries kept by the parents, observations of the parents and children in the home and laboratory, parent ratings and checklists of children's behavior, and persona- lity measures of the parents. Children were also observed in school, rated by teachers, and rated on measures of sex- role typing, cognitive and social development. Repeated measures MANOVAHs*were performed for all measures, as well 13 as a variety of correlational analyses. At two years following divorce, the girls from divorced and intact fami- lies showed no differences on a wide range of measures. Boys from divorced and intact homes showed few differences, but boys from divorced families still had more problems in relationships with peers and more problems in their rela- tionships with their parents. Cross-lagged correlations showed that more positive adjustment in children at two years was primarily related to aspects of the parental relationship. Agreement in childrearing, a positive atti- tude towards the spouse, low conflict between the parents, and frequent contact with the father were associated with positive mother-child interactions, and positive adjustment in the child. The prognosis for the children was poor when parental conflict continued, or when either parent showed poor individual adjustment. While children from low con- flict intact and divorced homes could not be distinguished at two years following divorce, boys from high conflict divorced homes showed more problems than any other group of children. These two studies, using very different methods con- cluded that children's long-term divorce adjustment was primarily determined by the pattern of post-divorce family relationships. The familial factors these studies identify, the decrease in parental conflict, agreement in parenting, and continued contact with the noncustodial parent, correspond to the proposed model for post-divorce family 14 reorganization. They both suggest that research on children's post-divorce adjustment must consider the functioning of the total family system. There exists one other study of children from divorced homes which included a two-year follow-up (Kurdek, Blisk & Siesky, 1981). The original sample included 58 white, middle class children aged 8-17 whose parents had separated approximately 4 years earlier. The study was primarily concerned with children's understanding and feelings about the divorce. Measures included a variety of questionnaires, as well as measures of locus of control, interpersonal reasoning, and the custodial parent's ratings of children's behavior and adjustment to the divorce. The noncustodial parent was not included in the research. These same measures were administered two years later to 24 children from the original sample. On the whole, the children showed good adjustment at each time, according to their own and their custodial parent's accounts. Their overall adjustment was positively related to their divorce adjustment, both being rated by their custodial parent. Children's adjustment was also related to their feelings about the divorce. Their most negative feelings related to the loss of the noncustodial parent and the changes in family rela- tionships. Some of the findings from this study conflicted with previous research on childrenfis adjustment to divorce. The authors found that positive divorce adjustment was related 15 to infrequent visitation, as well as having less competent parents. The authors suggest that two factors may have contributed to these unexpected findings. The children in this sample were primarily adolescents, and at a point in their development when they would be moving out from their families,and less involved with each of their parents. The other factor which was critical was the fact that data on the childrenwsadjustment and visitation was only collected from the custodial parents. The findings of this study were congruent with other research in several ways. The findings suggest that most children are able to adjust to the parental divorce and not have any long-term problems, according to the account of the custodial parent. In addi- tion, the children reported that the most problematic aspects of the divorce were the loss of the noncustodial parent,and the changes in family relationships. Cross-sectional and Correlational Studies. One of the most carefully executed studies of children'Sjpost-divorce adjustment was conducted by Hess and Camara (1979). The subjects were 16 divorced families with children aged 9-11, and 16 matched control families. The families were iden- tified through court records and participated in the research approximately one and a half to two years after the date of separation. Interviews were conducted with the children, each parent, and the children's teachers to collect data on the children's behavior. Additional data were collected through the use of behavior checklists 16 completed by the parents and teachers. Data analyses included comparisons of the children from divorced and intact homes on stress, aggressiveness, social relations, and work effectiveness as well as evaluations of the relative contribution of family process variables to the child outcome measures. The children from the divorced homes showed more stress and less work effectiveness than the children from the intact group. However, additional analyses demonstrated that these differences were primarily related to family process variables, and not family type. The quality of the parent-child relationships and the degree of parental harmony contributed the most to the variance