MlCHlGAN STATE UNWERS! SlTY LIB ll In.” Ml“.till.”.l'lHIMl 1293 00913 2014 l! This is to certify that the dissertation entitled INDIVIDUAL, FAMILY AND CHILD-REARING CHARACTERISTICS IN YOUNG ALCOHOLIC AND NONALCOHOLIC FAMILIES presented by W. HOBART DAVIES has been accepted towards fulfillment of the requirements for Ph .1) . degree in PSYCHOLOGY / Date MARCH 13L 1992 MS U i: an Affirmative Action/Equal Opportunity Institution 0-12771 L; LIBRARY Michigan State l University \_.__ ‘ fi— PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES mum on or before date due. 1 DATE DUE DATE DUE DATE DUE l Em %EL] 1 LL JWE l MSU le An Affirmative Action/Equal Opportunity Institution cha-pt INDIVIDUAL, FAMILY AND CHILD-REARING CHARACTERISTICS IN YOUNG ALCOHOLIC AND NONALCOHOLIC FAMILIES BY W. Hobart Davies A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Psychology 1992 / .C' Os.) ". (I .7 /,, ABSTRACT INDIVIDUAL, FAMILY AND CHILD-REARING CHARACTERISTICS IN YOUNG ALCOHOLIC AND NONALCOHOLIC FAMILIES By W. Hobart Davies Examined differences in demography, parental psychopathology, family environment, child behavior problems, and socialization attitudes between three groups of families: (1) neither-parent-alcoholic (9-32); (2) father-only-alcoholic (3-71); and (3) both father and mother alcoholic (n—S9). Effects of the level of alcohol dependence were also investigated. Alcoholic fathers were a population based sample of convicted drunk drivers identified through district courts. They had typically not been in treatment as yet. Control families were located via neighborhood canvassing in the same census tracts as the alcoholic families. Each family had a male child between 3.0 and 6.0 years and was intact at the time of recruitment. Compared to nonalcoholic families from the same neighborhoods, (a) families with alcoholic fathers were found to have higher levels of paternal psychopathology (both drinking-specific and nondrinking-specific), fewer resources, less parental achievement, and a less supportive family environment; and (b) families with two alcoholic parents were found to have higher levels of parental psychopathology (both mother and father), family conflict, and child behavioral difficulty. Among the alcoholic families, severely alcohol dependent fathers reported greater levels of psychopathology, child behavior difficulties, and family conflict and crises. Maternal alcoholism severity was associated with more antisociality, increased child behavior problems, and a less supportive environment. There was considerable evidence that parents' reported socialization attitudes were stable and consistent, and that they comprise a coherent cognitive schema. Consistent differences in socialization attitudes were found only in relation to socioeconomic status (SES) and its associated demographic factors. Lower SES was associated with more authoritarianism, less paternal nurturance, and greater parental disagreement about child-rearing. Socialization attitudes were essentially unrelated to the family's diagnostic classification. More work is needed in understanding the connections between socialization attitudes and parent-child interaction. The present results clearly indicate that young alcoholic families show pervasive trouble that would be expected to place the children growing up in these homes at risk, especially when both parents are alcoholic and/or the severity of alcoholism is greater. To Cheryl, Now and Always, and in Memory of my Grandfathers, John Frank Davies and Herbert L. Hobart "Here the heart May give a useful lesson to the head, And Learning wiser grow without his books." -William Cowper iv ACKNOWLEDGEMENTS I am indebted to Robert Zucker for years of guidance and collaboration. As my graduate advisor and director of the MSU Family Study, he has greatly influenced my ways of thinking and approach to psychology. His challenging theoretical formulations served as a constant reminder of the complexity of families and individuals. His support and willingness to follow the path of my interests and goals were essential to completing this project. My other committee members, Hiram Fitzgerald, Frank Floyd, and Ellen Strommen, were consistently helpful in their comments and supportive in their actions. Frank Floyd deserves further comment as a fixture throughout my graduate career. He was always the perfect committee member, and his influence has impacted all of my work. I also need to recognize the sustained influence of Robert Noll, first as research and clinical supervisor, and more recently as friend and collaborator. He remained a committee member in spirit after his departure from MSU and was very helpful with his observations from afar. Susan Refior deserves recognition for this research, as well as all work emanating from the Family Study. Not only does she face the daunting task of locating, recruiting, and retaining families for the project, but she also vigilantly reminded me (often as a solo voice) that this research is about families' lives, not about numbers, questionnaires, or data bases. The data collectors of the project at all levels deserve credit for the immense amount of work they do, often under trying circumstances. And all of the families in the project have my continuing admiration for what they put up with in order to make this work possible. I reserve my greatest thanks for my wife, Cheryl, whose love, patience, and encouragement exceed the bounds of my understanding. As thoroughly demented as the graduate school process is from the inside, the experience of the significant other can only be compared to watching an extended and brutal cockfight which one is not allowed to leave. This work was supported in part by grants from the National Institute on Alcohol Abuse and Alcoholism (AA 07065) and from the Michigan Department of Mental Health, Prevention Services Unit. vi TABLE OF CONTENTS List of Tables . CHAPTER I: Introduction and Review of the Literature The Alcoholic Family Outcomes for Children from Alcoholic Families . Parent-Child Relationships among Groups at High Risk for Alcoholism Families with Alcoholic Parents . Families of Adolescent Problem Drinkers . Families of Later-To-Be Alcoholics Families of Antisocial Adolescents Families with Depressed Parents . Effects of Socioeconomic Status . General vs. Specific Effects of Parental Psychopathology Statement of the Problem Hypotheses CHAPTER II: Method Subjects Procedure . Measures The Parent-Child Relationship . Demographic Variables . Parental Psychopathology vii xi 21 26 28 29 34 38 40 42 44 51 52 52 S8 S9 Alcohol Use/Problems Depression Antisocial Behavior . Overall Psychosocial Functioning Child Behavior Family Environment Data Analytic Strategy CHAPTER III: Results Missing Data and Outliers . Descriptive Statistics and Reliabilities Demographic, Individual Parent and Child, and Perceived Family Functioning Variables . Intercorrelations . Group Comparisons as a Function of Number of Alcoholic Parents . Group Comparisons as a Function of Severity of Parental Alcoholism Summary . Socialization Variables . Intercorrelations . Longitudinal Stability of CRPR Measures . Relationships of CRPR Measures to Reported Parent-to-Child Aggression Correlations with Independent Variables . Group Comparisons . Summary . viii 59 6O 62 63 63 63 65 67 67 67 7O 84 91 101 106 106 108 111 112 117 120 Further Explorations Extreme Groups on Variables other than Alcoholism . Group Comparisons on Individual CRPR Items CHAPTER IV: Discussion The Sample in Perspective . Family Members and the Family System in Young Alcoholic Families Demographic Characteristics . Paternal Psychopathology Maternal Psychopathology Child Behavior Family Environment Extent of Familial Alcoholism and Level of Cumulative Risk The Parent—Child Relationship . The Place of Socialization Attitudes in the Parent-Child Relationship Socialization Attitudes and Parental Psychopathology Socialization Attitudes and Family Characteristics Socialization Attitudes and Demographic Characteristics . Socialization Schemas, Family Characteristics, and Parental Psychopathology Limitations of the Current Study Future Directions . Conclusions . List of References . ix 123 123 124 126 127 127 128 129 131 132 133 135 135 138 139 140 141 143 144 145 . 149 10. ll. 12. l3. 14. LIST OF TABLES . Sociodemographic Characteristics of the Sample . Descriptive Statistics and Internal Consistency (Coefficient alpha) of Individual, Family, and Socialization Variables for Mothers . Descriptive Statistics and Internal Consistency (Coefficient alpha) of Individual, Family, and Socialization Variables for Fathers . Intercorrelations (Pearson R) between Demographic Variables . Intercorrelations (Pearson R) between Parental Psychopathology Variables . Correlations (Pearson R) between Mother and Father Psychopathology Variables . Intercorrelations (Pearson R) between Parental Ratings of Family and Child Functioning . Correlations (Pearson R) between Mother and Father Ratings of Family and Child Functioning . Correlations (Pearson R) between Demographic Variables and Mothers' Psychopathology and Family Functioning Variables Correlations (Pearson R) between Demographic Variables and Fathers' Psychopathology and Family Functioning Variables Correlations (Pearson R) between Mothers' Ratings of Family and Child Functioning and Maternal Psychopathology Correlations (Pearson R) between Fathers' Ratings of Family and Child Functioning and Paternal Psychopathology Mean Scores (and Standard Deviations) of Demographic Variables by Number of Alcoholic Parents Mean Scores (and Standard Deviations) of Maternal Psychopathology Variables by Number of Alcoholic Parents 48 68 69 71 72 74 76 77 79 80 82 83 85 87 15. 16. l7. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. Mean Scores (and Standard Deviations) of Paternal Psychopathology Variables by Number of Alcoholic Parents Mean Scores (and Standard Deviations) of Maternal Ratings of Family and Child Functioning by Number of Alcoholic Parents Mean Scores (and Standard Deviations) of Paternal Ratings of Family and Child Functioning by Number of Alcoholic Parents Mean Scores (and Standard Deviations) of Demographic Variables by Mother Diagnosis Level Mean Scores (and Standard Deviations) of Maternal Psychopathology Variables by Mother Diagnosis Level Mean Scores (and Standard Deviations) of Maternal Ratings of Family and Child Functioning by Mother Diagnosis Level Mean Scores (and Standard Deviations) of Demographic Variables by Father Diagnosis Level Mean Scores (and Standard Deviations) of Paternal Psychopathology Variables by Father Diagnosis Level Mean Scores (and Standard Deviations) of Paternal Ratings of Family and Child Functioning by Father Diagnosis Level Summary Table of Increases in Parent and Family Difficulties and Risks as a Function of Number of Alcoholic Parents and Severity of Parental Alcoholism Intercorrelations (Pearson R) between Socialization Variables Test-Retest Reliability (Pearson R) for Socialization Variables and Internal Consistencies (Coefficient Alpha) at Second Administration Correlations (Pearson R) between CRPR Socialization Variables and Specific Reports of Harshness of Discipline of Both Parents on the Conflict Tactics Scale Correlations (Pearson R) between Socialization Variables and Demographic Variables Correlations (Pearson R) between Socialization Variables and Mother and Father Psychopathology Variables Correlations (Pearson R) between Socialization Variables and Mother and Father Ratings of Family and Child Functioning Mean Scores (and Standard Deviations) of Socialization Variables by Number of Alcoholic Parents xi 88 9O 92 93 94 96 98 99 102 103 107 109 113 114 116 118 119 32. Mean Scores (and Standard Deviations) of Socialization Variables by Mother Diagnosis Level 121 33. Mean Scores (and Standard Deviations) of Socialization Variables by Father Diagnosis Level 122 xii CHAPTER I Introduction and Review of the Literature Children of parents who have alcoholism have consistently been found to be at increased risk for the development of psychopathology and other life problems (Searles & Windle, 1990; West & Prinz, 1987). However, little is known about the mechanisms through which this effect operates, or about the other child and family variables which serve as mediators or contributors to the process. Certainly the parent-child relationship is one important area which impacts the development of children in alcoholic homes.' Also important, however, are other characteristics of the family environment, including individual qualities of the parents and child. These broader characteristics likely have both direct effects on child development and indirect effects, through their impact on the parent-child relationship and other interactive processes. This cross-sectional study examines demographic characteristics, parental psychopathology (both alcohol-related and otherwise), family environments, child behavior problems, and self-reported child-rearing patterns in families with an alcoholic father compared to community control families. A further distinction is made between families with one alcoholic parent (the father) versus two alcoholic parents. The subjects for this study are drawn from a population of initially intact alcoholic and control families with male preschool-aged children. The results of this investigation will provide information about the effects of parental alcoholism on the early child-rearing environment. W The term ”alcoholic family" refers to family systems in which one or more family members are alcoholic (Steinglass, Bennett, Wolin, & Reiss, 1987). The application of the label to the whole system is in recognition of the degree to which the behaviors and interactions of all family members "become organized around, or distorted by, the presence of alcoholism” (Steinglass et al., 1987, p. 46). Interest in the system level variables at work in the family with an alcoholic parent or parents began to emerge in the 1950's and 1960's (Ablon, 1976). However, the study of these variables using adequate methodological technology is still very recent and the picture is far from complete. It has become customary in recent discussions of family effects of alcoholism to consider the developmental stage of the child and/or the family, and such will be the practice here. The clearest finding concerning the family environment has been that there is a higher level of conflict in alcoholic families compared to non-alcoholic families (Clair & Genest, 1987; Zucker, Weil, Baxter, & Noll, 1984). Alcoholic families have also been found to be less cohesive, less organized, and less oriented towards intellectual- cultural and moral-religious pursuits than families with no alcoholic parents (Clair & Genest, 1987; Zucker et al., 1984). The Clair and Genest data are based on retrospective accounts of 18 to 23 year old subjects with an alcoholic father and a non-alcoholic mother. The Zucker and colleagues data are drawn from the current project and are based on families with at least one child under six years. Despite a number of clinical reports and data from unrepresentative samples, methodological inadequacies prevent a 3 definitive answer as to whether heightened level of conflict is associated with increased interpersonal violence in alcoholic families (Reider, Zucker, Noll, Maguin, & Fitzgerald, 1988). Reider and colleagues have identified specific parent life troubles and family characteristics that are related to increased severity and intensity of spousal violence (Reider et al., 1988) and parent-child violence (Reider, Zucker, Maguin, Noll, & Fitzgerald, 1989) within the population of young alcoholic families. The systems level issue is that increased levels of conflict in alcoholic families create an atmosphere of tension which typically affects everyone in the home. Even young children have remarked at the contrast between their own home and the homes of their friends (Cork, 1969; Wilson & Orford, 1978). Clair and Genest (1987) suggest that having a parent who is an alcoholic can be conceptualized as a form of chronic stress. The family is often motivated to keep the parent's alcoholism a secret from people outside the family and it is frequently a taboo subject within the family as well (Ablon, 1976). This can result in patterns of communication which are based on lies and denials, a pattern which is reportedly learned by many children of alcoholics (Barnes, 1977). Ou comes or Child en ro Alcoho i ies It has become an accepted clinical "fact" that children of alcoholics are at increased risk for the development of a variety of life problems in childhood and adulthood. This consensus has emerged largely from clinical reports from social service agencies, and from much-publicized accounts of adults who grew up in such homes, rather 4 than from controlled research (Gomberg, 1989; Searles & Windle, 1990). More recently, methodologically adequate research has begun to fill in the picture, although it is far from completely understood. Children of alcoholics have been found to have elevated rates of alcoholism as adults, although the magnitude of the ratio varies from study to study (Russell, 1990). The highest risk (ninefold increase over controls) was found among adopted sons of severely alcoholic fathers with an early onset of serious criminality (Cloninger, Bohman, & Sigvardsson, 1981). The commonly cited figure of a fourfold increase in risk has been found in several studies not selected for severity or antisocial involvement (Cotton, 1979; Goodwin, Schulsinger, Hermansen, Guze, & Winokur, 1973) while even lower ratios have been reported in some samples (e.g., Drake & Vaillant, 1988; Midanik, 1983). Much of the variability here appears to be due to sample characteristics such as age, treatment status, antisocial behavior, and the pattern or type of alcoholism (see e.g., Zucker, 1987). Russell (1990) concludes that the commonly cited 4:1 ratio may be an overestimate for the entire population of children of alcoholics, but cautions that any number is likely to be an overestimate for certain subtypes of alcoholics and an underestimate for others. It follows that the degree of risk is due largely to the genetic component of the particular subtype of alcoholism being examined. The question of genetic versus psychological transmission (e.g., social learning) remains unresolved, and is probably an unanswerable and unimportant question when framed that way. Evidence exists for influence transmitted by both types of mechanisms and both appear to be at work (Cloninger & Reich, 1983; Goodwin, 1979; Peele, 1986; Zucker & 5 Gomberg, 1986), although the pathways of effect for genetic influences are much less well understood. With regard to familial nongenetic influences upon outcome, Zucker (1979) has proposed that there are three subcategories of influence through which parental behaviors can impact the drinking behavior of their children. Eigag are those pagaagai behaviors which serve as modeling alternatives to the child. Saaaag, the pagentai :ewara stgactuzaa that are available and used to shape the child according to the standards and values of the family. And finally, the quality and patterns of family affectional interaction. This schema also posits that there are both drinking specific and drinking non- specific factors that can influence the child's drinking behavior. 0n the one hand, any of these influences may be directly connected to the parent's use of alcohol, attitudes about alcohol, reward or punishment for use that are communicated, as well as the extent to which they demonstrate that alcohol use has reinforcing qualities for themselves. 0n the other hand, these influences may be related to factors that are quite non-specific to alcohol use. For example, parental modeling of aggression and reinforcement of positive attitudes towards aggression may lead the child to become more antisocial. This may in turn be a pathway that leads to the child's later alcohol involvement via involvement with an aggressive and heavy drinking peer group. Evidence exists for alcohol specific effects very early in the child's development. Relationships between preschoolers' knowledge about alcoholic beverages and experiences in the home, including the level of parental drinking, have already been shown in a community sample (Noll, Zucker, & Greenberg, 1990), and in a sample of alcoholic families and matched neighborhood controls (Noll & Zucker, 1983). 6 Having an alcoholic parent has also been found in various investigations to be associated with greater amounts of childhood psychosomatic complaints, impaired sense of reality, suicide, depression, interpersonal difficulties, failure to complete school, and being a victim of incest or physical abuse and neglect (Deutsch, DiCicco, & Mills, 1982). Bailey, Haberman, and Sheinberg (1965) found more negative behavioral symptoms among school-aged children of alcoholics, especially temper tantrums and fighting at school. Parental alcoholism is among the risk factors which have been identified for the development of conduct disorders in children (Kazdin, 1987; West & Prinz, 1987). The greater inconsistency and unpredictability of parental support and expectations is thought to affect the children's sense of trust, security, self-esteem, and confidence in others. This may lead to problems in identity formation, personality development, role performance and the ability to form relationships with others. Anecdotal reports from social agencies that deal with children have long supported this conceptualization (Jacob, Favorini, Meisel, & Anderson, 1978). Fighting and dissension between the parents can mar the child's perspective on marital relationships (Fox, 1962). The sense of secrecy within the family may limit the amount of contact that the children of alcoholics have with others outside the family, depriving them of important supports for development (Clair & Genest, 1987). These childhood problems have been reported to show continuity into adulthood. Although it seems from this overview that children of alcoholics are at increased risk for experiencing a range of life difficulties, recent reviewers have suggested that this risk has been overstated and 7 the heterogeneity in outcome understated (Russell, Henderson, & Blume, 1984; Searles & Windle, 1990; Seilhamer & Jacob, 1990). Searles and Windle point out that no clinical syndrome has been identified which is distinct for children of alcoholics, and the commonly described symptoms (e.g., decreased self-esteem, depression, difficulties in close relationships) are certainly not unique to this population. They go on to point out that in most outcome studies the majority of children of alcoholics fall within the normal range even when as a group they show deficits relative to comparison groups. In fact, it is commonly found that some children of alcoholics are functioning well above the average level of comparison children on the outcome variables of interest (Clair & Genest, 1987). Searles and Windle (1990) conclude that "most individuals emerge from these environments relatively intact psychologically and emotionally" (p. 3). However, recent work (Luthar & Zigler, 1991) has pointed out that research on resilience to childhood stress has often neglected the assessment of internalizing problems, and this is certainly the case in the literature on children from alcoholic homes. While it is important to know the degree to which alcoholic families engender problems in their children (the "main effect" of parental alcoholism), there is a second step to the process of understanding. Given the increased awareness of heterogeneity of outcome, it is vital to better understand the precise determinants of these differential outcomes. As West and Prinz (1987) point out, investigators have largely overlooked such contributing or mediating variables as the child's age at the onset of the alcoholism, the severity and pattern of the alcoholism, concomitant psychopathology in 8 the parents, and the family's socioeconomic status. Werner (1986) has presented data which illustrates the degree of heterogeneity of outcome in children of alcoholics as well as the importance of examining a broad range of variables in developmental context. In this longitudinal study on the island of Kauai which followed 698 children from birth to age 18, 49 of the children had an alcoholic parent. At age 18, 59% of these children showed no serious adjustment problems, despite the fact that most had been raised in chronic poverty. Factors which predicted problematic outcome included child sex (70% of the children with significant problems were male), maternal competence, the quality of early caregiving, the degree of family conflict in early childhood, and the child's intellectual abilities and academic performance. In another vein, only a few controlled studies to date have compared a group of children with alcoholic parents to children of both normal control and non-alcoholic, but otherwise disturbed parents. Thus, we cannot be sure of the extent to which children of alcoholics encounter similar or different difficulties compared to children whose parents suffer from other psychopathologies. This question will be dealt with in more detail below. The next section examines current knowledge about the parent-child relationship in alcoholic families and in several closely related groups. Two of these groups, families of adolescent problem drinkers and families of later-to-be alcoholics, have alcohol-specific connections and show considerable continuity to the central group of this study, parents in alcoholic families. We then turn to parent-child relations as they relate to two types of psychopathology which are 9 commonly associated with alcoholism, antisocial behavior and depression, and examine the evidence for the effects of social class on parenting. Finally, the question of specificity of parent diagnostic category in determining child outcome is addressed. Pazent-Child Relationships among Groups at High Riak for Alcoholiaa Familias wiga alcoholic parents. Ablon (1976) has suggested that the presence of children may intensify the problems of the alcoholic. Cognizance of failure in the responsibilities of parenthood and related guilt, as well as the increased daily hassles of child-rearing may present added pressure for excessive drinking. At a more general level, one would expect that the characteristics and behavior of the child would affect his/her environment and, hence, his/her development (Belsky, 1984). Cork (1969) interviewed 115 children of alcoholics. Their ages ranged from 10 to 16, and they represented a "wide range" of socio- economic (SES) levels. The children's parents had volunteered them in response to a newspaper advertisement. Twenty-eight of the children had only an alcoholic mother, and in ten of the families both parents were alcoholic. In this highly selective sample, she found that over 90% of the children felt rejected by their parents and thought that their parents behavior was unpredictable. Over 70% reported that they had lost respect for the alcoholic parent. O'Gorman (1971) compared adolescent children of unrecovered alcoholics (N-29), children of recovering alcoholics (N-23), and children of parents with no drinking problems (N-27). She found that the adolescents who had unrecovered alcoholic parents were significantly lower on perceived parental affection and attention than were the other 10 two groups. Swiecicki (1969) conducted a ten-year follow-up study in Poland on 100 alcoholic families and 100 non-alcoholic control families matched on living standard, SES, and number of children. The target children were adolescents when the study began, and were young adults at the time of follow-up. Significantly fewer of the children of alcoholics reported having a good relationship with their father. There were no significant differences in the reported quality of relationships with their mother or siblings. Significantly fewer of the children of alcoholics thought that they would try to raise their children as they had been raised. A unique perspective on this question is offered by data from the Cambridge-Sommerville Youth Study (McCord, 1988, 1990; McCord & McCord, 1960, 1962). This was a study begun in 1935 that was aimed at reducing delinquency and monitoring the child-rearing practices of lower SES families. Families were generally first contacted when the targeted boys were between 10 and 16 years old. Information was gathered over up to five years of contact and included notes from direct observations of the families in their homes by multiple observers who did not know that any research would be done with them and were not focused on the question of alcoholism. In addition, information was available on which of the parents were alcoholic. A drawback to these data is that they were not gathered systematically but were pulled from case notes. Nonetheless, clear results emerged from these "blindly" gathered clinical data. The alcoholic fathers were more likely to be living apart from the family. They were less likely to be affectionate with their sons (McCord, 1988) and tended to be either excessively punitive or lax in discipline (McCord & McCord, 1962). The mothers in alcoholic 11 families were more likely to exert little control over their sons and both parents provided less supervision of the child. The McCords (1962) suggested that alcoholics are typically unsure of receiving love and are also very unsure of how to go about giving it. These fathers played a passive role in the family, and when they were active they tended to be dictators rather than leaders. Other authors (Hecht, 1973; Sloboda, 1974) have reported anecdotally that discipline is inconsistent in the homes of alcoholics. Often there are no clearly defined standards or rules within the family, which contributes to the sense of instability created by the erratic behavior of the parent(s) (Barnes, 1977). It is not yet well established to what extent physical abuse and violence are present in the parent-child relationships in alcoholic families (Reider et al., 1988). Some studies have found evidence for increased physical abuse of children of alcoholics. For example, Tarter, Hegedus, Goldstein, Shelly, and Alterman (1984) reported a sixfold increase in the rate of physical abuse compared to nonalcoholic families. This was based on retrospective accounts from mothers when the children were adolescents. Although physical violence toward the child may be greater in these homes, the data also suggest that this is not the most common method that parents use to deal with their children. In fact, parental withdrawal appears to be more common than physical abuse of the children (Zucker, 1979). The additional issue of whether any increased level of violence is specifically related to alcohol use has received little attention. Research has shown that alcohol use facilitates aggressive behavior (Bushman & Cooper, 1990), but the more specific question of determinants of violence in the family context has 12 not yet been clarified. In cross-sectional analyses done on a subset (N-9l) of our current sample, Reider et a1 (1989), found that the level of violence between parent and child was most strongly related to parental aggression (involvement in antisocial behavior and aggression toward the spouse) and the extent of alcohol related difficulty across the life span. Age of child was also found to be negatively related to the level of violence toward the child, but for mothers only. An earlier stage of the current investigation examined the relationships between parents' self-reported parenting practices and the parents' life circumstances and levels of psychiatric symptomatology in a sample of only alcoholic families (Davies, 1989; Davies, Zucker, Noll, & Fitzgerald, 1989, 1991). Sociodemographic attributes were strongly associated with reports of child-rearing patterns, while parental psychopathology was less strongly related. The demographic variables of interest were parent age, intelligence, and education, and family SES and income. The nature of the relationships between demographic variables and socialization was as predicted from the literature, indicating that these variables have similar effects in alcoholic homes as in non-alcoholic homes. Parents in families with greater resources, both personal and economic, tended to be less authoritarian and more concordant with their spouses regarding child-rearing. The fathers in these families tended also to be more nurturant. Maternal nurturance appeared to be independent of these parent and family characteristics. Maternal involvement in antisocial behavior was associated with both more belief in authoritarian control and with less affective positivity. Although the mechanism behind this effect is not yet l3 understood, the answer may lie in a more damaged past for the more antisocial women, including impaired socialization experiences as a child (cf. Egeland, Jacobvitz, & Sroufe, 1988). Maternal nurturance was also negatively associated with her level of lifetime alcohol problems and current depression. Paternal depression and, to a lesser extent, lifetime alcohol problems were the most important psychopathological correlates of parenting for fathers. In contrast to mothers, involvement in ASB was unimportant. Taken together, these results point to a constellation of parent/family characteristics involving fewer resources, lower achievement, greater parental depression, and greater maternal antisocial involvements, which are associated with greater authoritarianism, decreased parental positivity, and increased parental conflict about child-rearing in alcoholic families who are still at a very early stage of family development. Thus variability in socialization attitudes among alcoholic families appears to be related primarily to non-alcohol-specific factors. Another line of research paints a different picture of the influence of alcohol-specific effects on the parent-child relationship (Seilhamer & Jacob, 1990). Seilhamer (Seilhamer, 1987; Seilhamer, Jacob, & Dunn, 1990) did time series analyses which examined the day-to- day relationships between the level of drinking by alcoholic fathers and various measures of the functioning and behavior of their adolescent sons. Her sample consisted of eight families. In six of the eight families, the boy's reported satisfaction with the father-child relationship was affected by the level of his father's drinking. 14 However, there was much variation in the direction of this effect, with it apparently being mediated by the son's perceptions of his father's drinking. In those families where increased drinking was associated with more positive reports from the son, the child perceived the father as either more amiable when drinking or unchanged. When there was a negative association, the children uniformly reported negative perceptions of their father's drinking. These results serve as a caution against assuming that the short-run effects of parental drinking on the parent-child relationship will always be negative. In addition, considerable correspondence was found between the marital and father-son subsystems in the specific area of reaction to paternal drinking. When children were negatively affected by paternal drinking, their mothers tended also to be negatively affected; when children were positively affected. mothers were either unaffected or affected positively. Jacob, Krahn, and Leonard (1991) examined parent-adolescent interactions during problem-solving discussions in intact families of untreated male alcoholics (N-44) and depressives (N-37) and nondisturbed men (N-40) in drinking and non—drinking conditions. The oldest child in each family was chosen as the target child. Both male and female children were included, but the sex distribution of the children was not presented. The psychiatric samples were "clean" (i.e, the fathers evidenced only one psychiatric diagnosis and the mothers none). There was also evidence that the families, and the fathers in particular, were higher functioning than is the norm for the psychiatric population of the same developmental time frame. Not only were all the families intact, but the majority of the fathers were functioning as the family's 15 primary wage earner. Group differences were most pronounced for father-child interactions, where the control dyads showed increased rates of congeniality and problem solving compared to both psychiatric groups. The rest of the findings are remarkable for the lack of differences between the groups across mother-child and triadic interactions and drinking versus nondrinking conditions. The authors draw attention to this lack of alcohol-specific effects on parent-child interaction and point out the nonspecific effect of parental disturbance on parent- child interaction. It is also worthy of mention that this article raised a number of questions about the ethics of using a "drinking” condition in a study which includes men with a known history of alcohol difficulties (Jacob & Leonard, 1991; Kendall & Beutler, 1991; Koocher, 1991). Seilhamer and Jacob (1990) reach the following conclusions based on research findings and theoretical models: 1) there is an ongoing association between parental drinking and the parent-child relationship; 2) this association involves a aaaaai relationship, in that drinking/ intoxication effects disturbances in the parent-child relationship; and 3) while parental alcoholism is assumed to cause negative outcomes for children in the long run, the quality of the parent-child relationship during day-to-day cycles of sobriety and intoxication may vary with drinking pattern and consumption level. (p. 182) The divergent pictures which emerge from these two literatures (epitomized by the Zucker and Jacob research teams) are no doubt due in 16 part to methodological differences. For one, age differences of the children and the families are substantial. The Michigan State University Studies are based on families with preschool age children and are restricted in focus to male children. Jacob et al.’ 3 studies are based on families with adolescent male and female children. The nature of the alcoholism in these families is probably also quite different. In the former group, families constitute a drunk driver population that are obtained through the District Courts. Co-occurring parent and family pathologies are common and the parents are likely to be depressed or antisocial (Noll, Zucker, Fitzgerald, & Curtis, 1990; Reider et al., 1989). In the latter group, families are volunteers to a newspaper advertisement. Families with mothers or fathers evidencing psychiatric problems other than the father's diagnosis of classification (i.e., alcoholism or depression) are eliminated. And, as a group, they probably tend to be higher functioning because, as noted above, most of the fathers continue to be the primary wage earner and all are in intact families at a later stage of family development those in the Michigan State studies. Finally, in the Jacobostyle studies, drinking patterns are measured at a micro level and are conceptualized as primary determinants of child and family outcomes, especially in the short run. By contrast the Zucker-type studies seek to measure a wider variety of parent and family variables and present a more systemic picture of family process. Drinking is measured by asking about recent average consumption and the effects of drinking are seen as a cumulative process. Clearly it would be impossible for either approach to yield the findings that have emerged from the other. Integration of these two 17 general approaches will be necessary before a final determination can be reached about the relative contributions of alcohol specific and non- specific effects on the parent-child relationship across different population groups and at different stages of family development. Much of the literature discusses the effects of parental alcoholism with an apparent assumption that alcoholism entails a self- explanatory and unitary behavioral entity (Wilson & Orford, 1978). Several important, yet unexplored, parent and family characteristics are likely to be significant sources of difference within the population of alcoholic families. For one, few studies have attempted to observe the differential impact of maternal versus paternal alcoholism as it affects male and female children (Deutsch et al., 1982; Williams & Klerman, 1984), the effects of living in a home with two alcoholic parents vs. only one, or the effects of different patterns of parental alcoholism. Data are particularly scarce concerning the effects of maternal alcoholism, although this situation is slowly changing. Consistent differences have been observed in the drinking patterns of male and female alcoholics, and in mothers' and fathers' roles within the family. These differences can be expected to alter the way in which the alcoholism affects the children. Thus, the alcoholic mother is more likely to be hidden, ignored, or protected by family members, to drink at home, and to exhibit drinking patterns that are closely tied to family dynamics, crises, and developmental transitions (Jacob & Seilhamer, 1987). Female alcoholism has also been more consistently related to depression (Rivers, 1989; Sclare, 1970) and other neurotic symptomatology (Fellios, 1989), which can also interfere with the mother-child relationship. The typically increased responsibility for 18 child-rearing held by the mother is likely to mean that the effects of her drinking are more severe (Richards, 1979). Cork (1969) found that adolescents view an alcoholic mother's drinking more negatively than an alcoholic father's. The children reported that heavy drinking among mothers was "not nice; not like a mother." Krauthamer (1974, 1979) found that alcoholic mothers tend to be more ambivalent, confused, and inconsistent about parenting than nonalcoholic mothers. The children of the alcoholic mothers were more cold, distrustful, rigid, reserved, submissive and dependent. Of course, these data are in keeping with the effects of having an alcoholic parent of either sex. However, data from Miller and Jung's (1977) longitudinal study do seem to indicate a greater incidence of psychopathology in the offspring of female alcoholics than male alcoholics. Williams (1987) found significant differences in the environments and child-rearing patterns of male and female alcoholic parents, but her findings are heavily confounded by the lower SES of her female alcoholics as well as a higher likelihood of having an absent or alcoholic spouse. Her study also included no control group and included a wide range of child ages (from birth to 18). Alcoholics are more likely to be married to alcoholics than to non-alcoholics (Hall, Hesselbrock, & Stabenau, 1983). Proposed explanations for this finding have included assortative mating and convergence of behavior after marriage. Regardless of the mechanism behind this effect, it has been shown to be a risk factor for the development of alcoholism and antisocial behavior in the couple's children (Merikangas, Weissman, Prusoff, Pauls, 5: Lechman, 1985; Williams, 1987). 19 Studies utilizing interviews with children of alcoholics have shown that a variety of patterns of parent-child relationships can exist in alcoholic families (Cork, 1969; Wilson & Orford, 1978). Families differed along dimensions such as the consistency of the children's feelings toward the alcoholic parent, the relationships of different siblings toward the parents within the same family, the ease with which the parents report they can relate to their children, and the relationship between the children and the nonalcoholic parent. These findings prompted Wilson and Orford to conclude that "parental alcoholism probably is not in itself highly predictive of the quality of parent-child relationship" (1978, p. 129). However, these conclusions are based on studies done with no control groups. Certainly it is important to document that there is variability in parent-child relationships in alcoholic families; but the question is left open whether there are general differences which can be identified that separate alcoholic families from nonalcoholic families in this regard (a main effect), and whether variables can be identified which would be useful in predicting the course of parent-child relationships in these families (either interactional effects or evidence that the same factors account for differences in developmental pathways across groups). Factors which Wilson and Orford (1978) suggest would be useful are the behavior and personality of both parents; the sex, age and temperament of the child; and factors of family structure, such as family size or parental separation. Also likely to be important are subcultural issues, social status, severity and type of parental alcoholism, and concomitant psychopathology other than alcoholism. It seems clear, then, that we are a long way from understanding 20 the whole picture of the effects of parental alcoholism on the parent- child relationship and the processes through which these effects operate. More attention needs to be paid to the differential effects of the parents, the degree and type of disturbance in both parents, the age and personality traits of the child, and family factors such as sibling structure and extrafamilial contacts. There is a comparative dearth of research on alcoholism, including family effects, among ethnic minorities (Harper & Dawkins, 1976; Watts & Wright, 1983) despite evidence for differences in drinking patterns and physical and psychological sequelae (Brown & Tooley, 1989; Harper, 1983; Neff, 1986). Thus findings of this review and this study are applicable only to white families at this time. Brisbane (1987) identified unique developmental risks for the oldest daughter in Afro-American alcoholic families. It is common for the oldest daughter to become a "second mother" who helps care for children and takes charge when necessary. When this occurs in the context of parental alcoholism, she may be left shouldering significant responsibility without adequate parental support. Brisbane argued that these children are likely to encounter difficulties with identity formation. This work illustrates the need for further research on the interaction of alcoholism, culture and ethnicity, and child development. We now turn to an examination of parent-child relationships in families of adolescent problem drinkers and families whose children develop alcoholism later in life. Given our assumption that there is a continuity of process that creates the intergenerational transmission of the problem, these two bodies of literature shed further light on the features of parent-child relationships that are associated with the 21 development of alcoholism. Eaaiiiea of adoiescept pgobiem dpipkegs. The Rutgers Community Study (Zucker & Barron, 1973; Zucker & Devoe, 1975; Zucker & Fillmore, 1968) examined the relationship of parental drinking characteristics, personality, and child-rearing practices to children's drinking and antisocial behavior. The adolescents in the study were a stratified sample of students (male and female) from the only high school in a Middle Atlantic community of 15,000. They were chosen to reflect the diversity of drinking patterns in the school. As many of their parents as possible were then recruited, and data on these domains were then collected from both groups. Child-rearing practices were measured using the Bronfenbrenner (1961) Parent Activity Inventory. Based on the parents' reports of their behavior, it appears that the mothers are more strongly implicated in the development of problem drinking behavior in their sons (Zucker & Barron, 1973). The mothers of heavy drinking and problem drinking boys were found to be heavier drinkers themselves as well as more cynical and more antisocial. Their child-rearing practices were marked by more open rejection (via ridicule and unfavorable comparison to other children) and less parental pressure. They tended to discipline by deprivation of privileges and property and through social isolation. Fathers were similar in that they were also heavier