of the child outcome measures. The authors concluded that the family relationships that emerge after divorce affect children as much or more than the divorce itself. They also stressed that ongoing relationships with each parent were particularly important. The research reviewed to this point has focused on divorced families two to six years after the separation. A study by Jacobson (1978a, 1978b, 1978c) investigated children's adjustment during the first year after separa- tion. The sample consisted of 30 families with 51 children ranging in age from 3-17. The parents had been separated at least one week but no more than 12 months. Data were collected through structured interviews with the custodial parent and children, and childrenfs adjustment was measured with the Louisville Behavior Checklist. During the first 17 year of separation, the factor which contributed the most to children's adjustment was the attention parents gave to the children to help them deal with the divorce. Unfortunately, most parents were so overwhelmed by their own needs that they were unable to attend to their children's needs. This is reminiscent of Hetheringtonis (1979) finding that the first year after divorce is a time of decreased quality in parenting and deteriorating parent-child relations. The next most important factor was loss of time with the father, and then interparent hostility. Children who received help from their parents in dealing with the divorce, who main- tained contact with their fathers, and who experienced a decrease in parental hostility, showed the best adjustment in this sample. These findings suggest that even in the first year following the separation, changes in the family system can lessen as well as heighten the detrimental impact of the divorce on children. One aspect of the post-divorce family system which has only recently been studied is the impact of the continuation of the marital subsystem on childrenis adjustment. Few studies have considered the possibility of the marital rela- tionship continuing after the divorce. Studies with intact families have demonstrated that parental conflict is asso- ciated with problems in children (Baruch, 1944; Johnson & Lobitz,1974; Leighton, Stollak, & Ferguson, 1971; Porter & (YLeary, 1980). Parental agreement has also been shown to foster healthy psychological development in children (Block, 18 Block, & Morrison, 1981). A number of studies have compared the adjustment of children from intact but unhappy or con- flicted homes with that of children from divorced homes. These studies have all found that the impact of an intact but conflict-ridden home is more harmful to children than the impact of divorce (Berg & Kel ly, 1979; McCord, McCord 5. Thurber, 1962; Nye 1957; Rutter, 1971; Whitehead, 1979). However, none of these studies considered the possibility of continued marital conflict within the divorced homes. One of the first reports to deal with the failure of divorce to end the marital relationship and the subsequent impact on children is a study by Westman and Cline (1971). The authors reviewed 105 consecutive divorce cases that involved children occurring during a two year period in one county in Wisconsin which involved children. Thirty-one percent of the cases were involved in 2-10 court actions in the two years following the divorce. The authors felt that while most couples did resolve their relationships through the divorce, one third did not and continued to fight through the legal system. The same authors also reviewed 153 consecutive admissions to a child guidance clinic, and found that 23 cases involved parental divorce. All of these guidance clinic cases showed evidence of turbulent interac- tion between the divorced couple or total loss of contact with one parent. None of these cases demonstrated a mutually satisfactory resolution of the marital relation- ship. Failure to resolve the marital relationship had a strong 19 affect on the families' interactions, and negative conse- quences for the children. While this study was lacking in statistical sophistication, it was important for bringing attention to the idea that divorce may not lead to the resolution of the marital relationship. Several studies have now investigated ongoing marital conflict in divorced families. Raschke and Raschke (1979) studied family conflict and its relationship to self-esteem with a sample of 259 children in grades three, six, and eight. Data on family type and family conflict were collected using questionnaires completed by the children. Children's self-concept as measured by the Piers-Harris self-concept scale was the outcome measure. While the authors found no relationship between parents' marital status and children's self-concept, self-concept was signi- ficantly related to children's perception of fighting in the family. Unfortunately, the researchers' questions con- cerning family conflict were vague, so that conflict in a single—parent home may have represented parental conflict, conflict with other adults in the household, or parent-child conflict. A more carefully designed study was conducted by Parrish, Dostal, and Parrish (1981). The self-evaluations of children from "happy" and "unhappy" divorced and intact homes served as the dependent variables. The judgement of a home as "happy" or "unhappy" was based on the childrenfs self-report. The subjects were a sample of 284 fifth 20 through eighth grade children who had completed evaluations of themselves, their homes,and their parents. There was no main effect for family type on