drinkers and more antisocial, but their child- rearing practices were related to their son's drinking only in that they too tended to use removal of supplies as a disciplinary technique (Zucker, 1976). The boys' reports present a starkly contrasting picture, indicating little or no relationship between the mother's practices and 22 the child's behavior and strongly indicating the influence of the father in the process. Only the mother's absence was related to the development of problem drinking. The fathers were seen by their problem drinking sons as affectively distant and unrewarding and as not having attempted to shape their sons as they were growing up (through achievement demands, instrumental companionship, and principled discipline). The fathers were not seen as malevolent or destructive, but rather as not having been engaged in those activities that build a strong relationship between father and son (Zucker & Barron, 1973). Family atmosphere measures indicated that there was a good deal of tension. There was also a good deal of open rebellion and resentment of parental intrusiveness. Zucker and Barron (1973) interpreted the seeming contradiction between parents' and son's reports to indicate that in the homes of heavier drinking adolescents, the mothers are indeed harsher and more rejecting. The unfavorable light in which the fathers are viewed is seen as reflecting the failure of the father to provide the child with the support that the mother has already failed to provide. Thus, in terms of developmental time, the perceptions of the father follow and may be caused to some extent by the mother's behaviors. The primacy of the mother's rejection is a reflection of the normatively greater influence of mothers in child rearing. The adolescent's drinking and antisocial behavior can be seen as a way to escape from a family setting that is not at all rewarding. According to information provided by the parents, heavier drinking among the girls was associated with heavier drinking by both parents, a personality style of aggressive sociability in the mother that is 23 oriented towards peers rather than family, little mother-child interaction, and few attempts by the mother to shape the daughter's behavior. Father absence was the only father variable other than his drinking that was related to the girl's drinking (Zucker, 1979). The heavy drinking girls reported a more negative picture of rejection, neglect, and lack of nurturance, affection, and companionship from both parents than do the heavy drinking boys. They felt that there was a lack of principled discipline from both parents, but saw the father's discipline as especially arbitrary. The girls' data seem to implicate both parents about equally, in contrast to the boys', which indicated a greater effect for the father or the same-sex parent. On the family atmosphere measures, alcohol consumption for the girls was related to perceived parental defiance and, again, to family tension (Zucker, 1976; Zucker & Devoe, 1975). Barnes (1977) found support for the general hypothesis that problem drinking among adolescents is associated with what she terms "incomplete, inadequate socialization within the family." Specifically, the problem drinkers are less likely to feel ”very close" to their family; more likely to feel loosely controlled (especially by their mother) and to feel rejected (especially by their father); and report more conflict in their relationship with their father. Barnes suggested that the inadequacy of the socialization process leads to a diminished ability to cope with problems, which, in turn, increases the likelihood of problem drinking. In a more refined study, Barnes, Farrell, and Cairns (1986) found that the level of parental support and parental control is related to the level of adolescent drinking. The development of nonproblematic 24 drinking behavior is facilitated by a high level of parental support and a moderate amount of control. Problem drinking among the adolescents was associated with low levels of maternal control and high levels of paternal control. Other investigators have shown that positive parent-child relationships are related to the absence of problem drinking. Wechsler and Thum (1973) found that adolescent nondrinkers and light drinkers have a greater perceived closeness to both of their parents. Demone (1973) found that nondrinking males reported a higher level of parental confidence and more issue-centered discussion between parents and child. Alexander (1967) found that the father's abstinence is related to rebellious, abusive drinking only when the father-child relationship is emotionally distant. Studies of family factors related to involvement with drugs in adolescence paint a similar picture. Increased drug involvement is associated with both parental personality attributes (e.g., psychiatric symptomatology, antisociality) and aspects of theparent-child relationship (e.g., identification, relationship satisfaction, time spent together)(Brook, Whiteman, Gordon, & Cohen, 1986). The model which best accounts for the data posits that parental personality attributes affect both the parent-child relationship and the adolescent's personality, and that the latter two have a direct impact on the child's involvement in drug use. The studies of the Brook group have consistently pointed to the importance of parent-child mutual attachment in determining drug involvement (Brook, Whiteman, & Gordon, 1983; Brook, Whiteman, Gordon, & Brook, 1981, 1984; Brook et al., 1986). The mother's influence in the process has also been consistently found 25 to be larger than the father's (Brook et al., 1986; Jessor & Jessor, 1974). The conclusion offered by Brook and her colleagues (Brook et al., 1986) sums up well the findings about the effects of the parent-child relationship in affecting drug problems (including alcohol problems) among adolescents: ...the psychological stability and conventionality of the mother are related to an affectionate and nonconflictual relationship with her child and to conventionality and psychological stability in the child. These qualities, in turn, are associated with lower stages of drug use. (p. 466) The same process occurs for fathers, but its impact appears to be weaker. Recent work by Shedler and Block (1990) indicates that adolescents who experiment with drugs tend to be better adjusted than either abstainers or heavy users. Wexler (1991) draws on previous research to argue that a distinction needs to be made between principled non-users (those who have made an active decision not to use drugs now or ever) and "simple" non-users (who are more similar to occasional users). It seems likely that these same distinctions would be useful with regard to categories of drinking in adolescence and young adulthood. While the results of some studies indicate an essentially parallel process in the development of problem drinking in male and female adolescents, the preponderance of studies indicate that there are important areas of difference (Zucker, 1979). This examination of the literature has focused on males, as will this investigation. The 26 existence of sex differences dictates that these findings should not be generalized to parent-daughter relationships. Eaailies of later-to-be alcoholica. A great deal of our knowledge about the development of alcoholism has been established through correlational designs. However, an increasing number of studies have utilized one of various longitudinal designs, which place the conclusions on firmer methodological ground and permit better understanding of the time ordering of events (Zucker & Noll, 1982). This discussion will focus primarily on the findings of these longitudinal studies. There have been six major longitudinal studies of the development of alcoholism (Jones, 1971; McCord & McCord, 1960, 1962; Monnelly, Hartl, & Elderkin, 1983; Ricks & Berry, 1970; Robins, Bates, & O'Neal, 1962; Vaillant & Milofsky, 1982). In their review of these studies, Zucker and Gomberg (1986) concluded that the parent-child interaction in the alcoholics' families of origin were characterized by inadequate parenting and a lack of parent-child contact. These interactions were described in such ways as involving inadequate or lax supervision, the absence of parental demands, parental disinterest, or lack of affection for the child. The levels of parental adequacy examined in these studies range from a grossly defined measure of the provision of such basic obligations as physical care and financial support (Robins, et a1, 1962) to more psychologically oriented measures of the level of family attachment (McCord & McCord, 1962). Nonetheless, a clear pattern emerges which links inadequate parent-child contact and parenting skills to the development of alcoholism in the children. This lack of parent- 27 child contact can occur with the parent in the household or as a result of parental separation or death (Zucker, 1979). If the focus is broadened beyond parent-child interaction, a lack of family cohesiveness and socially deviant behavior on the part of the parents are also indicated (Zucker & Noll, 1982). These parental influence findings do show substantial across-study variation in some areas (Zucker & Noll, 1982). Variation is most evident concerning the amount of parental deviance reported and the amount of harshness or rejection found in parent-child interactions. Some of these inconsistencies are due to the social class from which the samples were drawn (Zucker & Noll, 1982). The implicated parental behaviors will be much more destructive when they occur in lower SES families (Zucker, 1976). For example, in one study of alcoholic mothers it was found that children of alcoholic mothers who were from low SES families showed impaired emotional, intellectual, and academic functioning compared to controls. In contrast, none of these problems appeared in the high SES families, despite the fact that the parent- child relationships were seriously disrupted (Krauthamer, 1974, 1979). Other variations are traceable to problems in the study designs (Zucker & Noll, 1982). In addition, there is evidence from the cross-sectional studies for an interaction between sex of parent and sex of child which is not yet clearly understood (Zucker & Noll, 1982). This reinforces the limits on generalizability of results discussed above. The longitudinal findings on alcoholic men in many ways parallel the findings of the cross-sectional studies on alcoholic and adolescent problem drinking families. There is a lack of family cohesiveness, parental deviance, and inadequate or inconsistent parenting rather than 28 actively rejecting parental behavior (Zucker & Noll, 1982). The parents of prealcoholics do seem to be more uncaring and inconsistent in their supervision and discipline than the parents of adolescent problem drinkers (Zucker, 1976). Eami1iaa_a£_ap§iappial_agpia§gap;§. There is impressive evidence that the phenomenon of adolescent drinking represents one expression of more generalized antisocial behavior in adolescence (Donovan & Jessor, 1985; Zucker, 1979). The familial correlates of delinquent behavior are very similar to the findings discussed above for problem drinking. In fact, the Rutgers Community Study (Zucker & Barron, 1973; Zucker & Devoe, 1975; Zucker & Fillmore, 1968) set out to test the general hypothesis that family interactions which had previously been linked to impulsive and antisocial behavior in adolescents (e.g., weak family affectional ties, inadequate discipline, and low family cohesion) would also be predictive of problem drinking. This hypothesis received strong support. Jensen (1972) found that there is a relationship between delinquent involvement and such variables as parental support, parental supervision, and the affective quality of father-son relationships independent of the number of delinquent friends that the son reported. In a review of the delinquency literature, Cove and Crutchfield (1982) concluded that the level of parent-child attachment is the strongest predictor of delinquency. An extensive review of the longitudinal and cross-sectional literature on the effects of family factors on the development of delinquency has been conducted by Loeber and Stouthamer-Loeber (1986). They concluded that the best predictors were variables associated with 29 socialization, such as a lack of parental involvement and supervision, parental rejection, and a lack of parent-child involvement. Of medium strength were background factors such as the parents' marital relationship and the level of the parents' criminality. Weaker predictors were a lack of parental discipline, parent health, and parent absence. These findings hold for longitudinal studies as well as those comparing delinquent/nondelinquent and aggressive/nonaggressive children. Data from normal families show less importance on socialization factors and more importance on parent-child rejection. Within those families which did have a delinquent child deficits in parenting skills were related to more serious delinquency. The McCords' finding about the disciplining style of alcoholic fathers (that they tend to be either erratically punitive or lax) is parallel to the parenting characteristics that they found to be associated with the development of delinquent and eventual criminal behavior (McCord & McCord, 1962). Thus we should expect that the parenting characteristics that we hypothesize to be related to the development of problem drinking and alcoholism are similarly predictive of eventual delinquency and adult antisocial behavior. Famiiies with depressed parents. In recent years, alcoholism (alcohol abuse/alcohol dependence) has been clearly linked to other forms of psychopathology, most notably depression and antisocial behavior (Helzer & Pryzbeck, 1988; Penick, Powell, Liskow, Jackson, & Nickle, 1988). This clustering occurs despite the fact that there is independence of transmission (Merikangas et al., 1985). Given these comorbidities, the contribution of each type of symptomatology should be examined for potential impact upon parenting in alcoholic families, both 30 directly and in interaction with alcohol specific variables. Comparatively little is known about the socialization practices of antisocial parents. The evidence that we do have is drawn from studies addressing the development of antisocial behavior in children (see above) where parental antisociality and socialization deficits have both been found to be risk factors (e.g., Loeber & Stouthamer-Loeber, 1986). Direct investigation of the parent-child relationships of antisocial parents would be desireable. A great deal more attention has been paid to parent-child relationships among depressed parents (see review by Downey & Coyne, 1990). Like alcoholism, depression is seen to have a genetic component to its transmission, while the discordance with the genetic model suggests that environmental aspects are also important (Kidd & Weissman, 1978). The effects of growing up with a depressed parent are in some ways quite similar to those seen in children of alcoholics. They have been reported to evidence an elevated rate of depression and other psychiatric disorders, as well as problems in other life areas, such as interpersonal problems, physical and emotional complaints, behavioral and school problems, and attentional and cognitive disturbances (Billings & Moos, 1983; French, 1983; Hirsch, Moos, & Reischl, 1985; Kuyler, Rosenthal, Igel, Dunner, & Fieve, 1980; Welner, Welner, McCrary, & Leonard, 1977). Children of depressed parents appear to be especially vulnerable to internalizing behavior problems (Lee & Gotlib, 1989). Lee and Gotlib (1989) suggest two processes which may lead to the transmission of these difficulties: 1) increased self-focus of the parent leads to a lack of awareness of the child; and 2) these children 31 may be deprived of the opportunity to develop appropriate conflict- resolution and affect-regulation skills. The latter point is supported by observational studies of depressed families which have shown that depressive symptoms may be functional in reducing aversive exchanges between family members (Hops et al., 1987; Kochanska, Kucynski, Radke- Yarrow, & Welsh, 1987). Depressed families (i.e., families with a depressed parent) are characterized by more conflict, and less cohesion, expressiveness, and organization. There is less emphasis on independence and on intellectual-cultural, moral-religious, and shared recreational activities (Billings & Moos, 1983). Billings and Moos (1983) argue that this family environment is the result of a complex interaction between the family environment and the individual family members. They do not claim that the parental depression "causes" these environmental effects. As we would expect, the degree of impairment of the child is associated with the level of stress and support experienced by the family members (Billings & Moos, 1983; Hirsch et al., 1985). Specific deficits have been identified in the child-rearing environments of depressed mothers. They have been rated as less affectively involved and less responsive to their children than normal mothers beginning at very young child ages (Goodman, 1987). As with much of the alcoholism research, many of the studies on depressed families are methodologically weak. Few studies have examined the relative effects of different types or severities of depression. Susman and colleagues (Susman, Trickett, Iannotti, Hollenbeck, & Zahn- Waxler, 1985) have used the Child-Rearing Practices Report (CRPR)(Block, 1980) to examine self-reported child-rearing patterns in mothers with 32 current or past major or minor depression as compared to abusive and normal mothers. They found that mothers with current major or minor depression reported high levels of inconsistency in discipline and control. In addition, mothers with current major depression were more likely to use guilt and anxiety induction techniques to control their children, and were more likely to express feelings of disappointment about their children. Mothers with either current or past minor depression and past major depression were remarkably similar to the normal mothers. This suggests that the problems associated with depressed parenting are closely tied to the parent being in the depressed state, rather than being due to the personality traits of the depressed parent. However, none of the depressed groups were nearly as disparate from the control group as the abusive mothers. Stoneman, Brody, and Burke (1989) examined the relationships between depression, marital satisfaction, and various aspects of parenting, but with a particular emphasis on inconsistency. They used a volunteer sample of 47 two-parent families with two same-sex children. The children ranged from 4.5 to 9.5 years old. Although this sample was not chosen to include clinically depressed parents, several of the parents were in this range as measured by the Beck Depression Inventory. The CRPR was used as the measure of self-reported parenting. The mother-son dyad was the only parent-child combination which did not show at least moderate associations between reported depression and marital satisfaction and inconsistency in parenting. Paternal depression, but not maternal, was also found to be associated with increased parental disagreement about discipline. Jacob and Leonard (1986) examined the psychosocial functioning of 33 children of depressed, alcoholic, and control fathers. Their sample consisted of around 45 fathers from each group, each of which had at least one child between 10 and 18. Sons of both alcoholics and depressives were rated by their parents as higher than controls on behavior problems, internalizing behavior, and social incompetence. For behavior problems and internalizing and externalizing behavior, the daughters of depressives were rated higher than the daughters of alcoholics, who, in turn were rated higher than the daughters of controls. Although the majority of adolescents in all groups were not in the range of severe impairment, 23% of the alcoholics and 15% of the depressives had at least one child in the range of severe impairment. None of the children of controls were in this range. Comparisons were made between the functioning of alcoholic fathers with and without a child in the range of severe impairment. They found that having a child in the impaired range was associated with more alcohol related difficulties in the preceding month, higher self- reported depression scores, and more concomitant psychopathology. The mothers of these children also reported more psychopathology. Similar analyses between the groups of depressed fathers revealed few differences. The finding that child behavior problems among children of alcoholic fathers are associated with more paternal life problems (both alcohol related and unrelated) has recently been replicated (Seilhamer 6 Jacob, 1990). The influence of the parent-child relationship in the etiology of adult depression has been examined primarily in retrospective studies. Crook, Raskin, and Eliot (1981) compared reports of early parental behavior from 714 hospitalized depressives and 387 non-depressed adults. 34 Depressives were more likely to report parental rejection and control through derision, negative evaluation, and withdrawal of affection. Maternal rejection was more strongly associated with adult depression for females than for males, while paternal rejection presented a sex-by- race interaction. Holmes and Robins (1988) examined perceptions of early home environment by 46 adult depressives, 54 adult alcoholics, and 100 adult controls. Reports of unfair, harsh, and inconsistent discipline were strongly related to both depression and alcoholism. There were also evidence of continuity in psychiatric diagnosis from parent to child that was specific to each diagnosis. The connection between early disciplinary practices and adult pathology was strongest for subjects who reported psychopathology among their own parents, but continued to exist for those subjects who reported no parental pathology. When controlling for SES of family of origin, effects were weakened but still significant. There was evidence that poor discipline was a better predictor of depression in middle-class children and alcoholism in low SES homes. Effecga of §ocioeconomic status. As pointed out above in relation to both parental alcoholism and parental depression, family SES is a related and potentially confounding factor when examining child-rearing practices and parent-child relationships. Indeed, social class has been the most studied determinant of the parent-child relationship (Belsky, Hertzog, & Rovine, 1986). Handel (1970) has concluded, "The differences in the parentochild relationship between the middle-class and the working and lower classes arise from systematically different life experiences which, in turn, lead to differences in basic outlook and 35 life style" (p. 97). He argues that these differences are expressed through a greater emphasis on parental control and proscription among lower SES parents and more awareness of issues of parental support and paternal involvement among middle-class parents. These differences have also been attributed to the application of the value system of the parent's work environment to child-rearing (Bronfenbrenner 6: Crouter, 1982; Kohn, Naoi, Schoenbach, & Schooler, 1990). Pittman (1985) has argued that parent-child relationships in wealthy families are also qualitatively different from other groups. The effects of poverty in particular have received a great deal of attention (Halpern, 1990; McLoyd, 1990). As Hamburg (1985) has concluded: Poverty does not harm all children, but it does put them at greater developmental risk, through the direct physical consequences of deprivation, the indirect consequences of severe stress on the parent-child relationship, and the overhanging pall of having a depreciated status in the social environment. (p. 4) Despite this attention, the relationships between the complex variables of poverty, child-rearing, and child development are far from clear and a great deal of variability remains unexplained (Halpern, 1990; Sameroff, Seifer, Barocas, Zax, & Greenspan, 1987). Many authors have argued that the search for sociological connections in this area is misdirected if the goal is to understand the effects on the parent-child relationship (Belsky et al., 1986; Block, 1971; Bronfenbrenner, 1986; Elder, 1981; Zigler, Lamb, 6: Child, 1982). For example: 36 Although the global characterization of the family's socioeconomic status remains a powerful predictor of parental and child functioning, what is gained in variance explained is often lost in conceptual specificity. . . .(The) construct is too broad and provides too little insight into processes of influence. (Belsky et al., 1986, p. 155) Similarly, Elder (1981) states that the effects of social class "cannot be understood until the structural variable is conceptualized as a set of psychological processes or mechanisms that cause the outcome to be explained" (p. 81). Despite these well-directed observations, social class continues It to be one of the most studied determinants of child development. will continue to be, and indeed cannot be ignored, until we are able to account for it through psychological variables. This must happen through a combination of (1) identifying psychological variables which can account for the variance currently attributed to variability in SES; and (2) directly examining psychological process variables that are associated with differences in SE8. Some progress has been made in each of these areas. From the first direction, parental behaviors have been found in some studies to account for more variance than SES or other sociodemographic factors (Braithwaite 6: Gordon, in press; Wolf, 1966). The present investigation will also examine relative contributions of parenting behaviors, other parent characteristics, and sociodemographic factors. From the other direction, Susman et a1. (1985) present data collected from a normal group of lower SES mothers (to serve as a 37 control group for the abusive mothers discussed above). These mothers' patterns differed sharply from that seen in the middle-class control The lower SES group expressed difficulty in positively group. encouraging their children, talking to their children about feelings, enjoying and finding satisfaction in them, and in trusting their children to make decisions. This is in keeping with the results from the earlier stages of this investigation. Davies et a1. (1990) found that sociodemographic variables were directly associated with parent's reports of their child- rearing practices and the degree of dyadic concordance about child- rearing. It is important and encouraging that variability in SE8 has been found to be related to socialization ideologies, since this is clearly a psychological variable which affects child development. However, this picture will not be complete until we understand how this linkage comes about and how and when the connection is broken. In view of the continued finding that socioeconomic variables are robust predictors of the quality and outcome of the parent-child relationship, research into the determinants of parent—child relationships must utilize a multifactor approach that includes both sociodemographic and pertinent psychological variables. Parental psychopathology and the processes associated with lower SES probably both make contributions to problems in child development, and probably have a cumulative and/or interactive effect in combination. Early data from the Rochester Longitudinal Study (Sameroff, Seifer, Zax, 6: Barocas, 1987) suggest that the effects of having a parent with psychiatric problems may be more pervasive than the effects of low SES. 38 Their findings suggest that the latter may just delay development, while the former may involve qualitative distortions in development. 0 G v c c e o e We do not yet know whether parents with psychopathology influence child development through specific effects caused by those conditions, or whether there is a generalized effect on development of having a disturbed parent. As Jacob and Leonard (1986) stated, "the absence of psychiatric comparison groups has made for difficulties in attributing any observed impairments to alcoholism per se rather than to a general parental psychopathology effect" (p. 374). A number of the studies discussed above have compared different parental diagnostic categories and their effects on the parent-child relationship and child outcome (Hirsch et al., 1985; Holmes & Robins, 1988; Jacob & Leonard, 1986; Lee & Gotlib, 1989; Susman et al., 1985). Taken together, these studies provide evidence for parental psychopathology as a generalized risk factor for child development. Harder and colleagues (Harder, Kokes, Fisher, 6: Strauss, 1980) have also found evidence that the usual level of parent functioning may be more critical than the presence of specific disorders in determining the effects on the children. However, we have also seen evidence for specific effects of diagnosis (Holmes 6: Robins, 1988), of severity of symptoms (Jacob & Leonard, 1986), and of current versus past experience of symptoms (Susman et a1. , 1985). Hirsch et a1. (1985) found that adolescent children of arthritic parents were similar in risk status to adolescent children of depressed parents (i.e. , that negative parental life events were associated with child symptomatology), suggesting the need to go beyond parental psychological problems per se to a more general definition of parental 39 This finding is in apparent contradiction to work by Lee life problems . The latter study found that children of nondepressed and Gotlib (1989) . medical patient mothers did not present more behavior problems than children of nondepressed nonpatients, and presented significantly less than children of depressed or nondepressed psychiatric patients. Lee and Gotlib suggest that this may be due to the failure of Hirsch et al. to screen their arthritic sample for psychiatric problems. It is also possible that arthritic parents are different from other groups of medical patients in the ways that they interact with their children. It may be that an answer to the question lies in an examination of the degree to which the medical condition in question is chronic, impedes daily functioning, and leads to psychological self-absorption. There is little research on the relative, cumulative, or interactive effects of maternal and paternal psychopathology. As discussed in relation to alcoholism, we would expect the effects of maternal pathology on the child to be greater given the still typically greater responsibilities of mothers in child care, as well as the mother's role as "kin-keeper" and emotional specialist (Cohler, 1984). The question of interplay between parental difficulties is important given evidence that maternal and paternal psychopathology are highly interrelated in multiproblem families (Reider et al. , 1988). At the same time, one might also anticipate that the degree of relative contribution of each parent would shift as the child gets older. There are different approaches that have been used in attempting to evaluate the impact of specific effects of parental disorders (in our One approach is to work with samples that are case alcoholism). For example, the Jacob and relatively free of other psychopathologies 40 Leonard (1986) study included only families in which the father satisfied Research Diagnostic Criteria (RDC) for the diagnostic category he was in and met the RDC for no other mental disorder, and the mother did not satisfy the RDC for any current major disorder or have a history of alcohol-related problems. The difficulty with this approach is that the majority of alcoholic families have difficulties in a variety of areas, so the findings from such research are of limited usefulness in helping to understand the process for most alcoholic families. Instead, the approach taken by our research group, and in the work preliminary to this study (Davies et al., 1991), has been to draw a community (i.e., "uncleaned") sample and to attempt to measure each of the dimensions which seem likely to be important. Additionally, symptoms have generally been treated dimensionally rather than categorically. This is in recognition of considerable heterogeneity of frequency and severity of symptoms in groups meeting and failing to meet standard diagnostic criteria, and that this variability is likely to be meaningfully related to the dependent variables of interest. The present investigation strikes a compromise between using dimensional and categorical psychopathology variables. Standard nosological criteria are used to define groups based on the presence/ absence of alcoholism in both parents and the severity of parental alcoholism, but alcohol-specific and non-specific symptoms are also assessed dimensionally to allow more refined assessment of these dimensions. Statement 0 the roblem There is considerable evidence that parent-child relationships in alcoholic families are problematic and that this may, in turn, 41 contribute to maladaptive child outcomes. It is also clear that there is considerable heterogeneity in parent-child relationships in these families and in child outcome. This study builds on previous work (Davies et al., 1991) which showed that variability in socialization attitudes among young alcoholic families was related to sociodemographic characteristics, and, to a lesser extent, to parental psychopathology. This study extends the previous work by: (1) including community control families; (2) assessing alcoholism status of both fathers and mothers; (3) expanding the range of variables assessed across the family system; and (4) making greater use of formal diagnostic criteria which facilitate comparison to other samples. No other work to date has explicitly examined child-rearing patterns and the family environment in alcoholic families at such an early stage of family development. It follows that we also know little about differences in the family environment and socialization practices between young families with one versus two alcoholic parents, although this has repeatedly been identified as a variable which deserves more attention. The sample for this investigation is unique and allows an examination of early processes in a sample which is at considerable risk for developing problems as the family and child develop. It is a community based sample which approximates the population of families with drunk-driving parents and preschool-aged sons. In addition to being more representative of alcoholic families than samples drawn from treatment populations, it allows examination of the parent-child relationship at an earlier developmental level than the great majority of existing research. The controls are drawn from the same census 42 tracts as the alcoholic families, and are identified through neighborhood canvassing. This ensures a more representative sample than methods that involve self-selection. The theoretical model underlying this research is one of risk cumulation (Cicchetti & Rizley, 1981; Sameroff & Chandler, 1975). Parental alcoholism, other parental psychopathology, lower family resources, disturbed family environments, and child characteristics which lead to aversive parent-child interactions are all seen as risk factors that may lead to a negative parent-child relationship. The presence of multiple risk factors would be expected to further increase such risk, while the absence of a particular risk factor can be seen as protective. The literature on alcoholic families suggests that the alcoholic families in this sample should differ from controls on each of these "risk" variables. This leads to an expectation that the alcoholic families as a group would report less optimal socialization practices, and further, that the group differences would be due both to the parent's alcohol diagnosis (or parents' diagnoses) and to other, alcohol-nonspecific characteristics of these high-risk families. Hypogheaes (1) Nonalcoholic families will have lower levels of parental psychopathology, family conflict and unsupportiveness, and child behavior problems than families with an alcoholic father and nonalcoholic mother. Similarly, families with only an alcoholic father will have lower levels of parental psychopathology, family conflict and unsupportiveness, and child behavior problems than families where both the mother and father are alcoholic. (2) Parents (both mothers and fathers) from nonalcoholic families will (3) 43 report lower levels of authoritarianism and higher levels of nurturance and dyadic concordance than parents from families where the father is alcoholic. Families where only the father is alcoholic will, in turn, show the same pattern of differences (i.e., less authoritarianism, more nurturance and concordance) compared to families in which both the mother and father are alcoholic. Controlling for demographic, parental psychopathology, and family environment variables which differ significantly across groups will help explain differences between groups on the child-rearing variables. CHAPTER II Method Subjacga The subjects for this study were drawn from the Michigan State University Longitudinal Study (Zucker, 1987; Zucker, Noll, & Fitzgerald, 1986), a longitudinal study examining the factors that contribute to the development of alcoholism and other conduct problems in the offspring of alcoholic men. All families who had completed data collection were included in this study. The sample consists of 133 alcoholic families and 32 community contrast families recruited from the same neighborhoods as the drunk drivers. All of the families in this study were of nonhispanic caucasian heritage. The limited ethnic/racial composition was dictated by the fact that census data in the target area indicated that other ethnic and racial groups would represent under 10% of the sample. Given the extensive literature demonstrating a substantial relationship between patterns of alcohol involvement and ethnic/racial status, and that it was not possible to effectively control for such differences in analyses, it was decided that these groups should be excluded rather than have them contribute to error. Alcoholic subjects were recruited from local district courts using a drunk driver population. All men who 1) were convicted for DWI with a blood alcohol level (BAL) of at least .15% (150 mg/100 ml), or with a BAL of .12% which represented at least the second alcohol related arrest; 2) at the time of contact were in intact families; and 3) had biological sons between 3.0 and 6.0 years were considered potential 44 45 candidates for the study. All such candidates were asked to give their permission to have their names released to project personnel. If they agreed, the families were visited in their home by project staff, who further screened the family for suitability and, if appropriate, recruited them into the project. To date 79% agreed to have their names released; of those contacted by project staff, 91% agreed to participate. The selection criterion of BAL of .15% or multiple alcohol related arrests indicates that these men have developed significant tolerance for alcohol. Questionnaires and interviews were later administered to ensure that the father met formal diagnostic criteria for alcoholism utilizing the Feighner criteria (Feighner et al., 1972). Maternal alcoholism (alcohol abuse/dependence) was assessed, but was neither a requirement nor a basis for exclusion. Control families were recruited from the same census tract as the alcoholic families. Door-to-door survey techniques were used to locate families with children of similar age (1.5 years), sex, and sibling structure, and, when possible, family SES. These families were also screened for the presence of parental alcoholism or other drug abuse/dependence, and those reporting evidence of such difficulties were excluded. A sizable proportion of potential controls (typically fathers rather than mothers) fail the criteria for being non-alcoholic or non- drug-involved. Group assignments for purposes of this study were made based on diagnostic criteria for lifetime alcohol abuse or alcohol dependence from DSM-III-R (American Psychiatric Association, 1987). These diagnostic classifications formed the basis of family assignment into 46 one of three discrete groups: (1) neither parent was alcoholic (p - 32); (2) only the father was alcoholic (p - 74); or (3) both parents were alcoholic (a - 59). The other possible group, that is, families where the mother but not the father was alcoholic, was excluded because only eight families fell into that category. This subset was allowed in the early pilot work for the study (Zucker et al., 1983) but were later systematically excluded from participation once the full scale study (Zucker et al., 1986) began. The use of the DSM-III-R nosological system also allows the investigation of group differences as a function of severity of parental alcoholism. DSM-III-R specifies four levels of diagnostic severity based on the number of symptom areas experienced. 1co buse criteria are met with symptoms in two groups (and in some cases one, depending on the number of occurrences). Mild depepdange is defined as three symptom areas, and mpderata dapepdenae four to six symptom areas. Sevepe gapepdapga criteria specify greater than six areas or less than six with a pattern of symptoms indicating significant trouble over an extended period of time. Examination of the number of parents at each diagnostic level indicated that in some cases grouping individuals from adjacent categories was necessary to prevent extreme variations in cell sizes. For both mothers and fathers, the categories of abuse and mild dependence were combined, and for mothers the categories of moderate and severe dependence were also combined. This resulted in the following groups and cell sizes for mother diagnostic level: (1) nonalcoholic with nonalcoholic husband (11 - 32); (2) nonalcoholic with alcoholic husband (a - 74); (3) maternal alcohol abuse/mild dependence (p - 31); and 47 (4) maternal moderate/severe dependence (9 - 28). Thus, for comparisons involving maternal diagnostic level, two of the three original groups remain the same; the mother alcoholic group has been divided about equally between abuse/mild dependence and moderate/severe dependence. For paternal diagnostic level, the groups and cell sizes were as follows: (1) nonalcoholic (p,- 32); (2) abuse/mild dependence (9 - 33); (3) moderate dependence (p - 52); and (4) severe dependence (a - 48). The redefinition of groups based on fathers' severity level bears less resemblance to the preceding analyses, as the three alcoholic groups are drawn from both the father-only and both-parents-alcoholic groups. It is important to note that while the families who remain in the contrast group showed no evidence of parental alcoholism, many showed a variety of other problems (as would be expected given the community- based recruitment procedures). These problems often included other parental psychopathology and/or chronic poverty. Sociodemographic characteristics of groups based on the number of alcoholic parents in the family (zero, one, or two) are shown in Table 1. These data confirm that this is largely a working class to lower middle class sample as a whole, and suggest that the alcoholic families have fewer resources than the comparison families. While significance tests are left for the results section, these descriptive data suggest lower levels of family SES, income, and education. Twenty percent of father-only-alcoholic families and 22% of both-parents- alcoholic families reported annual family income in the range of or less 48 Table l S e o r 1 Ch racter st cs B e be Alcoho a en 5 M SD Range Child Age (months) Neither Parent Alcoholic (N) 55.16 11.64 35 - 77 Only Father Alcoholic (F) 52.61 11.71 36 - 82 Both Parents Alcoholic (B) 53.07 14.80 32 — 85 Mother Age (years) Neither Parent Alcoholic 31.94 4.22 24 - 41 Only Father Alcoholic 29.30 4.22 21 - 40 Both Parents Alcoholic 29.81 4.46 22 - 41 Father Age (years) Neither Parent Alcoholic 34.34 5.67 25 - 48 Only Father Alcoholic 31.20 4.60 23 - 47 Both Parents Alcoholic 32.00 4.96 23 - 44 Family SES' Neither Parent Alcoholic 41.00 17.35 20.25 - 80.00 Only Father Alcoholic 30.26 13.30 17.50 - 79.00 Both Parents Alcoholic 28.62 9.97 15.50 - 64.65 Mother Education (years) Neither Parent Alcoholic 13.53 2.30 7 - 18 Only Father Alcoholic 12.82 1.90 9 - 18 Both Parents Alcoholic 12.59 1.65 9 - 18 (table ggpgipaea) 49 M SD Range Father Education (years) Neither Parent Alcoholic 14.25 2.14 12 - 20 Only Father Alcoholic 12.65 2.09 7 - 19 Both Parents Alcoholic 12.47 1.76 9 - 16 Religious Preference Mother Father N F B N F B Protestant 56.3% 43.8% 54.2% 37.5% 34.2% 31.0% Catholic 28.1% 21.9% 25.4% 34.4% 13.7% 22.4% Other 9.4% 5.5% 3.4% 6.3% 5.5% 1.7% None 6.3% 28.8% 16.9% 21.9% 46.6% 44.8% Number of Marriages N F B N F B o 3.2% 1.4% 5.2% 0.0% 2.7% 1.7% 1 83.9% 90.5% 79.3% 93.8% 82.2% 69.0% 2 12.9% 8.1% 12.1% 6.3% 13.7% 27.6% 3 0.0% 0.0% 3.4% 0.0% 1.4% 1.7% (table continues) Number of Children 5 Family Income 1. Under $4,000 2. 3. 4. 5. 10. $ 4,000 $ 7,000 $10,000 $13,000 . $16,000 . $20,000 . $30,000 $50,000 - $ 7,000 $10,000 $13,000 $16,000 - $20,000 — $30,000 - $50,000 $75,000 Over $75,000 50 6.3% 65.6% 18.8% 9.4% 0.0% 0.0% 6.3% 18.8% 59.4% 14.9% 58.1% 21.6% 5.4% 0.0% 4.1% 10.8% 21.7% 28.4% 23.7% 47.5% 18.6% 8.5% 1.7% 1.7% 10.2% 23.7% 30.5% aDuncan TSEI2 (Stevens & Featherman, 1981) 51 than the weighted average poverty level for a family of four ($12,092; U.S. Bureau of the Census, 1990), compared to only 6% of controls. Family SES was lower in alcoholic families and also showed less variability compared to controls. Parents in alcoholic families were more likely to report having no religious affiliation. These data suggest that a sizeable portion of the families in the sample, particularly the alcoholic families, can be considered at-risk solely on the basis of poverty, lack of parental resources, and potentially, a lack of support from activities such as church participation. All families in the study receive monetary compensation for their participation. Currently the amount of compensation is $200 for Wave 1 data collection. Pgocedure Each participating family provides information through questionnaires, direct observation sessions, and interviews (Zucker, Noll, 6: Fitzgerald, 1986). The data are collected during the course of an eight session contact schedule which includes 18 hours of contact with project personnel. The majority of data collection takes place in The family comes to the university campus once for the family's home. Data collection is videotaping of a structured interactional task. accomplished by a trained team of graduate and undergraduate students. Data collectors are blind as to the family's status (alcoholic or control). 