children's self evaluations. However, boys from unhappy divorced homes had lower self- concepts than all other groups, while females from intact homes had higher self-evaluations than all other groups. Children from divorced homes had more negative evaluations of their parents than controls from intact unhappy homes. These studies suggest that one major task which must be accomplished following parental divorce is the resolution of the marital relationship. When this relationship continues, the marital conflict continues and the consequences for the children are negative. When the divorced family is able to resolve the marital conflict, the children are free from the stress of the hostility. Children from both divorced and intact homes that are free from marital conflict show similar behavior. However, the legal act of divorce and the resolution of the marital relationship are two different events which must be distinguished. This distinction is one which has been largely overlooked in research on divorce. Studies Comparing Different Custody Arrangements. Some research has investigated the impact of different custodial arrangements on children's post-divorce adjustment. One problem with this approach is that it confuses legal arrangements with the psychological experience of the family. There can be great variation in the amount of contact between family members in any custody arrangement. 21 Some chidren in single-custody arrangements have more contact with each parent than some children in joint-custody arrangements. The legal definition of custody has no clear correspondence to the amount of time with each parent, or the psychological importance of that contact. Another problem with this approach is that it tends to treat divorce as a specific, isolated, time limited event, rather than as a complex process which unfolds through time (Hetherington, 1979; Wiseman, 1975). Despite these limitations, much of the research on children and divorce has been conducted in this way and must be considered. A number of studies have been concerned with the general impact of various custody arrangements on children's overall adjustment to divorce. Santrock and Warshak (1979) studied social development in 60 boys and girls aged 6-11 from father-custody, mother-custody, and intact families. The parents in the divorced families had been separated an average of 2.9 years. Data were collected by means of interviews, self-report scales, projective tasks, and obser- vations in the laboratory. They found that father-custody boys were more socially competent than father-custody girls; mother-custody girls were more socially competent than mother-custody boys. The boys in father-custody homes had higher self-esteem than intact family boys, and boys in mother-custody homes had less anxiety than intact family boys. Father-custody girls, however, showed less social competence than intact family girls,and mother-custody girls 22 showed increased anxiety. The authors concluded that children, especially boys, benefited from living with their same-sex parent following parental divorce. A study by Stephens and Day (1979) also attempted to investigate the impact of mother or father custody on girls from divorced families. Their subjects were 23 girls ranging in age from 2-23 years, whose parents had been separated an average of 11.7 years. They found no differences between the two divorced groups or a control group on measures of self-concept or sex-role identity. However, it is hard to make sense of their results given the large age range of their sample, their small sample size, and their failure to control for the age when the separation had occurred. A problem with each of these studies is their failure to obtain data on children's relationships with the non- custodial parent. This critical variable was included in a study by Lowenstein and Koopman (1978) which investigated mother and father custody. Their subjects were 40 single parent mothers and fathers with sons aged 9-14. All the parents had had custody of their children at least one year. Data were collected by means of self-report instruments and questionnaires administered in the subjects' homes. There was no main effect for sex of the custodial parent on the boy's self-esteem. However, boys who saw their noncustodial parent once a month or more had higher self-esteem than those who saw their noncustodial parent less frequently. The 23 sex of the custodial parent was less important than having continued contact with both parents. There is also one report on joint custody and its affect on children's post-divorce adjustment. Abarbanel (1979) studied four families that had agreed to joint- custody arrangements using a case study approach. She included no objective measures in her research, but did interview all the family members involved and observed the children in each of their homes. She concluded that the commitment to share responsiblity helped all members of the family, both children and adults. The children were able to continue their relationships with each parent and showed less loss than is typical following parental divorce. She concluded that joint custody had benefited the children because they did not experience the sense of loss of a parent, and because each parent felt less burdened and more able to provide quality parenting. Father-Absence Studies. A large amount of research has been devoted to studying the impact of one variable, "father-absence", on one aspect of childrenfls development-- their sex-role development. Thorough critiques of this literature have been presented by Herzog and Sudia (1968) and Sprey (1967). In the following section, only those "father-absence" studies which studied children from divorced homes will be reviewed. A number of father-absence studies with boys from divorced families have demonstrated that boys whose parents 24 divorce before they were age six have a less masculine identity as measured by projective tests of sex-role orien- tation (Biller, 1969; Hetherington, 1966), exhibition of aggression in doll play (Santrock, 1970), and exhibition of aggression in recreational activities (Hetherington, 1966). Father-absence, however, even at a young age has not been shown to affect sexual preference or sex-role adoption (Biller, 1971). Studies with boys who were past age 6, or whose parents divorced after they were age 6, have not demonstrated any deficits in sex-role identity (McCord, et al.,1962; Santrock, 1977). One of the most carefully executed father-absence studies was completed by Hetherington (1972) and studied the impact of the time and reason for father-absence on adolescent girlsfl personality development. The subjects were lower-middle class white girls aged 13-17 who attended a community recreation center, from divorced, widowed, or intact families. In the divorced group, only girls with minimal father contact were studied, and none of the girls from the widowed or divorced homes had male siblings. Data on the girls included observational measures of the girls at the center, measures of nonverbal behavior during the inter- view, ratings based on individual interviews with the girls and their mothers,and scores on a number of personality measures. In general, there were few deviations in the measures of sex-role typing. However, disruptions were evident in the girls interactions with male interviewers. 25 Girls from divorced homes exhibited open and responsive non- verbal behavior as well as proximity and attention seeking behavior. The girls from widowed homes exhibited inhibi- tion, avoidance, and restraint with males. For each father- absence group, earlier father absence was associated with more severe effects. Extreme father absence at an early age does have consequences for adolescent girls' heterosexual behavior. What seems more remarkable is how few differences there actually were between the groups, which were specifi- cally selected to emphasize extreme father-absence. Studies of father-absence have illustrated how one particular aspect of childrenfls develOpment, their sex-role identity, may be adversely affected by parental divorce. This is most likely to occur if the divorce is before the child is age six, and if the paternal absence is extreme. The effects may emerge at different ages depending upon the sex of the child. In general, studies of father-absence have found few differences in sex-role development in children from divorced homes. One reason for this has been their failure to control for the amount of contact with the father. In addition, the impact of a father's absence may be both direct, in terms of the decreased time the child has with the father, and indirect, in terms of the ongoing relationship between the mother and father and changes in the mother-child relationship. Finally, the impact of father-absence may be more general and less specific than simply affecting the child's sex-role development. The 26 research on children's responses to divorce suggest that the major problems include general behavioral problems, depression, loneliness, and school problems (Hetherington, et al., 1978; Wallerstein and Kelly, 1980). As a research strategy for understanding the impact of divorce on children, focusing on "father-absence" seems too simplistic to capture the magnitude and complexity of the changes which a divorce produces in a child's total world. Although research on divorced families has demonstrated that the ongoing family system plays a major role in deter- mining children's adjustment, a model for the process of family reorganization has not been applied in an emprical study. Family systems theory suggests that the divorced family must accomplish the tasks of transforming the marital relationship, while adapting and maintaining the parental and parent-child subsystems. Children's adjustment in divorced families will be determined to a large degree by the family's ability to accomplish these changes. Although almost all children display behavioral problems immediately following parental divorce, most return to their expected patterns of development as the new family system emerges and stabilizes. When the marital relationship is able to resolve its conflicts, when the parental and parent-child subsystems are able to adapt to meet the changing needs of the family members, the children are able to continue their lives with little risk of ongoing emotional problems. The study presented in the next chapter is an attempt to test 27 empirically a theoretical model for post-divorce family organization which seems able to account for findings from a wide range of prior studies on children's adjustment, as well as integrating reseach on children'scdivorce adjustment into a larger theoretical framework. METHOD Subjects The subjects were obtained with the assistance of the Friend of the Court of Ingham County, Michigan, who permitted the researcher to identify all couples with children aged 5-11 who had obtained a divorce in Ingham County during the years 1979-1981. Letters were sent to 631 custodial parents requesting that they return an enclosed postcard indicating if they were interested in participating in research on the impact of divorce on children and families. See Appendix A for a copy of the initial contact letter. The seventy-six parents who responded positively were individually contacted by telephone or personal letter and the research procedure explained to them in more detail. Of these 76 custodial parents, eight did not have telephones; although several strategies were utilized to contact these parents, the project was not successful in collecting data from this group. Five parents changed their minds about participating when they were contacted by the researcher, or when the research assistant tried to set up a time to collect the data. Two of the parents were lost because they moved or lived too far away for the data to be collected expediently. 