52 Measuges e t- Relation h The Child-Reagipg {pactigea Rapppt (CRPR) (Block, 1965) utilizes 91 statements which are administered in a Q-sort format with a forced- choice, seven-step rectangular distribution. The Q-sort methodology consists of asking the respondent to sort cards, on which the items are typed, based on how descriptive they are of their own child-rearing practices. This results in data which is different from questionnaire data in that items are rated relative to one another (ipsative data) rather than rated on an absolute scale (normative data)(Block, 1961). However, Jack Block (1957) has demonstrated empirically that ipsative data can be used in the same manner as normative data (i.e., for making comparisons across and within groups of individuals, as it is in this study). The CRPR uses the same items for both mothers and fathers, and the parents are instructed to respond according to how they actually behave with the target child. The administrator of the instrument works with the respondent to ensure that they understand the instructions and correctly complete the procedure. The item pool was constructed from three sources. First, items were written which reflected the observed behaviors found to differentiate groups of mothers with different child-handling techniques (Block, Jennings, Harvey, 6: Simpson, 1964). To supplement this, a review of the socialization literature was conducted and additional items were written to tap dimensions not included in the observational study. Finally, Block added further items based on conversations with European colleagues. This extended the coverage of the item domain and 53 reduced the culture-boundedness of the item pool. To encourage more accurate descriptions, the items are phrased whenever possible in the active voice and emphasize a behavioral orientation. The CRPR provides information on four different domains of socialization: 1) how positive and negative emotions are expressed, handled, and regulated in the child and in parent-child interactions; 2) how parental authority is conveyed, and the specific forms of discipline and control strategies used to achieve socialization; 3) the ideals and goals of the parent with respect to the child's ambitions, aspirations, and accomplishments in life; and 4) the parent's values regarding the development of the child's autonomy, independence and emergence of self (Susman et al., 1985). The CRPR has been used longitudinally to examine the relationship between parent reported child-rearing practices during early childhood and adolescent self-esteem, intelligence, aspects of moral judgement, and ego-control and resiliency (Vaughn, Block, 6: Block, 1988); adolescent creativity (Harrington, Block, 6: Block, 1987); and drug use It in adolescence (Block, Block, 5: Reyes, 1988; Shedler 6; Block, 1990). has also been used cross-sectionally to examine child-rearing attitudes in physically abusive families (Seagull et a1. , 1988; Susman et al. , 1985; Trickett 6: Susman, 1988); in families with depressed parents (Stoneman, Brody, 6 Burke, 1989; Susman et al., 1985); and in families who have a child with cancer (Davies, Noll, 6: Bukowski, 1991; Davies, Across a broad variety of N011, DeStefano, Bukowski, 6: Kulkarni, 1991). populations and cultures (Lin & Fu, 1990; Seagull et a1. , 1988), investigations using the CRPR support the sensitivity of this measure to 54 evaluate current parental socialization practices as well as the ability of this approach to predict future adaptations of children. There have been three primary approaches used in research with the CRPR: (1) use of individual item scores as variables; (2) scale building through factor analysis; and (3) development of prototype sorts through expert ratings which can then be compared to the respondents' distributions. Early efforts at establishing expert consensus on the dimensions that the literature suggests are important in alcoholic families met with little success, so the prototype sort approach was rejected. Factor analysis was chosen because it offered the possibility of tapping higher order constructs than would be possible through analysis of individual items. CRPR research using factor-analytically derived scales has typically made use of the scales that Block (1965) provided in the manual for the instrument. It was not specified exactly what the sample was for this procedure, but it seems likely that it was a sample of "normal" mothers, as was the case with the other results provided in the This analysis yielded 21 factors, 13 of which contain fewer manual . No data were provided on the reliability of these scales. than 4 items. Confirmatory factor analysis of the CRPR on a sample from this project revealed that Block's factor structure was not replicated for Indeed, Block either the mothers or the fathers (Davies et al. , 1991). (1965) argued that there is no reason to assume that a given factor structure will be appropriate beyond the sample from which it was Earlier work (Davies et al. , 1991) has established a factor derived. Exploratory factor analyses structure for a sample from this project. (principal components extraction, varimax rotation) were conducted 55 separately for mothers and fathers (N-160 families, both alcoholic and Examination of the scree plot (Cattell, 1966) indicated that control). It nine factors should be extracted for mothers and eleven for fathers. was decided that factors would be considered adequate if they displayed acceptable internal consistency (alpha > .60) and accounted for greater than five percent of the variance after rotation. Applying these additional criteria reduced the number of factors to two each for mothers and fathers. This factor structure replicated when only items with loadings greater than .30 on the final factors from the first analyses were included. This finding suggests that this is a valid factor structure rather than an artifact of the method (e.g., items that loaded together on the first analyses may have done so not because they were highly related to one another but because they were unrelated to all the other items). There was no item overlap on the factors. For mothers these two factors were ”Positive vs. Negative Affective Parenting" (22 items, alpha - .82) and "Authoritarianism vs. Child-centeredness" (18 items, alpha - .78). For fathers they were "Positive vs. Disengaged Affective Parenting” (35 items, alpha - .91) and "Authoritarianism vs. Liberal Parenting" (10 items, alpha - .69). Thus for both parents these two factors tap the general domains of support and control, the two major tasks of parenting (Sears, Maccoby, 6: Levin, 1957). Despite this similarity, there are sufficient differences The to warrant the use of separate factors for mothers and fathers. negative roles are different in content and reflect differences in In addition, perceived roles and responsibilities of the two genders. despite the general parallelism of content, there is only 53% item 56 overlap for the Positive Affective Parenting dimension, and only 57% item overlap for the Authoritarianism dimension. It is interesting to note that the general domains tapped by these factors bear considerable resemblance to higher-order factors created by combining factors from Block's factor structure using both empirical (Deal, Halverson, 6: Wampler, 1989) and conceptual (Kochanska, Kuczynski, 6: Radke-Yarrow, 1989) groupings. The interrelationships of these scores have been examined in a subset of the current sample (Davies et al., 1991). Mothers' Affective Parenting and Authoritarianism scores correlated -.42, while the same Mothers' and fathers' scores for scores for fathers correlated -.22. Authoritarianism were positively correlated (r, - .25), while Affective Parenting scores were unrelated (; - .02). Two test-retest reliability studies have been conducted with this instrument (Block, 1980). The first involved 90 undergraduate students (most of whom presumably were not parents) enrolled in a child psychology course. They completed the CRPR at the beginning and end of The average correlation the course, a time span of eight months. The second study between the two tests was .707 (range .38 to .85). used 66 Peace Corps volunteers as subjects, who completed the third- person version of the CRPR at the beginning and end of their duty, an interval of three years. The third-person version has the same items as the first-person, but items are phrased so that they apply to the respondent's own parents. Maternal descriptions showed an average correlation of .64 (range .04 to .85) while paternal descriptions had an average correlation of .65 (range .13 to .85). 57 In the nine-year longitudinal study with the CRPR discussed above (Roberts et a1. , 1984) significant correlation coefficients were obtained for 73% (66/91) of the items for mothers and 56% (51/91) for fathers. This reveals considerable constancy/stability in response, especially when taking into account that most of the changes in responses were in keeping with developmentally appropriate changes in child-rearing (e.g. , increasing autonomy for the child, decreasing parental control of the child). Although further testing of the instrument's reliability would be desirable, particularly with samples of parents, the existing evidence suggests that there is acceptable stability and consistency for the ratings. Much of the validity attributable to the CRPR comes from the straightforward and logical way in which it was developed. Its validity has been assessed experimentally by having the participants in the original observational study from which the CRPR was derived fill out the instrument four years later. Of the 112 mothers in the original study, 76 agreed to take part in the follow-up. Considerable coherence was found between the self-report descriptions provided by mothers and the observed behaviors recorded earlier (Block, 1980). The similarity between CRPR self-reports and observed parenting behaviors has recently been replicated cross-sectionally (Kochanska et a1. , 1989) and longitudinally (Kochanska, 1990) in studies which limited their scope to CRPR content to that which could most easily be operationalized. Less focused studies on the CRPR and similar instruments have yielded less positive results (Holden 6 Edwards, 1989). In addition to the factor scores obtained for each parent on the it is also possible to compute a within-couple Wm CRPR , 58 This is obtained by correlating the ipdegg (PAI) across all items. mother's and father's responses to each item (Block, Block, 6 Morrison, The PAI has been shown to be related to the development of 1981). competence across different developmental trajectories for boys and girls (Block et al., 1981; Vaughn, Block, 6 Block, 1988), to positive characteristics of parent-child interaction (Gjerde, 1988), and to marital survival (Block et a1. , 1981). Recent data have suggested that high PAI scores are associated with "qualities that therapists and other 'healthy' family professionals generally ascribe to well-functioning, families" (Deal et al., 1989, p. 1031) and that the PAI can be conceptualized as a general measure of parental effectiveness in the family. The PAI constitutes our closest approximation to measuring the Cohen parenting alliance (Cohen 6 Weissman, 1984) in these families. and Weissman define the parenting alliance as that component of the marital relationship which is distinct from the libidinal relationship. It is in actuality more complex than the simple agreement of the couple about child-rearing beliefs, encompassing the respect of the spouses for each other's contribution and the way the parents take care of the business of parenting. However, use of the PAI provides us with information about an aspect of the parents' relationship which is distinct from the spousal relationship. Demographic variables The demographic variables of interest here are family socioeconomic status (SES), family income, years of education for each The index of the age of the parent, and the age of the child. parent, SES used here is the Revised Qanpap §ocioecpnomic 12962: (TSEIZ; Stevens 59 6 Featherman, 1981). This measure was selected based on work by sociologists suggesting that occupation-based measures represent a more contemporary indicator of SES that is sensitive to changes in occupational attainment (Featherman 6 Hauser, 1977; Mueller 6 Parcel, 1981; Nock 6 Rossi, 1979). Family SES is defined here as the father's SES when his is higher than the mother's, or the average of the mother's and the father's when hers is higher. Pare ta 3 cho atholo Alcohol uselproblems. Several instruments are administered individually to both parents which provide information about current use of alcohol and problems associated with such use, as well as about drinking history. Parents are given an extensive Drinking and Drug History (DDH; Zucker, Fitzgerald, 6 Noll, 1990), the Short Form of the Michigan Alcoholism Screening Test (SMAST; Selzer, 1975), and are also asked about their drinking practices during the Diagnostic Interview Schedule (DIS; Robins, Helzer, Croughan, 6 Ratcliff, 1981; Robins et a1. 1985), an extensive clinical interview. The Lifetime alcohpl ngblems Score (LAPS; Zucker, 1991) incorporates information on the primacy (onset), variety, and life invasiveness of problems associated with drinking based upon data from the DDH and DIS. LAPS consists of three component subscores: (a) the primacy component, which is the squared inverse of the age at which the respondent reported first drinking enough to get drunk; (b) the variety component, which is the number of areas in which drinking problems are reported, adjusted for current age; and (c) the life percent component, which is the interval between the earliest and most recent drinking problems, again adjusted for current age. Scores are standardized 60 separately for males and females within our project sample. This measure is unrelated to current alcohol consumption in problem drinking samples and has been shown to be a valid indicator of differences in long-term severity of drinking difficulty in a wide variety of areas (Zucker, 1991). Information from the DDH is used to compute an index of current drinking which reflects the gnangity, fragnengy, and vagiabiiigy of their drinking. The measure used here is a derivative of Cahalan, Cisin, and Crossley's (1969) Alcohol Consumption Index QFV measure called QFV-R (Zucker 6 Davies, 1989). This measure utilizes the rationale of the original scoring system, but rather than combining the Quantity-Variability classification with the Frequency classification to yield a five-category classification, the score is obtained by multiplying the QV class times the approximate number of drinking episodes per year (based on the reported average frequency). This yields a 0 to 21,000 score which is then subjected to a logarithmic transformation (base ten). This revision of the scoring system greatly increases the sensitivity of the measure and so increases the information that the score provides about the relative level of current This is necessary given that according to the original drinking. scoring system, almost all the alcoholic men in the current sample would be classified as "heavy drinkers" or ”abstainers" (approximately one fourth of the alcoholic men were "dry" at the time of first-wave data collection (Davies, 1989)). Depression. Several measures of parental depression are Self-reported depression is measured using the Short Form of collected. This version the gag}; Qapzaasipn Inventory (BDI; Beck 6 Beck, 1972). 61 contains 13 groups of statements concerning different areas of functioning known to be affected by depression (e.g., appetite, sleep habits, mood). The respondent indicates the statement in each group that best describes how he feels on that day. Scores on the long and short forms of the BDI have been found to correlate between .89 and .97 (Beck et al., 1988). However, there is some evidence from factor analytic studies that the short form may pick up on one cognitively oriented symptom dimension, while the long form reflects more than one symptom dimension, including some noncognitive symptom clusters (Beck, Steer, 6 Garbin, 1988). The internal consistency of the long form of the B01 has been examined in at least 25 studies (Beck et al., 1988). For psychiatric populations, the mean coefficient Alpha was .86, with a range of .76 to .95. For nonpsychiatric samples, the mean coefficient Alpha was .81, ‘with a range of .73 to .92. Beck et a1. (1988) claim comparable internal consistency for the short version, but no data are presented. Examinations of the validity of the B01 have shown impressive results in the areas of content, concurrent, discriminant, and construct validity. The Hamilton Rating Scale £0; Depression (HRSD; Hamilton, 1960), an instrument for the clinical rating of depression, is done following the DIS by the clinician who conducts the interview. This rating covers a variety of behavioral, affective, somatic, and psychological dimensions associated with depression, and the score is based on the subject's responses, as well as the clinician's judgments. The cilinician makes ratings of the level of depression at the current time arui for the time when the person was most depressed (worst-ever). Initerrater reliabilities have ranged from .80 to .90 (Hamilton, 1969). 62 Acceptable interrater reliability (Bech et al., 1975) and criterion validity (Carroll, Fielding, 6 Blashki, 1973) have been demonstrated. The HRSD was developed as an observationally-based rating scale to be completed by a treatment provider in an institutional setting. Recent work (Whisman et al., 1989) has demonstrated that graduate students administering the instrument in the context of the DIS displayed excellent agreement with psychiatrist ratings. Interrater reliability on the current project based on ratings done after a review of DIS protocols and audiotapes demonstrated acceptable reliability. Ratings for current depression correlated .78, and ratings for worst- ever depression correlated .80. Antispcial behavior. Antisocial behavior (ASB) on the part of the parents is measured using the Antisocial Behavior Checklist (Zucker 6 Noll, 1980), a 46-item revision of an earlier version used in the Rutgers Community Study (Zucker 6 Barron, 1973). The questionnaire asks the frequency of the respondent's participation in a variety of delinquent, criminal, and antisocial activities (e.g., talking back to parents, having an affair, committing armed robbery) and includes items relevant to childhood and adulthood. A series of reliability and Validity studies with samples ranging from male and female college students to male and female state prison inmates has shown that the instrument has adequate test-retest reliability (.91 over four weeks) and internal consistency (coefficient alpha - .93) (Zucker 6 Noll, 1930). It has also been shown to differentiate among groups with major histories of antisocial behavior (inmates) versus individuals with minor Offenses in district court versus university students (Noll 6 Zucker, 63 1980). Separate scores are computed for antisocial involvements in childhood and adulthood. Overall psychosociai functioning. Each parent's overall level of psychosocial functioning in the past year is rated using Aaia_y of DSM-III-R (American Psychiatric Association, 1987). Ratings are done by the clinician who conducted the Diagnostic Interview Schedule based on information obtained about the subject's recent experiences in the realms of social, occupational, and leisure time functioning. Ratings are done on a lOO-point scale which ranges from grossly impaired to superior. Reliability of ratings, assessed via the procedure described above for the HRSD showed acceptable interrater reliability (; - .68). Child behavior The Chiid Behaving Checklisg (CBCL; Achenbach 6 Edelbrock, 1983) is perhaps the most carefully standardized and widely used objective measures of child adjustment. Based on parent reports of the frequency of discrete behaviors, the instrument yields standardized scores on eight narrow band subscales and two broad band subscales (i.e., externalizing and internalizing behavior problems), as well as information on social competence. Test-retest reliability averages .87 and extensive criterion-referenced validity studies have been conducted (Achenbach 6 Edelbrock, 1983). This study will utilize the Total Behavior Problems Score, which incorporates both internalizing and externalizing behaviors. F ' env ro ent The Eaniiy_§nyi;pnnang_§aaia (FES; Moos 6 Moos, 1976, 1981) is an empirically based taxonomy of family social environments as perceived by the family members themselves. It is reportedly the most widely used 64 measure of family characteristics (Oliver, May, 6 Handal, 1988). Based on true-false responses to a 90-item questionnaire, the FES provides ten subscale scores in the content areas of relationships, orientation towards personal growth, and system maintenance. The subscales have demonstrated internal consistency ranging from .64 to .79, test-retest reliability ranging from .68 to .86, and average subscale intercorrelations around .20, indicating that they measure distinct, though somewhat related aspects of family social environments (Moos 6 Moos, 1976). Earlier analyses of FES data in the current project revealed significant relationships between levels of current alcohol consumption and the conflict and cohesion subscales (Reider, 1987). Differences between alcoholic and nonalcoholic families have been demonstrated on the conflict and moral-religious subscales (Zucker, Weil et al., 1984). The FES variables used in this study will be three higher order, factor analytically derived scales (Kronenberger 6 Thompson, 1990) which are composites of the original subscales. These factors were derived based on data from.a sample of families with chronically ill children, ‘but have been shown to generalize to normal population samples of both adults and adolescents. The three factors are (1) Supportive, the sum of Cohesion, Expressiveness, Independence, Active-Recreational Orientation, and Intellectual-Cultural Orientation; (2) Conflicted, composed of the Conflict score minus COhesion and Organization; and (3) Controlling, which is the sum of Control, Achievement Orientation, and.Moral-Religious Emphasis, minus Independence. This three factor solution accounted for over 50% of the variance in all three samples (Kronenberger 6 Thompson, 1990). The Supportive 65 and Conflicted factors differentiated families of chronically ill children with and without behavior problems. The authors point out that the Conflicted and Controlling factors are in keeping with existing theoretical models of family functioning involving two factors (e.g., Olson, Sprenkle, and Russell's (1979) Cohesion and Adaptability). They go on to suggest that the identification of a third factor related to supportiveness may allow a bridge to the literature on the buffering characteristics of social support (e.g., Cohen 6 Wills, 1985). The Family Cgisis List (FCL; Patterson, 1982) is a 40-item list of family events across a variety of areas from minor daily hassles to major life troubles and tragedies. The parent is asked to endorse all events which have occurred in the immediate family in the last six months. The FCL was developed at the Oregon Social Learning Center to document the association between family stress and patterns of family coerciveness. The count of number of items endorsed will be the measure used for this study. Data Analytic Strategy Overall correlational relationships between the variables (both independent and dependent) in the study will be examined first. Then main.effects for number of alcoholic parents and severity of parental alcoholism on the demographic, parental psychopathology, and family environment variables will be explored using multivariate analysis of 'variance (MANOVA). In addition to providing information important in its own right, these results will permit these differences to be controlled for in subsequent analyses. MANOVA will then be used to examine whether there are differences ion.the child-rearing variables as a function of these group divisions. 