28 29 Table 1. Custodial Parent Socioeconomic Status Class Lower Middle Upper Unknown Number 15 9 26 6 2 3 Percentage 24.6 14.8 42.6 9.8 3.3 4.9 Individual appointments were arranged to collect the data at the homes of the 61 remaining parents who were interested in participating in the research with their children. The participating custodial parents had a mean age of 33.0 years (SD = 5.1 years), and 57 percent had completed some college education or more. The parents had been separated an average of 3.5 years (SD = 1.5 years) and had been divorced an average of 2.6 years (SD = 1.1 years). Socioeconomic status was assessed using Hollingshead's Two Factor Index of Social Position (1965). The custodial parents were primarily middle class or lower; 82% were in Hollingshead's lower three classes. Ninety-five percent of the custodial parents were women and 20% of the custodial parents had remarried. The nature of the relationship with the Friend of the Court prohibited a comparison of the parents who agreed to participate with those who did not. Table 1 contains a summary of the custodial parent socio- economic data. 30 The noncustodial parents were contacted after obtaining verbal consent from the custodial parent. Individual telephone calls or personal letters were sent to the non- custodial parents, requesting their participation in the research project. If the noncustodial parents agreed to participate, an appointment was set to collect the data at their home at a time that was convenient for them. Although efforts were made to contact the former spouses of the 61 participating custodial parents, only 38 noncustodial parents participated in the study. Eight of the noncustodial parents did not have telephones and either did not respond to the initial contact letter or failed to respond to additional attempts to contact them; 10 were unwil ling to participate; 3 had moved out of state; 2 said they would participate but all efforts to arrange an appointment for the data collection failed. The mean age of the participating noncustodial parents was 34.4 years (SD = 5.1 years), and 58 percent had completed some college education or more. According to Hollingshead's Index of Social Position, the noncustodial parents were primarily middle class or lower; 86.8% were in Hollingshead's lower three classes. Table 2 contains the noncustodial parent socioeconomic data. Ninety-four percent of the noncustodial parents were men and 28 percent had remarried. The noncustodial parents had been separated an average of 3.5 years (SD = 1.6 years) and had been divorced an average of 2.1 years (SD = .9 years). 31 Table 2. Noncustodial Parent Socioeconomic Status Class Lower Middle Upper Total Number 7 10 16 3 2 38 Percentage 18.4 26.3 42.7 7.9 5.3 100 The compositions of the custodial and noncustodial parent groups were very similar with respect to the demographic variables. The age and educational levels of each group were nearly identical. There was a larger percentage of custodial parents in the lowest socioeconomic class. However, the distribution of participants in the lowest three classes was very similar, with a cumulative frequency of 82 and 86 percent for the custodial and noncustodial parents, respectively. Both groups had been separated the same number of years. The largest difference between the groups was the prOportion that had remarried- While 20% of the custodial parents had remarried, 28% of the noncustodial parents had remarried by the time they participated in this study. Data were collected from 61 children. Their parents had been separated an average of 3.5 years (SD = 1.5 years) and had been divorced an average of 2.6 years (SD = 1.1 years). The children, at the time of the testing, ranged in 32 Table 3. Age and Sex of Children Age SEX 6 7 8 9 10 11 12 TOTAL Male 2 3 6 15 4 6 1 37 Female 5 2 8 1 5 3 0 24 TOTAL 7 5 14 16 9 9 1 61 age from 6-12 years (M =8.8 years, SD = 1.6 years); there were 37 boys, and 24 girls in the sample. Table 3 provides a breakdown of the children by age and sex. Individual letters were sent to teachers of each of the 61 children in the study after signed consent was obtained from the custodial parent. Two teachers refused to partici- pate as a result of personal beliefs they held regarding research. Four teachers simply failed to return the research questionnaire. Data was collected from 56 teachers. The initial letter to the teachers is presented in Appendix B. Design The study used a cross-sectional, quasi-experimental design (Campbell & Stanley, 1963L. The naturally occuring variables of time since the divorce and children's sex were used to approximate a two-factor experimental design. Time since divorce was selected as a factor over time