66 If significant differences are discovered, analysis of covariance will be used to control for variables which the earlier analyses indicated differed between groups and which are known to be related to child- rearing practices (e.g., SES, parental depression). CHAPTER III Results Missing Data and Outliers Before beginning analyses, a thorough screening was done for missing data and outlying variable values. All missing data were estimated via regression analyses on the available data. For no variable was data missing on more than five subjects (or 3% of the sample). Outliers were defined as nonadjacent values falling outside a normal distribution superimposed on the frequency distribution histogram. Each outlying value was assigned a value adjacent to the closest non-outlying value while maintaining the rank order of subjects on each variable. Desc tive tatistics a d Re iabili ies Tables 2 and 3 display descriptive statistics and internal ‘*—) consistency estimates for independent and dependent variables for mothers and fathers respectively. Coefficient alpha estimates of internal consistency were generally acceptable and ranged from moderate to very good. For mothers, an alpha less than .70 was obtained only for the Controlling factor from the FES. For fathers, alphas less than .70 were obtained for CRPR Authoritarianism and, again, for FES Controlling. None of the variables showed internal consistency estimates less than .60. Demo c d v ual rent n C d a d e ce ved Fam WM Correlations between the independent variables for the sample as a whole will be examined first. All correlation coefficient significance ‘values in this study have been adjusted using Holm's (1979) procedure to 67 Table 2 s t ve Stat st cs and Inter Co 51 tenc Coef icient ha 0 v 1 F i and Social a i n V ab e o e alpha 5 SD Range t v dua ctionin Lifetime Alcohol Probs. - 9.83 2.04 7.75 15.49 Current Drinking (QFV-R) - 1.70 1.09 0 4.20 Beck Depression Inventory .77 3.39 3.46 0 16 Hamilton Depression-Current .88 6.44 6.57 0 28 Hamilton Depression-Worst .89 16.12 10.59 0 50 Antisocial Behavior .85 12.28 7.92 1 43 Axis V of DSM-III-R - 58.48 11.89 29 82 Ennily gnd Child Functioning Family Environment Scale Supportive .79 28.90 6.09 8 42 Conflicted .81 9.72 4.92 1 24 Controlling .62 18.48 4.20 4 27 Family Crisis List .77 16.74 6.24 2 35 Total Child Behavior Problems .94 32.54 16.18 1 75 figcinlizgtion Variables Nurturance .82 124.04 12.09 79 144 Authoritarianism .75 52.72 11.63 30 91 Agreement Index (Dyadic) - .52 .15 .09 .82 Table 3 esc t ve Statistics and nte nd vidu 1 Famil Co 5 ste c nd Sociali ation V riables o Fathers oeffic ent 1 h of alpha ! SQ Range Father Individual Functioning Lifetime Alcohol Probs. - 9.92 2.43 5.31 15.82 Current Drinking (QFV-R) - 2.21 1.42 0 4.32 Beck Depression Inventory .77 2.78 2.91 0 13 Hamilton Depression-Current .86 5.10 4.98 0 23 Hamilton Depression-Worst .89 13.72 9.88 0 43 Antisocial Behavior .92 20.86 13.37 2 70 Axis V of DSM-III-R - 55.56 11.65 30 82 anily and Child Functioning Family Environment Scale Supportive .73 27.74 5.35 12 41 Conflicted .80 9.27 4.74 1 23 Controlling .62 18.87 4.17 6 27 Family Crisis List .83 15.01 6.52 3 35 Total Child Behavior Problems .94 29.31 15.52 2 68 Sociaiization Variabies Nurturance .89 204.04 12.42 160 226 Authoritarianism .67 33.76 7.90 19 56 70 control the overall Type I error rate. This procedure maintains the overall Type I error rate at alpha-.05, but affords more statistical power than the traditional Bonferroni procedure (Holland & Copenhaver, 1988). intercorrelations. Intercorrelations between the demographic variables are shown in Table 4. Significant positive correlations were obtained between family SES, family income, mother and father age, and mother and father education. Child age was not significantly related to any of these variables. Intercorrelations between the psychopathology variables for mothers and fathers are shown in Table 5. For mothers, a general pattern emerges indicating that the domains of alcohol-related difficulties, depression (self-reported and clinician rated), antisocial behavior (ASB), and ratings of impaired psychosocial functioning are consistently and positively related. The main exception to the pattern is the QFV-R index of current drinking, which is only significantly associated with lifetime alcohol problems and involvement in ASB, suggesting that for these women drinking may be serving an antisocial function (as opposed to drinking which is driven by negative affect). The only other exceptions are a nonsignificant relationship between antisocial involvements and the current Hamilton depression rating; and a nonsignificant relationship between LAPS and current depressed mood on the BDI. The consistent pattern of results is particularly striking given that these ratings include four self-report measures and three clinician ratings. Table 4 71 Intezgggrelations (Pearson R) between Demographig Vagiabies (fl-165 families) 1. Family SES .41 * 2. Family Income 3. Mother Age 4. Mother Education 5. Father Age 6. Father Education 7. Child Age .31 * .51 * .30 * .57 * .03 .42 * .41 * .35 * .41 * .00 .29 * .70 * .32 * .12 .20 + .51 * .04 .27 * .12 * n,< .05, + n < .10, two-tailed. Holm's procedure. Significance values adjusted via 72 Table 5 intercorrelations (Pearson R) between Parental Psychopathology Variables (3-165 families) nothgrs 2 3 4. 5 6 7 1. Lifetime Alcohol Probs. .42* .12 .22* .26* .57* -.28* 2. Current Drinking (QFV-R) -.14 .11 .00 .35* -.08 3. Beck Depression Inv. .27* .33* .32* -.39* 4. Hamilton Depr.-Current .67* .17 -.56* 5. Hamilton Depr.-Worst .29* -.56* 6. Antisocial Behavior -.37* 7. Axis V of DSM-III-R Fathers 2 3 4. 5 6 7 1. Lifetime Alcohol Probs. .10 .35* .39* .43* .63* -.51* 2. Current Drinking (QFV-R) .19+ .21* .00 .11 -.21* 3. Beck Depression Inv. .39* .40* .36* -.47* 4. Hamilton Depr.-Current .67* .33* -.61* 5. Hamilton Depr.-Worst .32* -.52* 6. Antisocial Behavior -.46* 7. Axis V of DSM-III-R * n < .05, + n < .10, two-tailed. Significance values adjusted via Holm's procedure. 73 For fathers an impressively consistent picture again emerges suggesting important commonalities between the four domains and seven 'variables tapped, and again demonstrating an exception for the measure of current drinking. Correlations between mother and father psychopathology variables are shown in Table 6. For all seven variables, the mother and father scores were significantly related to one another. Mothers' histories of alcohol related problems were associated with six of the seven father variables (the exception being the worst-ever Hamilton rating). Mothers' self-reported depressed mood was associated with fathers' drinking problems, current and worst-ever Hamilton depression, and lower psychosocial functioning. Mother's current and worst-ever Hamilton depression ratings were significantly negatively correlated with father's Axis V rating, and the worst-ever rating was significantly related to his BDI score and current Hamilton score. Mothers with more involvement in ASB had husbands with increased alcohol difficulties, higher levels of current depression, and lower overall functioning. Mother's Axis V ratings were negatively related to father LAPS, BDI, and both Hamilton ratings. Axis V ratings of the couples were particularly highly correlated. These results provide considerable evidence that parental psychopathology occurs in the context of a disturbed family system and that there is little specificity with regard to particular symptom patterns in these families. That is, drinking specific and nonspecific symptoms and overall functioning are all related to one another. 74 Table 6 gomrelggions (Pearson 3) between notngr and Eatbe; Psychopathology ‘Variables (fl-l65 families) Fathers Mothers 1. 2. 3. 4. 5. 6. 7. 1. Lifetime Alc. Probs. .30* .28* .30* .27* .21 .23+ -.33* 2. Current Drinking .04 .46* .01 .01 -.12 -.01 -.06 3. Beck Depression .23+ .10 .41* .28* .27* .19 -.29* 4. Hamilton-Current .04 .16 .19 .23* .12 .11 -.34* 5. Hamilton-Worst .06 ' .09 .24+ .26* .25* .11 -.33* 6. Antisocial Behavior .31* .19 .16 .27* .17 .31* -.33* 7. Axis V -.23+ -.03 -.41* -.30* -.26* -.20 .61* * n < .05, + n < .10, two-tailed. Significance values adjusted via Holm's procedure. 75 Intercorrelations between parent ratings of family and child functioning for mothers and fathers are shown in Table 7. For mothers, reports of overall child behavior problems were significantly associated with all four measures of the family environment (negatively with the FES Supportive factor and positively with the FES Conflicted and Controlling factors and the level of family crises). Scores on the FES Supportive factor were strongly negatively correlated with the FES Conflicted factor. Finally, scores on the Family Crisis List were positively associated with the FES Conflicted and Controlling factors. The FES Controlling factor was unrelated to the other two FES factors. The pattern was much the same for fathers, although the strength of associations were somewhat weakened. The FES Controlling factor was not significantly related to any of the other variables, including child behavior problems. However, ratings of behavior problems were significantly associated with the FES Conflicted factor and the Family Crisis List. Again, FES Supportive and FES Conflicted were highly negatively correlated, and the reported level of environmental conflict was positively associated with the number of crises. Correlations between mother and father reports of family environment and child behavior are reported in Table 8. Mother-father correlations were significantly positively related for all five variables, although the highest correlation was only .46 (FES Conflicted), indicating considerable non-shared variance. For both combinations, each gender's reports of Supportive environments were negatively related to the other's reports of a conflicted atmosphere. Mothers' FES Conflicted scores were also negatively related to fathers' 76 Table 7 nter e at o 5 ea so R etwee a ent R t n s 0 am nd Child Ennggioning (fl-165 families) Mothegg 2 3 4. 5 1. FES Supportive -.53* -.08 -.14 -.28* 2. FES Conflicted .09 .39* .30* 3. FES Controlling .19+ .21* 4. Family Crisis List .32* 5. CBCL Total Child Behavior Probs. Fathers 2 3 4. 5 1. FES Supportive -.47* -.15 -.10 -.17+ 2. FES Conflicted .02 .36* .35* 3. FES Controlling .08 .03 4. Family Crisis List .25* 5. CBCL Total Child Behavior Probs. * n < .05, + 2 < .10, two-tailed. Significance values adjusted via Holm's procedure. 77 Table 8 Cnrrelations (Pearson R) between Mother and Father Ratings of Family and 1d F tion n (fl-165 families) Fathers gothers 1. 2. 3. 4. 5. 1. FES Supportive .38 * -.24 * -.02 -.ll -.03 2. FES Conflicted -.28 * .46 * .03 .25 * .08 3. FES Controlling -.04 .14 .39 * .11 .05 4. Family Crisis List -.10 .19 .03 .36 * .14 5. CBCL Total Child Beh. Probs. -.27 * .21 -.02 .17 .36 * * n < .05, two-tailed. Significance values adjusted via Holm's procedure. 78 reports of family crises. Finally, mothers' reports of child behavior problems were negatively associated with fathers' FES Supportive factor. Table 9 shows correlations between the family demographic variables and mothers' reports of personal and family functioning. Ages of mother, father, and child were unrelated to any of these variables, while the mothers' alcohol problem history, level of current drinking, and reports of conflicted and controlling environments and family crises were unrelated to family demography. Higher levels of family SES, income, and parental education were associated with lower levels of maternal antisocial behavior, higher levels of psychosocial functioning, and reports of a more supportive home environment. In addition, higher family income was associated with lower reported levels of depressed mood and lower ratings of current and worst-ever depression. Maternal education was negatively associated with ratings of current and worst- ever depression. The corresponding relationships between paternal reports and demography are shown in Table 10. Again, the ages of all three family members were unrelated to the other variables except that older fathers reported less antisocial involvements. No demographic correlates were discovered for fathers' level of current drinking, clinician ratings of worst-ever depression, the FES Controlling factor, or the level of family crises. Higher family SES, income and paternal education were associated with fewer lifetime alcohol problems and antisocial involvements and higher psychosocial functioning and reports of a more supportive home environment. Higher family income was also associated with a lower level of reported depressed mood and lower clinician Table 9 r e i n Pea son betwe n e 0 Va ables nd ot ers' Esygnongthology nnd Fanily Functioning Varinbles (fl-165 Families) Mother Child SES Income Age Educ. Age e a un tioni Lifetime Alc. Probs. .14 .15 .02 -.17 -.06 Current Drinking (QFV-R) .01 .00 .04 -.08 .05 Beck Depression .16 .27 .09 -.21 .02 Hamilton-Current .17 .27 .07 -.26 * -.20 Hamilton-Worst .17 .25 .05 -.32 * -.12 Antisocial Behavior .32 .34 .21 - 36 * .00 Axis V .33 .44 .18 .42 * .06 i a d h n t oni FES Supportive .25 .24 .12 .35 * -.03 FES Conflicted .12 .09 .01 -.12 .09 FES Controlling .01 .14 .04 -.14 -.01 Family Crisis List .00 .09 .03 -.07 -.05 Child Behavior Probs. .12 .22 .18 -.08 .10 * n < .05, two-tailed. procedure. Significance values adjusted via Holm's 80 'Table 10 Cor t ns P arson betwee emo r h c Va ’ab es and Fathe s' Esychonntnology nnd Ennily Ennggioning Variables (fl-165 Families) Family Father Child SES Income Age Educ. Age te a ct ni Lifetime Alc. Probs. -.29 * -.28 * -.18 -.30 * -.10 Current Drinking (QFV-R) -.09 -.01 -.13 -.05 .04 Beck Depression -.21 -.27 * -.01 -.17 .05 Hamilton-Current -.15 -.26 * .01 —.14 -.02 Hamilton-Worst v.16 -.16 .06 -.ll -.04 Antisocial Behavior -.32 * -.36 * -.21 -.29 * -.01 Axis V .38 * .44 * .07 .34 * -.03 Ennily and Child anctioning FES Supportive .25 * .26 * .11 .22 .03 FES Conflicted -.23 -.16 .01 -.12 .03 FES Controlling .02 -.15 -.07 .03 .02 Family Crisis List -.14 .02 -.12 .05 -.05 Child Behavior Probs. -.22 -.19 -.10 -.18 .11 * n < .05, two-tailed. Significance values adjusted via Holm's procedure. 81 ratings of current depression. Correlations between mothers' reports of the family environment and psychopathology variables are shown in Table 11. Mothers' reports of less supportive and more conflicted environments, more family crises, and more child behavior problems were associated with higher levels of depressed mood, antisociality, and impaired psychosocial functioning. Reports of a conflicted environment and more crises were associated with higher ratings of current and worst-ever depression, while more supportive environments were associated with lower levels of lifetime depression. The mothers' history of alcohol problems was positively associated with the levels of family crises and child behavior problems. Ratings of a more controlling environment were associated with lower Axis V ratings. The corresponding correlations between paternal psychopathology and paternal family ratings are shown in Table 12. Fathers' reports of degree of control in the environment showed no significant associations with any of the paternal functioning variables, while level of current drinking was unrelated to all of the family environment measures. Paternal reports of greater conflict, crises, and child behavior problems were associated with higher levels of lifetime alcohol problems, antisociality, current depressed mood, ratings of current and historical depression, and psychosocial functioning. Paternal reports of supportiveness were negatively associated with level of depressed mood and positively associated with Axis V ratings. 82 Table 11 o e t Pea so etwee e ' o ami nd Ch d W (Ii-165) Mothers' FES FES FES Supp. Confl. Control. FCL CBCL Lifetime Alc. Probs. -.11 .17 .02 .24 * .27 * Current Drinking (QFV-R) .03 .00 -.04 .10 .08 Beck Depression -.46 * .49 * .16 .26 * .33 * Hamilton-Current -.19 .26 * .20 .23 + .15 Hamilton-Worst -.22 + .29 * .19 .35 * .22 + Antisocial Behavior -.36 * .28 * .12 .31 * .34 * Axis V .34 * -.23 * -.28 * -.28 * -.30 * * n < .05, + n < .10, two-tailed. Significance values adjusted via Holm's procedure. 83 Table 12 C o s earson R between a e ' t s o Famil n d W (ti-165 Fathers' FES FES FES Supp. Confl. Control. FCL CBCL Eaterngl Functioning Lifetime Alc. Probs. -.07 .31 * -.16 .27 * .29 * Current Drinking (QFV-R) -.06 .16 -.06 .ll .11 Beck Depression -.32 * .48 * .01 .33 * .40 * Hamilton-Current -.16 .26 * .08 .23 + .16 Hamilton-Worst -.09 .28 * .04 .28 * .20 Antisocial Behavior -.07 .35 * -.09 .27 * .29 * Axis V .23 + -.36 * -.05 -.28 * -.21 * n < .05, + n < .10, two-tailed. Significance values adjusted via Holm's procedure. 84 o co ar so 5 as function 0 numbe o a coho c aren s. For all of the group comparison analyses, overall type I error was controlled by first conducting a multivariate analysis of variance (MANOVA) across the variables in each domain. A significant effect at this level indicates that the groups do show differences across those variables and allows further univariate analyses of variance (ANOVAs) to be conducted and interpreted. Where the MANOVA was nonsignificant, the univariate ANOVAs were conducted due to the exploratory nature of the study, but any significant univariate results were interpreted with caution. Duncan Multiple Range Tests (DMRT) were conducted to determine which specific group means differed significantly from one another. As indicated in Table 13, the MANOVA for the demographic variables was significant across the groups determined by the number of alcoholic parents in each family. Univariate ANOVAs indicated significant group differences on mother and father age, father education, and family SES and income. Maternal education showed a nonsignificant trend, while there was no evidence of group differences for child age. For all of the variables showing significant ANOVAs (i.e., mother and father age, father education, and family SES and income), the DMRT indicated that the neither-parent-alcoholic group was significantly higher than either of the groups where the father was alcoholic. Thus, the parents in families recruited as controls had greater resources and achievement and the parents were older than in the families with an alcoholic father (whether or not the mother was also alcoholic). This effect occurs despite the labor intensive census tract approach used in locating controls, and the systematic effort to achieve demographic 85 Table 13 ,~;, 0 -s . S anda d Dev : _o; o 0emo:ra- ic Va i:b1es . ybe g: Alcnnolig Parents (N—165 Families) Alcoholic Parents Neither Father Both Duncan (n-32) (BF74) (n-59) Multiple (N) (F) (B) 2? Range Test (Multivariate E(14,312) - 2.16 **) Mother Age 31.94 29.30 29.81 4.28 * N > B,F (4.22) (4.22) (4.46) Father Age 34.34 31.20 32.00 4.52 * N > B,F (5.67) (4.60) (4.96) Child Age 55.16 52.61 53.07 .45 - (11.64) (11.71) (14.80) Mother Education 13.53 12.82 12.59 2.58 + N > B (2.30) (1.90) (1.65) Father Education 14.25 12.65 12.47 9.33 *** N > F,B (2.14) (2.09) (1.76) Family SES 41.00 30.26 28.62 10.06 *** N > F,B (17.35) (13.30) (9.97) Family Income 7.56 6.47 6.28 4.59 * N > F,B (1.26) (2.05) (2.25) *** n < .001, ** n < .01, * n (.05, + n < .10. 'Univariate Analysis of Variance with (2,162) degrees of freedom. 86 comparability. Table 14 shows group comparisons for the maternal individual functioning variables. The MANOVA for these comparisons was highly significant. As we would expect from the way the groups were formed, the alcoholic mothers showed large differences on the drinking specific variables (lifetime problems and current consumption) compared to both of the other groups. Large differences were also apparent for involvement in antisocial behavior, although now all three groups were significantly different from one another on the DMRT, with the nonalcoholic women married to alcoholic men falling between the other two groups. Alcoholic mothers showed levels of worst-ever depression that were significantly higher than women who were nonalcoholic but married to alcoholic men. Finally, the alcoholic women showed lower Axis V ratings than the women in the neither parent alcoholic group. In general, then, the alcoholic women appeared to be functioning less optimally than the other women in the sample, whether or not their husbands were alcoholic. Group comparisons on paternal functioning variables are shown in Table 15. Again, the MANOVA was highly significant, and large differences were obtained for the drinking specific variables. Both groups of alcoholic fathers reported more drinking problems than the nonalcoholic men. Alcoholic men married to alcoholic women reported higher levels of current drinking than those married to nonalcoholic women, and both groups reported greater current consumption than controls. The two groups of alcoholic men were higher than controls and not significantly different from one another on ratings of current and Table 14 Mea c res and Standard Deviat o s 87 O aternal Ps cho a ho o Vngigbles by Nnnbe; of Alcoholig Parents (N—165 Families) Alcoholic Parents Neither Father Both Duncan (n-32) (n-74) (BF59) Multiple (N) (F) (B) E? Range Test MQEhglfi (Multivariate E(l4,312) - 15.27 ****) LAPS 8.38 8.76 11.96 131.52 **** N,F < B (.67) (.96) (1.75) QFV-R 1.13 1.51 2.24 14.93 **** N,F < B (.84) (.91) (1.19) BDI 2.75 3.41 3.71 .80 - (2.90) (3.60) (3.56) HRSD-Current 5.22 5.74 7.97 2.61 + - (6.52) (6.10) (7.00) HRSD—Worst 14.97 14.46 18.83 3.11 * F < B (11.65) (8.76) (11.67) ASB 6.91 10.49 17.44 29.37 **** N < F < B (4.63) (5.87) (8.67) AXIS V 63.47 58.81 55.36 5.13 ** N > B (13.04) (10.90) (11.65) **** n < .0001, ** n < .01, * n <.05, + n < .10. 'Univariate Analysis of Variance with (2,162) degrees of freedom. Table 15 co 5 and Sta dard Deviat on Var ab 5 b Number of Alcohol c 88 arent of Pat r a Ps cho at 01 (N—l65 Families) Alcoholic Parents Neither Father Both Duncan (n-32) (n-74) (n-59) Multiple (N) (F) (B) E? Range Test Enthegg (Multivariate §(14,312) - 11.59 ****) LAPS 6.55 10.78 10.67 70.39 **** N < F,B (1.00) (1.99) (1.85) QFV-R 1.56 2.13 2.67 7.07 ** N < F < B (.95) (1.53) (1.34) BDI 1.62 2.72 3.49 4.49 * N < B (1.70) (3.02) (3.11) HRSD-Current 2.03 5.45 6.34 8.85 *** N < F,B (2.52) (4.99) (5.34) HRSD-Worst 8.53 14.53 15.51 5.96 ** N < F,B (7.52) (9.90) (10.19) ASB 10.44 22.51 24.42 14.41 **** N < F,B (6.23) (14.99) (11.21) AXIS V 67.38 53.34 51.95 27.23 **** N > F,B (8.85) (10.63) (10.16) **** n < .0001, *** p < .001, ** n < .01, * n < .05. aUnivariate Analysis of Variance with (2,162) degrees of freedom. 89 worst-ever depression and impaired psychosocial functioning and self- reports of involvement in antisocial behavior. 0n self-reported depressed mood, alcoholic men married to alcoholic women reported more distress than nonalcoholics. Neither group was significantly different from the father-only group. Across all seven variables, alcoholic fathers demonstrated more within group variability (as indicated by standard deviations) than the nonalcoholic fathers. This tended to be the case for mothers as well, but there were several exceptions. Table 16 displays group comparisons on maternal reports of family and child functioning. The MANOVA was significant. Alcoholic mothers reported more conflicted environments than nonalcoholic women married to nonalcoholic men, while the mean for the father-only group fell between the other two and was not significantly different from either. The alcoholic women also reported more child behavior problems than nonalcoholic women married to alcoholics. Neither group was significantly different from the neither parent alcoholic group, although the mean for this latter group was much closer to the mean of the father-only group. Examination of univariate results that showed a trend toward significance indicated that the nonalcoholic women married to nonalcoholics tended to report more supportive family environments than the other two groups. In addition, the alcoholic women reported more family crises than nonalcoholic women married to alcoholic men. The FCL mean for the neither parent group was actually the lowest of the three, but was not significantly different due to an increased standard deviation. Table 16 ea c es a d ndard ev atio 90 o Mate na tin s o F nd C d Fun ionin b Number of lcoh li rents (N-165 Families) Alcoholic Parents Neither Father Both Duncan (n—32) (n-74) (n359) Multiple (N) (F) (B) E? Range Test Mothers (Multivariate E(lO,3l6) - 1.89 *) FES Supportive 31.06 28.46 28.27 2.57 + N > F,B (4.38) (6.61) (6.05) FES Conflicted 8.09 9.38 11.02 4.13 * N < B (3.70) (5.43) (4.54) FES Controlling 19.00 18.24 18.51 .36 - (3.72) (4.44) (4.17) Fy Crisis List 15.56 16.03 18.27 2.89 + F < B (5.96) (5.88) (6.62) CBCL Problems 30.31 29.80 37.20 3.95 * F < B (14.42) (14.65) (18.02) * n <.05, + n < .10. aUnivariate Analysis of Variance with (2,162) degrees of freedom. 