since separation because of the organization of the files at the 33 Table 4. Distribution of children by time since divorce and sex of child, custodial parent sample Time Since Divorce Sex of Child 1 2 3 4 Total Male 7 5 10 10 36 Female 4 10 6 4 24 Total 11 18 16 16 56 Note. Missing data reduced the final n to 56. Table 5. Distribution of children by time since divorce and sex of child, noncustodial parent sample Time Since Divorce Sex of Child 1 2 3 4 Total Male 5 4 7 5 21 Female 3 5 4 3 15 Total 8 9 11 8 36 Note. Missing data reduced the final 2 to 36. Friend of the Court during the initial identification of research families. Time since divorce was calculated by subtracting the date of the divorce decree from the time of the data collection. There were 4 groups for time since the divorce, 12 months or less, 13-24 months, 25-36 months, and 34 37-48 months. Tables 4 and 5 contain the distribution of the children by time since divorce and sex of child for each group of parents. Since the purpose of this research was to explore process of change within the family system of divorced families, no control group was used. Studies of divorced families have utilized both intact families and widowed families as control groups. The inclusion of these control groups has been crucial when the question of concern was how divorced families are different from or similar to other families (Altus, 1959; Felner, Farber, Ginter, Borke & Cowen, 1980; Hetherington, 1972; Tuckman & Regan, 1966). However, the use of intact or widowed families as controls makes less sense when the research specifically explores the process which unfolds within families over time as a conse- quence of divorce. Since this research was not concerned with generalizing the results to other types of families, the inclusion of a control group seemed less critical. Procedure The research instruments were individually administered to the participants at their homes. When the data were collected from the custodial parent and child, a research assistant met with them together to answer any questions they had and to obtain their written consent. The parent was then handed a packet containing the parental research instruments. While the parent worked independently, the research assistant worked individually with the child in a 35 Table 6. Research Measures Respondent Measure Children Child Questionnaire Child Report of Parental Behavior Inventory (CRPBI) Parents Parent Questionnaire O'Leary-Porter Scale (O'Leary) Child-Rearing Practices Report (CRPR) Child Behavior with Parent Inventory (CBPI) Child Behavior Checklist (CBCL) Teachers Teacher's Report Form separate room or part of the house. Every effort was made to collect the data privately, in a room away from other family members. When data were collected from the noncustodial parent, the research assistant again obtained consent from the parent, answered any questions, and handed the parent the research instruments to complete on their own. Most children completed the research questionnaires in less than an hour, and all of them completed the question- naires in less than an hour and a half. The parents generally completed their instruments in less than one and a half hours, and all of the parents completed their instru- ments in less than two and a half hours. Table 6 contains a summary of the research instruments which will be described below. 36 Instruments Child Questionnaire. This was an 8 item questionnaire designed to measure childrenfs global perceptions of the marital, parental and parent-child relationships, as well as how these had changed since their parents' divorce. The child rated each item on a five point scale which was repre- sented graphical ly for them on a card with a series of faces ranging from sad to happy. A few trial questions were asked to be sure that the child understood the response procedure. The research assistant then read each questionnaire item and marked down the child's responses. This instrument provided variables measuring the child's perceptions of the marital, parental, and parent-child subsystems. A copy of this instrument and its response card is in Appendix C. Child's Report gf Parental Behavior Inventory (CRPBI). The 108 item version (Schluderman & Schluderman, 1970) of Schaefer's CRPBI was selected to measure children's percep- tions of their relationships with each parent. Factor- analytic studies of the CRPBI have identified three factors which it measures, Acceptance-Rejection, Psychological Autonomy-Control, and Firm Control-Lax Control (Goldin, 1965; Schaefer, 1965a; Schaefer & Bayley, 1967). The CRPBI was designed as a self-report questionnaire for children in adolescence or older. For each item, the subject rates whether a statement is like, somewhat like, or not like the identified parent. This procedure had to be modified since the children in this study were too young to 37 read this instrument on their own. The questions were read to the children, who responded by pointing to pictures which represented whether the statement was like, somewhat like, or not like the identified parent. A copy of the CRPBI and the response card is in Appendix D. The research assistant trained the child with a few practice questions before administering the actual questionnaire items. The children completed the instrument twice, once for the custodial and once for the noncustodial parent. The factor scores from the CRPBI provided measures of the parent-child relationship. Parent Qpestionnaire. This questionnaire is designed to measure the parents'