91 Table 17 shows group comparisons on paternal reports of family and child functioning. The MANOVA is again significant, but the univariate differences are modest. The alcoholic fathers who are married to alcoholic women reported a more conflicted family environment than either of the other groups. The ANOVA for child behavior problems showed a trend toward significance, but the DMRT revealed no significant differences between groups. Grou com a isons as a un t on o seve t of a ent l alcoholism. Group comparisons which include breakdowns of both male and female alcoholics by the severity of their diagnoses according to DSM- III-R criteria follow. Since these are a refinement of the above analyses which affect only some of the groups (e.g., the neither parent alcoholic group is unaffected by the increased diagnostic specificity), it is appropriate to concentrate on what these analyses tell us above and beyond the first round of group comparisons. Comparisons of demographic variables based on maternal diagnostic level are shown in Table 18. The MANOVA was significant. There were no significant differences between the more and less severe groups of alcoholic women. Significant differences were maintained between the alcoholic and nonalcoholic families, but there was no consistent pattern regarding which group or groups were significantly different from controls. In general, then, the demography of the alcoholic families does not seem to vary as a function of maternal diagnostic class (including nonalcoholic mothers married to alcoholic men). Comparisons of maternal personal functioning based on maternal diagnostic level are shown in Table 19. The MANOVA is again highly 92 Table 17 Me es and St nda ev o s e a Ratin s 0 am d Qnilg Eunctioning by Numbe; 9f Alcoholig Barents (N—165 Families) Alcoholic Parents Neither Father Both Duncan (n-32) (n-74) (BF59) Multiple (N) (F) (B) 5? Range Test Fathers (Multivariate F(10,3l6) - 1.93 *) FES Supportive 27.84 28.61 26.61 2.33 F > B (5.75) (4.56) (5.91) FES Conflicted 7.94 8.69 10.73 4.84 ** N,F < B (4.71) (4.46) (4.79) FES Controlling 20.12 18.65 18.47 1.84 - (4.18) (3.93) (4.39) Fy Crisis List 13.72 15.38 15.24 .78 - (5.11) (6.81) (6.83) CBCL Problems 24.06 30.26 30.97 2.34 + - (12.55) (16.44) (15.41) ** g < .01, * p_ <.05, + 2 < .10. aUnivariate Analysis of Variance with (2,162) degrees of freedom. 93 Table 18 e c re an tandard Dev at o s o emo a hic Variable b er Dingnn§i§_Lgygl (3&165 Families) Husband Abuse/ Moderate/ None Alcoholic Mild Severe Duncan (n-32) (n-74) (n-3l) (n-28) Multiple (N) (H) (A) (S) E? Range Test (Multivariate £(21,446) - 1.85 *) Mo. Age 31.94 29.30 29.35 30.32 3.09 * N > H,A (4.22) (4.22) (4.62) (4.30) Fa. Age 34.34 31.20 32.23 31.75 3.04 * N > A (5.67) (4.60) (5.13) (4.84) Child Age 55.16 52.61 54.42 51.57 .54 - (12.64) (11.71) (15.41) (14.21) Mo. Educ. 13.53 12.82 13.00 12.14 2.75 * N > S (2.30) (1.90) (1.81) (1.35) Fa. Educ. 14.25 12.65 12.74 12.18 6.62 *** N > A,H,S (2.14) (2.09) (1.79) (1.70) Fy. SES 41.00 30.26 29.77 27.36 6.85 *** N > H,A,S (17.35) (13.30) (10.34) (9.57) Fy. Income 7.56 6.47 6.56 5.96 3.50 * N > H,S (1.26) (2.05) (2.15) (2.36) mu 2 < .001, * n < .05. aUnivariate Analysis of Variance with (3,161) degrees of freedom. 94 Table 19 Mean co es and S andard Dev at on Maternal Ps cho atholo Variables by Mother Qiagnosig Level (N-l65 Families) Husband Abuse/ Moderate/ None Alcoholic Mild Severe Duncan (n-32) (n—74) (n-3l) (n-28) Multiple (N) (H) (A) (S) 3' Range Test Mothers (Multivariate {(21,446) - 12.04 ****) LAPS 8.38 8.76 11.09 12.93 120.69 **** N,H < A,S (.67) (.96) (1.47) (1.53) QFV-R 1.13 1.51 2.14 2.35 10.16 **** N,H < A,S (.84) (.91) (1.06) (1.33) BDI 2.75 3.41 3.52 3.93 .60 - (2.90) (3.60) (3.60) (3.57) HRSD-C 5.22 5.74 7.06 8.96 2.16 + N,H < S (6.52) (6.10) (6.73) (7.26) HRSD-W 14.97 14.46 18.55 19.14 2.08 - (11.65) (8.76) (10.40) (13.13) ASB 6.91 10.49 15.71 19.36 21.40 **** N S (13.04) (10.90) (10.74) (12.01) **** R < .0001, ** 2 < .01, + 2 < .10. “Univariate Analysis of Variance with (3,161) degrees of freedom. 95 significant. As was found before, nonalcoholic mothers differed from alcoholic mothers on the two drinking variables (LAPS and QFV-R); however, these variables did not differentiate the two groups of alcoholic mothers. On the other hand, all four groups were significantly different from one another on antisocial activities, with the predictable order of numerical increase from nonalcoholic to moderate/severe dependence. Both groups of nonalcoholic women had higher Axis V ratings than the moderate/severe group. Axis V ratings of the abuse/mild group were actually closest to the group of nonalcoholic women married to alcoholic men. The effect for worst-ever depression disappeared in the reanalysis, perhaps due to the relatively large standard deviations for this variable. Finally, following a trend in the univariate ANOVA for rated current depression, the DMRT indicated that the moderate/severe alcoholic women were more depressed than either group of nonalcoholic women. For all seven variables, the means of the moderate/severe group indicated the most impairment, numerically if not significantly. Comparisons of maternal reports of family functioning and child behavior are displayed in Table 20. The MANOVA was significant. Reports of the supportiveness of the family environment were significantly lower for mothers meeting moderate/severe criteria than for mothers in the abuse/mild group and the neither parent alcoholic group. Mothers in the moderate/severe group reported more child behavior problems than either group of nonalcoholic mothers. Both groups of alcoholic mothers differed from the neither parent alcoholic group on the reported level of family conflict. These findings clarify that for family supportiveness and child behavior problems, it is the 96 Table 20 ea ores a d Standa d ev ati te a1 ati s o a 1 nd Qnilg Eunctigning by Mother Diagnosis Level (N—165 Families) Husband Abuse/ Moderate/ None Alcoholic Mild Severe Duncan (n-32) (n-74) (n-3l) (n-28) Multiple (N) (H) (A) (S) I? Range Test Motnets (Multivariate £(15,434) - 1.90 *) Supportive 31.06 28.46 30.29 26.04 4.32 ** N,A > S (4.38) (6.61) (5.83) (5.57) Conflicted 8.09 9.38 10.81 11.25 2.78 * N < A,S (3.70) (5.43) (4.18) (4.96) Controlling 19.00 18.24 18.26 18.79 .32 - (3.72) (4.44) (4.45) (3.90) Fy Crises 15.56 16.03 17.90 18.68 2.00 - (5.96) (5.88) (5.86) (7.46) Beh Probs 30.31 29.80 35.35 39.25 2.93 * H,N < 8 (14.42) (14.65) (16.46) (19.71) ** 2 < .01, * n < .05. aUnivariate Analysis of Variance with (3,161) degrees of freedom. 97 moderate/severe group of alcoholic mothers that account for the differences observed earlier between families of alcoholic and nonalcoholic mothers. Taken together, the analyses involving level of maternal diagnosis suggest that impairments and risk factors indicated for the families where the mothers are alcoholic are due in large part to elevated risk for families with more severely symptomatic mothers. Families of mothers with alcohol abuse or mild dependence appear be firmly between those of nonalcoholic mothers and those with increased alcohol trouble, and show few significant differences compared to either group. Group comparisons involving level of paternal alcohol diagnosis follow. Comparisons on the demographic variables are shown in Table 21. Again, significant differences tend to be between alcoholic and nonalcoholic families. The only differences between the alcoholic subgroups appear for family income, where the moderate group reported a higher level of income than the severe group. Group comparisons on paternal personal functioning are shown in Table 22. As one would expect, the MANOVA and all of the univariate ANOVAs are significant. LAPS scores show a consistent and significant rise from nonalcoholic men to those with severe dependence. Alcoholics with mild and moderate dependence reported significantly greater current alcohol consumption than controls, while men with moderate dependence reported higher consumption than those with severe dependence. The latter finding probably indicates that the men who have experienced the greatest difficulties related to alcohol use are the ones most likely to decide on their own that they need to quit and/or to be court ordered 98 Table 21 Me es nd Standa Dev 0 mo ic Va 1 bles b Father Dingnn§i§_Lgygl (£9165 Families) Abuse/ None Mild Moderate Severe Duncan (BF32) (n-33) (n-52) (n-48) Multiple (N) (A) (M) (S) if Range Test (Multivariate F(2l,446) - 1.95 **) Mother Age 31.94 29.36 29.65 29.50 2.70 * N > M,S,A (4.22) (3.79) (4.50) (4.54) Father Age 34.34 31.79 31.37 31.60 2.75 * N > A,S,M (5.67) (3.86) (4.97) (5.17) Child Age 55.16 56.76 51.56 51.46 1.69 - (11.64) (13.93) (13.16) (12.20) Mo. Educ. 13.53 12.97 12.81 12.46 2.09 N > S (2.30) (1.88) (1.80) (1.71) Fa. Educ. 14.25 12.45 12.88 12.31 6.92 *** N > M,A,S (2.14) (2.12) (1.88) (1.88) Family SES 41.00 29.71 31.66 27.10 7.61 *** N < M,A,S (17.35) (9.98) (14.81) (9.08) Fy. Income 7.56 6.30 6.82 5.98 4.52 ** N,M>S;N>A (1.26) (2.02) (1.98) (2.32) me 2 < .001, H 2 < .01, v: n < .05. ”Univariate Analysis of Variance with (3,161) degrees of freedom. Table 22 Mea cores 99 d tandard Deviatio s of Paternal Ps c o a h o Vatietles by Father Diagnosis Level (fl-165 Families) Abuse/ None Mild Moderate Severe Duncan (n-32) (n-33) (n-52) (n-48) Multiple (N) (A) (M) (S) E“ Range Test Fathers (Multivariate £(21,446) - 14.96 ****) LAPS 6.55 8.94 10.38 12.34 130.88 **** NS (.95) (1.05) (1.40) (1.72) BDI 1.62 2.30 2.60 4.08 5.71 ** N,A,M < S (1.70) (2.52) (2.80) (3.46) HRSD-C 2.03 5.21 5.12 7.06 7.29 *** N < M,A,S (2.52) (3.76) (5.08) (5.86) HRSD-W 8.53 13.88 12.00 18.92 9.05 **** NM>S;N>A (8.85) (10.78) (9.21) (10.72) “he 2 < .0001, m n < .001, ** n < .01. ‘Univariate Analysis of Variance with (3,161) degrees of freedom. 100 into treatment. Whether and how many of these men stay ”dry" over the long term, and what effect that has on them and their families needs to be tracked longitudinally. Severely dependent alcoholics reported significantly greater depressed mood than any of the other groups. Ratings of current depression were significantly lower for the nonalcoholics. Lifetime ratings of depression for the severe dependence alcoholics were significantly higher than for all other groups. In addition, mildly dependent alcoholics were rated as significantly more depressed at their worst time than nonalcoholics. Reports of antisocial behavior indicated that the severe group was significantly higher than the other three and that all alcoholic groups were significantly higher than the nonalcoholic men. Axis V ratings for the nonalcoholics were significantly higher than any of the alcoholic groups. Moderately dependent alcoholics were also rated significantly higher on Axis V than those in the most severe group. Taken together, the general pattern across all four areas (alcohol use/problems, depression, ASB, and global functioning) was that nonalcoholic men differed from those with mild and moderate dependence, who in turn differed from those with severe dependence. The major exception was for current alcohol consumption, where the lowest levels were reported by nonalcoholics and those with severe dependence. Group comparisons for paternal reports of family and child functioning are displayed in Table 23. The MANOVA for these comparisons was significant. Univariate results indicated that men with severe alcohol dependence reported more family level difficulty than men in the other groups. These men reported significantly higher levels of family 101 conflict and family crises than the other three groups and more child behavior problems than nonalcoholics or men with moderate alcohol dependence. Summary. Comparisons were conducted of demographic, individual parent, family system, and child behavior variables as a function of the number of alcoholic parents in each family (zero, one, or two) and severity of parental alcoholism. Table 24 summarizes these results in terms of significant changes in difficulties and risks. The plus (+) and minus (-) signs in each of the three columns for both mothers and fathers indicate significant differences in risk and trouble as we move across a continuum from nonalcoholic families to families with alcoholic fathers only (column 1); to families where both parents are alcoholic (column 2); and finally to families where the parent is severely alcohol dependent (column 3). This is not meant to imply that families are assumed to move through this continuum. In fact, the families in each group are likely qualitatively distinct. Rather this is a continuum of presumed risk based on a model of cumulation of risk and protective factors for individuals or single families (e.g., Cicchetti & Rizley, 1981; Schneider-Rosen, Braunwald, Carlson, & Cicchetti, 1985; Zucker, in press; Zucker & Fitzgerald, in press). In general, the number and magnitude of significant effects decreased as the focus moved away from the individual parent to the family system. Clear differences were seen for parental resources and achievement, and for parental psychopathology and global psychosocial functioning between groups. The general pattern was for impaired 102 Table 23 ea co es and Standard Dev a n o Pater a atin d Qnile Ennctiening by Eethet Diagnosis Level (N-165 Families) Abuse/ None Mild Moderate Severe Duncan (n-32) (n-33) (n-52) (n-48) Multiple (N) (A) (M) (S) if Range Test Eetnets (Multivariate £(15,434) - 2.40 **) Supportive 27.84 28.85 28.02 26.62 1.22 - (5.75) (4.62) (4.61) (6.20) Conflicted 7.94 8.21 8.75 11.46 5.43 ** N,A,M < S (4.71) (4.31) (4.04) (5.11) Controlling 20.12 19.15 18.62 18.12 1.61 - (4.18) (4.47) (3.85) (4.20) Fy Crises 13.72 12.58 14.96 17.58 4.75 ** A,N,M < S (5.11) (6.59) (6.15) (6.95) CBCL 24.06 29.58 27.17 34.94 3.85 * N,M < S (12.55) (17.18) (14.43) (15.92) ** n < .01, * n < .05. aUnivariate Analysis of Variance with (3,161) degrees of freedom. 103 Table 24 Summety Table of Incremental Differences in Riek as e Ennction of Nnnber of Aleeholic Patente and Severity of Parental Alcoholisn Mothers Fathers 1.° 2.” 3.c 1. 2. 4.6 Pa en A cohol ia osti Statu Father + + + + + ++ Mother 0 + ++ o + o/+ Demography - SES + + - Education + + - Income + + + - Age + + Eeyenopethelogy + LAPS + + + + QFV-R + + + - + BDI + + + HRSD-C + + HRSD-W + + + + ASB + + + + + + Axis V + + + Child Behavior + Total Probs + + + (table continuee) 104 Mothers Fathers 1.° 2.b 3f 1. 2. a.d Ey Envirenment - Support + + + + Conflict + + + + Crises + + Significant Risk IncrementsZLevel 5 9 3 10 4 9 Nete. + indicates increased risk/difficulty; - indicates decreased risk/difficulty. See text for further explanation of table. aFather-only alcoholic compared to nonalcoholic families. bBoth mother and father alcoholic compared to father-only-alcoholic families (or to nonalcoholic families if they were not significantly different than father-only-alcoholic families). cMother moderately/severely alcohol dependent compared to other families with alcoholic mothers. dFather severely alcohol dependent compared to other families with alcoholic fathers. 105 functioning in alcoholic families compared to controls, although nonalcoholic women married to alcoholic men looked more like the control women than like the alcoholic women with alcoholic husbands. The former groups differed from controls in being more antisocial, but still less so than the alcoholic women. Mother and father reports both indicated that greater family conflict compared to controls was associated with both parents being alcoholic. Alcoholic women also reported more child behavior problems and more family crises compared to nonalcoholic women with alcoholic husbands. The refinement of groups based on severity of parental alcoholism confirmed that families of parents with severe alcohol dependence tended to be the most impaired according to the effects discussed above, although it was rare that mean scores between alcoholic subgroups were found to be statistically different. There were, however, three important exceptions to the above generalizations: (1) severely dependent alcoholic men reported significantly lower levels of current alcohol consumption than other alcoholic men; (2) significant differences were found between parent psychopathology and global functioning variables, with increased alcohol severity being associated with impaired functioning beyond alcohol-specific symptomatology; and (3) men with severe alcohol dependence reported perceptions of the family environment and child behavior which differed significantly from all other groups and exceeded the moderate differences obtained between control families and alcoholic families as a group. 106 Va b e Given concerns that have been raised about the psychometric acceptability of instruments of child-rearing attitudes (Holden & Edwards, 1989), it seemed prudent to further explore the reliability and validity of the CRPR before beginning analyses with this instrument. The relatively complex nature of any Q-sort instrument, and the limited academic and intellectual background of much of the present sample, indicate that it is a legitimate question whether coherent and consistent (internally and externally) responses were obtained. This general area was explored in three ways: (1) intercorrelations between the factor scores and the PAI were computed cross-sectionally to check for general consistency of response (e.g., nurturance and authoritarianism should be negatively correlated (Davies et al., 1991)); (2) test-retest stability for that portion of the sample with available longitudinal data was examined for the factor scores, the PAI, and the overall distribution of individual items; and (3) relationships between the CRPR scores and parents' reports of specific behavioral parent-child conflict-focused interactions were assessed. lntereottelations. Table 25 presents the intercorrelations between the four CRPR factors and the Parent Agreement Index (PAI). A11 combinations of variables were significantly correlated in the anticipated direction with the exception of maternal and paternal nurturance. For both parents, nurturance was negatively associated with authoritarianism, while scores on the PAI were positively associated with nurturance and negatively associated with authoritarianism for both parents. These results are congruent with the Deal et a1. (1988) 107 Table 25 lnteteotteletiene (gearson R) between §oeializatien Vatiablee (N—165 families) 2 3 4. 5 1. Mother Nurturance -.34 * .08 -.19 * .42 * 2. Mother Authoritarianism -.17 + .33 * -.51 * 3. Father Nurturance -.35 * .62 * 4. Father Authoritarianism -.53 * 5. Parent Agreement Index * p < .05, + p < .10, two-tailed. Significance values adjusted via Holm's procedure. 108 position regarding the PAI-~name1y that it is associated with agreement to a standard of positive, effective parenting rather than simply representing concordance within a couple independent of content. Put another way, couples do not agree to use ineffective or destructive parenting techniques. Yet it is also clear from Table 25 that none of the socialization variables are redundant with one another and in all cases there is considerable nonshared variance. Lengitudinal Stability ef CRPR Meaeutee. Second administrations of the CRPR were available from the longitudinal component of the project for one or both parents in 68 families. These data allow examination of the stability of CRPR variables over time. The mean intertest interval was 16.5 months, with a range of 8 to 27 months. Previous test-retest analyses using the CRPR have used non-parent populations (Block, 1965) or had extended intertest intervals (i.e., eight years in the case of Roberts et al., 1984). Such an extended interval complicates the stability question, as considerable developmental changes will have taken place in the child and parent. The coefficient alpha estimates (see Table 26) were lower for all four scales than in the first administration (recall that the factor analysis was conducted on a subset of that sample) but all reliabilities were still reasonably acceptable. Table 26 also displays the test- retest correlations for the four factor scores and the Parent Agreement Index. These analyses indicate that there is considerable stability for all five measures. Further analyses assessed the stability of the overall Q distribution from one data point to the next. A correlation was 109 Table 26 es -'-te Re- ab it fea -.1 ' o_ ,-c : l_a 01 V: anle a . ter Co 5 stencie oeff cie t A1 ha at eco dministratio Test-Retest i Alma Mother Nurturance .73 *** .72 Authoritarianism .65 *** .65 Mean Intraparent Q-deck Agreement8 .68 - Father Nurturance .62 *** .61 Authoritarianism .64 *** .58 Mean Intraparent Q-deck Agreement .62 - PAI .71 *** - Nete. Mean intertest interval was 16.5 months (range 8-27 months). Sample sizes were n-66 for mothers, ne6l for fathers, and BF57 for the PAI. fiMean item-by-item correlation coefficient across the sample of mothers (or fathers). ,Computation is analogous to the cross-sectional PAI. See text for discussion of interpretation. *** p < .001, two-tailed. Significance values adjusted via Holm's procedure. 110 computed for each individual parent across all 91 items (this is analogous to the PAI computation, and assesses the relative consistency of responses to individual items) (see Table 26). For mothers, the mean value of these correlations was .68, and for fathers it was .62. Both of these values were significantly greater than the mean level of interparent agreement (mother vs. father, I - .50) at the first data point (for mothers, t(58) - 11.82, p < .05; for fathers, t(59) - 7.73, p < .05). Taken together, these results indicate that there is considerable stability in the responses of parents to individual items on the CRPR, as well as in the higher-order socialization variables being used in this study. More generally, and more importantly, these results provide strong evidence for a consistent internal plan or model of parenting which is a coherent cognitive structure or schema. A schema has been described as: that portion of the entire perceptual cycle which is internal to the perceiver, modifiable by experience, and somehow specific to what is being perceived. The schema accepts information as it becomes available at sensory surfaces and is changed by that information; it directs movements and exploratory activities that make more information available, by which it is further modified. (Neisser, 1976, p. 54) Thus, the attitudes that these parents hold about child-rearing and child development are consistently expressed and should be expected to interact with the environment in an orderly way in determining parent-child interaction and influencing child development. 111 Deletionships of CRPR neasures to tepetted perent-to-cnilg aggtession. One further psychometric investigation of the CRPR was undertaken. The concurrent validity of the instrument was assessed by comparing the factor scores to parents' reports of problem-solving interactions with the child. Scores on the four CRPR factor scores and the PAI were correlated with scores based on parental reports on the Conflict Tactics Scale (CTS; Strauss, Gelles, & Steinmetz, 1980). The CTS is administered as an interview during the DIS, and focuses on means of resolving conflict with other family members. The interview format and specific behavioral focus of the CTS provide a rough assessment of the external validity of the CRPR which is relatively free of method variance. The two instruments were collected at different visits under very different formats. The CTS information is gathered as an interview which is imbedded in the Diagnostic Interview Schedule. The placement of the instrument allows for maximum rapport between the subject and the interviewer (see Reider, 1987, for full details of data collection). The context of this visit contrasts sharply with the self-report nature of the CRPR administration. The two visits were nearly always at least two weeks apart, and in many cases the interval exceeded several months. Two scores are derived from the CTS for each parent for the preceding twelve months (Reider et al., 1989). Sevetity is the highest level of physical violence reported based on a Guttman scale of CTS events which reflects escalating conflict and violence. Cnnulative intensity is the product of level of violence and frequency of occurrence summed across all levels of violence. Work just completed on a sample from this project (Cruise, 1991) 112 has shown that the combination of these two scores across both parents yields a four-item cluster (alpha - .68) which taps the harshness of parental discipline. The psychometric properties of this cluster are superior to similar clusters constructed for mothers and fathers separately. The correlations of this measure of family-level harshness of discipline and the CRPR variables are shown in Table 27. Harsh discipline was positively related to authoritarian attitudes of both parents, and negatively related to paternal nurturance and interparent agreement. While the size of the correlations are modest, they do indicate that there is coherence to the CRPR parental reports and that there is some (albeit low order) connection between these scores and parents' discrete behavioral reports. Cotrelatione witn independent vetieblee. We turn now to correlations between the dependent variables (the socialization domain) and the other variables in the study. This section of the study can be seen as a replication and extension of Davies et a1. (1991). It is an extension in that a wider range of independent variables has been sampled, and because control (i.e., nonalcoholic) families have been included here, whereas the earlier study was based solely on alcoholic families. Correlations between family demographic variables and the socialization variables are shown in Table 28. Maternal nurturance was unrelated to any of the demographic variables. Associations were found between maternal and paternal authoritarianism and parental concordance 113 Table 27 Co e o Pears n be we R R oc at on Va 'ables and S ec c Re orts of Harshness f isci ine of at Parents on he Qeniliet Taeties Scele (Nr96 Families) t Mother Nurturance -.l7 Authoritarianism .30 * Father Nurturance -.35 ** Authoritarianism .21 + PAI -.23 + ** p < .01, * p < .05, + p < .10, two-tailed. Significance values adjusted via Holm's procedure. Table 28 Co re a ons Pearso betwee Varieties (N—165 families) 114 Soc a1 zat n Va iables and Demo ra ic Mother Father Nurtur. Author. Nurtur. Author. PAI Family SES .03 -.32 .12 .21 .20 Family Income .01 -.32 .19 .34 * .22 + Mother Age .05 -.33 .21 .41 * .31 * Mother Education .02 -.43 .25 * .30 * .20 Father Age .05 -.22 .12 .37 * .19 Father Education .09 -.21 .17 .27 * .20 Child Age .13 .00 -.23 + .00 -.16 * p < .05, + p < .10, two-tailed. Holm's procedure. Significance values adjusted via 115 and the demographic factors of family income, mother age, and mother education. Paternal authoritarianism was also related to paternal age and education. In all these cases, higher authoritarianism and lower agreement were associated with fewer resources, lower achievement, and younger age of the parent. Paternal nurturance was negatively associated with child age-~that is, fathers tended to report being more affectionate with younger sons. Child age was not related to any of the other dependent variables. Table 29 lists the correlations between the socialization variables and maternal and paternal functioning (psychopathology). Maternal nurturance was negatively associated with mother's level of involvement in antisocial behavior, while paternal nurturance was negatively related to the father's level of self-reported depressed mood. Maternal authoritarianism was negatively related to the mother's Axis V rating. Paternal authoritarianism showed no significant relationships to any of these independent variables. The level of parental agreement was positively related to the mother's Axis V ratings. At first glance, these results may appear not to replicate the earlier study based only on alcoholic families, at least with regard to the parental functioning domain. However, in most cases the magnitude of the correlations does not differ substantially; rather the reduction in power due to the correction procedure is more pronounced in the current investigation because there are more comparisons being made. The only striking reduction in effect size is for the relationship between paternal nurturance and paternal lifetime alcohol problems, 116 Table 29 0 ti ns ea so betw en oc a atio Variables d ot e Eetnet Esyehopathology Variables (N—l65 families) Mother Father Nurtur. Author. Nurtur. Author. PAI na 5 cho atho 0 Lifetime Alc. Probs. -.11 .04 .06 -.02 .05 Current Drinking (QFV-R) -.12 .13 .15 .04 .01 Beck Depression -.15 .11 -.10 .03 -.19 Hamilton-Current .07 .08 -.05 .11 -.06 Hamilton-Worst .00 .13 -.07 .07 -.03 Antisocial Behavior -.23 + .22 .00 .14 -.08 Axis V .08 -.29 * .14 -.19 .24 + te n s o athol Lifetime Alc. Probs. .02 -.01 -.09 -.04 .01 Current Drinking (QFV-R) -.06 .05 -.09 .09 -.11 Beck Depression .03 .08 -.26 * .05 -.18 Hamilton-Current -.12 .09 -.17 -.02 -.11 Hamilton-Worst .04 .08 —.03 -.12 .03 Antisocial Behavior .00 .09 -.17 .02 -.07 Axis V .04 -.15 .19 -.05 .15 * p < .05, + p < .10, two-tailed. Significance values adjusted via Holm's procedure. 117 which dropped from t - -.26 in the first study to t - -.09 in the present study. The general finding of the first study did replicate-— namely that family demography is a more consistent correlate of, and accounts for more variance in, parenting attitudes (with the exception of maternal nurturance) than does psychopathology. However, the strengthening of associations that might have been expected from broadening the nature of the sample by including controls, and thus increasing the variability of the independent (and perhaps the dependent) variables, was not observed. Correlations between the socialization variables and parental reports of family and child functioning are shown in Table 30. Again, the number of significant effects and their magnitudes are modest at best. Maternal and paternal nurturance were negatively associated with their own reports of child behavior problems. Maternal reports of the level of family supportiveness were negatively related to her level of authoritarianism and positively related to the level of parental concordance. Paternal authoritarianism showed nosignificant relationships to the family system and child variables. gteup eenparisons. Table 31 shows comparisons of the socialization variables as a function of the number of alcoholic parents. The MANOVA was nonsignificant for these variables, indicating that on the multivariate level these groups did not differ. Exploratory univariate analyses revealed a significant ANOVA only for maternal nurturance. Alcoholic mothers reported less nurturant behavior than nonalcoholic women married to alcoholic men. The lack of significant results in this area is surprising given the consistent group 118 Table 30 . ,- : 0 Fear 01 ' betwee . _=_ z. _. Va__=b -s a,. . h-_ and Eetnet Ratings pf Family end Chile Ennctioning (N—165 families) Mother Father Nurtur. Author. Nurtur. Author. PAI Mate Ratin FES Supportive .16 -.29 * .14 -.14 .23 FES Conflicted -.20 -.01 -.10 -.02 -.09 FES Controlling -.03 .16 -.02 .11 .00 Family Crisis List -.03 -.03 .00 .00 .07 Child Behavior Problems -.22 + .07 -.14 .09 -.15 Pate a Ratin s FES Supportive .17 -.18 .16 -.11 .20 FES Conflicted -.06 .01 -.12 -.11 -.09 FES Controlling -.02 .08 -.04 .20 -.13 Family Crisis List -.01 -.01 -.15 -.07 -.06 Child Behavior Problems -.17 .06 -.23 + .05 -.13 * p < .05, + p < .10, two-tailed. Significance values adjusted via Holm's procedure. Table 31 Me Ennbep of Alcohelic Patents (N-165 Families) c es a d tan d ev'at s 119 0 0C a at 0 Va ab es b Alcoholic Parents Neither Father Both Duncan (n-32) (n-74) (n559) Multiple (N) (F) (B) E? Range Test (Multivariate £(lO,316) - 1.28, ns) Mo. Nurturance 124.69 126.24 120.92 3.34 * F > B (12.63) (10.10) (13.55) Fa. Nurturance 203.41 203.28 205.34 .50 - (12.83) (12.52) (12.16) Mo. Authoritar. 52.41 52.16 53.59 .26 - (10.85) (11.78) (10.99) Fa. Authoritar. 32.38 33.62 34.69 .92 - (8.32) (7.60) (8.04) PAI .51 .51 .52 .06 - (.15) (.15) (.15) * p (.05. °Univariate Analysis of Variance with (2,162) degrees of freedom. 120 differences across demographic and parental individual functioning variables, and the relationships already articulated between these domains of the family system and socialization attitudes and parental concordance. The absence of a main effect for number of alcoholic parents on the child-rearing variables contradicts the study hypotheses. Thus, there is no indication that the proposed analyses of covariance should be pursued. Comparisons of the socialization variables based on severity of maternal diagnosis are shown in Table 32. The MANOVA was nohsignificant as we would anticipate from the earlier results. Exploratory analyses again showed the only differences for maternal nurturance, this time a nonsignificant univariate.trend with more severely alcoholic mothers reporting less nurturance than nonalcoholic women whose husbands are alcoholic. Finally, comparisons for the socialization variables as a function of paternal alcoholism severity are shown in Table 33. All multivariate and univariate analyses revealed no significant differences across groups on these variables. Summary. The strongest correlations to the socialization variables were obtained for the demographic variables, and most clearly for authoritarianism for both mothers and fathers. The level of parental concordance was positively associated with the age of the mother and showed a nonsignificant positive trend with family income. Parental nurturance was generally unrelated to demographics, the exception being a nonsignificant trend suggesting that the fathers of 121 Table 32 d da d Deviat s o i lizatio Va ab es b 05 v (3-165 Families) Husband Abuse/ Moderate/ None Alcoholic Mild Severe Duncan (n-32) (n-74) (n-31) (n-28) Multiple (N) (H) (A) (S) E? Range Test (Multivariate £(15,434) - .96, ns) Mo. Nurt. 124.69 126.24 121.71 120.04 2.31 + H > 8 (12.63) (10.10) (14.12) (13.08) Fa. Nurt. 203.41 203.28 206.58 203.96 .55 - (12.83) (12.52) (12.62) (11.70) Mo. Auth. 52.41 52.16 53.19 54.04 .20 - (10.85) (11.78) (13.66) (10.05) Fa. Auth. 32.38 33.62 34.94 34.43 .63 - (8.31) (7.60) (7.37) (8.85) PAI .51 .51 .52 .52 .05 - (.15) (.15) (.18) (.12) * p < .05, + p < .10. aUnivariate Analysis of Variance with (3,161) degrees of freedom. 122 Table 33 -:1 _ . -s 3,. _aqda . Ievia. .,_ ._ . : _:t on V: ab e b Eetnet Diagnoeis Level (N—165 Families) Abuse/ None Mild Moderate Severe Duncan (n-32) (n-33) (n-52) (n-48) Multiple (N) (A) (M) (S) E? Range Test (Multivariate {(15,434) - .80, ns) Mo. Nurt. 124.69 124.52 123.15 124.23 .14 - (12.63) (12.45) (10.47) (13.39) Fa. Nurt. 203.41 206.36 204.46 202.42 .70 - (12.83) (10.15) (11.94) (14.06) Mo. Auth. 52.41 52.12 52.85 53.21 .07 - (10.85) (11.60) (11.74) (12.36) Fa. Auth. 32.38 33.61 35.65 32.75 1.61 - (8.32) (7.91) (7.88) (7.46) PAI .51 .53 .52 .50 .28 - (.15) (.16) (.14) (.16) aUnivariate Analysis of Variance with (3,161) degrees of freedom. 123 (alder boys reported less nurturance with them. Correlations of the socialization variables with the parental psychopathology and family functioning variables showed very few significant relationships. Paternal nurturance showed significant negative relationships with father's current depressed mood and his reports of child behavior problems. Maternal nurturance showed only nonsignificant trends suggesting that more nurturant mothers were less antisocial and reported fewer child behavior problems. Maternal authoritarianism and interparent agreement both showed significant relationships with mother's current psychosocial functioning (negative and positive respectively) and her perceptions of the supportiveness of the family environment (positive and negative). Paternal authoritarianism showed no significant associations to variables in these domains. No differences were found on reported nurturance or authoritarianism of either mothers or fathers or on interparent concordance between the three groups of families (i.e., those with zero, one, and two alcoholic parents). This remained true when refining the groups to take into account the severity of parental alcoholism. Thus parental alcoholism per se appears to have little or no main effect on the socialization attitudes of parents. This finding is surprising given the consensus in the literature that this is a distinctive group of families, especially with regard to the child-rearing environment. Furtnet Exploratione tre e ou s on var a s othe an a coho s . Given that divisions of the sample based on alcohol diagnoses (and thus alcohol problems) did not yield differences in reported attitudes toward child- rearing, other possible divisions of the sample were explored based on 124 cflmaracteristics which have been reported or postulated to be related to socialization practices. Specifically, cases were selected which represented the high and low quartiles for mothers and fathers separately on child behavior problems, family conflict, and family «arises (or stress). A second division of each subsample was made based <3n.whether the family fell above or below the mean on an aggregated family demography variable. This was computed for each family by summing the standard (z) scores for family SES, family income, mother and father education, and mother and father age. Two-by-two ANOVAs were then performed on the socialization variables to determine whether the selection variable, family demography, or both contributed to variation in parenting attitudes. As one would expect from the correlations between demographics and family characteristics shown earlier, the cell sizes were not equal, but none violated the design assumptions for ANOVA. Across each of these analyses, the results replicated the correlational results and there were no significant interactions; given this consistency, these results are not presented here. ou o arisons o v t m . Given the lack of group differences in the CRPR factors for parental alcoholism, the possibility arises that the differences that set these families apart occur on dimensions other than the nurturance and authoritarianism factors developed through factor analysis. As noted above, group comparisons on individual items is a commonly used technique with the Q- sort methodology. This approach provides another avenue into the general question of socialization attitudes in alcoholic families. The approach taken to pursue this possibility was discriminant 125 function analysis. Across several divisions of the 91 Q-sort items (based on expert predictions, selection of items showing significant univariate effects, and random selection of items) no coherent discriminant functions emerged which provided orderly prediction of cases by diagnosis groups. We can therefore conclude that differences in parenting attitudes were not found to be systematically related to parental alcohol diagnosis, even when we go beyond the factor scales and examine combinations of individual items. CHAPTER IV Discussion The S e e s ec v Before discussing the results, it seems prudent to highlight the unique characteristics of the sample and limits on generalizability that need to be kept in mind. The alcoholic families in this study were identified when the father was convicted of drunk driving. High participation rates assure that this sample is closely representative of the population of drunk drivers with young male children living initially in intact families. These fathers are typically not yet in treatment. The fact that these men are convicted drunk drivers indicates that their alcoholism is more heavily combined with antisociality (Cloninger, 1987; Zucker, 1987) than is true of other types of alcoholism. From a typological standpoint, an analysis of the first 100 cases from the study sample showed that 60% were classifiable as Type II (antisocial) alcoholics according to Cloninger's (1987) typology; 25% were classified as Type I (nonantisocial, later onset) and 14% were indeterminate. Thus, this population of alcoholic men is most reflective of the subset of alcoholics known to be most damaged, with the most psychosocial comorbidity, and the earliest onset (see Regier et al., 1990). The use of convicted drunk drivers also means that alcoholic men with greater resources (who are less likely to be convicted if arrested) are underrepresented. The nonalcoholic families were identified through neighborhood canvassing in the same census tracts where the alcoholic families lived. Comparison families were screened out if either parent had a history of 126 127 life difficulties related to alcohol or other drug abuse. These controls are not subject to the interpretation difficulties presented by volunteer samples, "snowball" samples, or other samples of convenience. Finally, the target child in each family is between 3-0 and 6-0 years at the time the family is recruited. This is a significantly earlier developmental family stage than has been investigated previously in relation to familial effects of substance abuse (or family difficulties more generally). More importantly, this investigation is drawn from what will be the first of multiple waves of data collection at three year intervals. This study should be conceived as laying the groundwork for subsequent investigations of these families and their parent-child relationships over time. Fam ember and t e F 1 S ste n You 0 01 c es Demegtephit_tnetetteti§tie§. The parents in alcoholic families in this sample consistently showed fewer resources and less achievement than parents in nonalcoholic families drawn from the same neighborhoods. This was true of family socioeconomic status (SES), family income, and the educational level of the parents. The nonalcoholic parents were also two to three years older on average, despite the fact that the target children showed no differences in age. Thus the entry into parenthood may already be compromised, as well as contributed to, by parental impulsivity or a failure to delay gratification. This may explain some of the SES differences and raises the possibility of greater maturity on the part of the nonalcoholic parents as they face the challenges of child-rearing. This line of reasoning is also supported by the positive association between mother's age and the level of parental concordance about child-rearing. 128 No differences in demographic status were discovered between father-only-alcoholic and both-parents-alcoholic families. Further comparisons incorporating information on the severity of parental alcoholism revealed virtually no significant differences between alcoholic sub-groups. The only exception was that families of severely alcohol dependent (AD) men reported significantly less family income than did families of moderately AD men. The substantial differences between alcoholic and nonalcoholic families from the same neighborhoods suggests that alcoholic families should tend to stand out even within their neighborhoods. Indeed, this has often been the experience of project staff. These homes often appear "run-down” or "cluttered" compared to the other houses on their street. Of course, the distributions of appearance of alcoholic and nonalcoholic homes are far from non-overlapping. It is also possible that the two groups of families may be in the same neighborhoods for different reasons. Some of the control families may be in relatively deprived neighborhoods only temporarily, and will work themselves out into other environments. The alcoholic families are probably less likely to be on such a positive trajectory. It will be interesting to follow changes in the families' SES and neighborhood ecology longitudinally. Eaternal psychopathology. Given that paternal psychopathology (in the form of alcoholism) was the basis for selection for the study, the finding of large group differences should not be surprising. However, these differences extend well beyond the domain of alcohol-related difficulty. Alcoholic men reported more impairment or were rated as more impaired on measures of current and lifetime depression, 129 antisociality, and overall psychosocial functioning. Whether or not the wife was alcoholic generally appeared to make little difference for the alcoholic men. However, the alcoholic men with alcoholic wives did report higher levels of current drinking and depressed mood than did the alcoholic men with nonalcoholic wives. Severity of paternal alcoholism was an important correlate across all domains of psychopathology and clearly indicated that the severely alcohol dependent (AD) men were simultaneously experiencing a good deal of other life trouble. In addition to the highest levels of lifetime alcohol problems, they reported higher levels of current depressed mood, lifetime depression, and antisociality than did alcoholic men with lesser levels of dependence. They were also rated as having more impaired psychosocial functioning. The severely AD men reported significantly less current drinking than did the moderately AD men, and their group mean was squarely between the means of the other alcoholic groups and the nonalcoholic group. This suggests that the life difficulties they report related to their drinking (personal, legal, financial, and social) have led many of them to cease or curtail their drinking, at least temporarily. The permanence of this behavior change and its effect on the men and their families is important to follow longitudinally. Materna s cho atholo . As with the fathers, the differences in alcohol-specific maternal psychopathology were between alcoholic and nonalcoholic women, regardless of their husbands' diagnostic status. Alcoholic women married to alcoholic men were more alcohol involved, had greater histories of depression, were more antisocial, and were lower in overall level of mental health. They were clearly the most damaged 130 group of women in the study. For nonalcoholic women, having an alcoholic husband was associated only with their own greater antisociality. When examining possible effects pertaining to severity of maternal alcoholism, the alcoholic groups did not differ significantly on level of lifetime alcohol problems or current alcohol consumption, but the differences are in the anticipated direction and are of the same order of magnitude as the father group differences. The moderately/severely AD women did report more antisociality than those in the mildly AD/abuse group. The differences found between alcoholic and nonalcoholic women were found to be due to the moderately/severely AD women. A difference also appeared on rated current depression between the moderately/ severely AD women and both groups of nonalcoholic women. Thus, the most severely alcohol dependent women were experiencing more symptomatology in all areas except alcohol-specific ones. The lack of relationship between the drinking specific variables, especially lifetime alcohol problems, and severity of diagnosis is puzzling given that there is considerable content overlap between LAPS and the bases of alcohol diagnosis, and hence group assignments. This may be due in part to the relative lack of severely AD women, who had to be combined with moderately AD women for statistical purposes, and a resulting loss in statistical power. This diagnosis represented over a quarter of the sample for men, and accounted for the most striking group differences. It is also conceivable that the use of lifetime diagnoses becomes important here. The alcoholic women as a group may have become more responsible after the birth of their child or children, which would 131 have reduced differences between the groups when examining variables related to current functioning. gnild peneviet. The relationship of child behavior to parental alcoholism was not a major focus of this investigation, as only one variable tapping this domain was included (Total Child Behavior Problems from the Achenbach Child Behavior Checklist). Nonetheless, the results obtained were consistent with the literature on older children (see West & Prinz, 1987) and particularly suggest an effect of cumulated parental risk in determining child behavior problems. Specifically, alcoholic mothers (i.e., mothers in both-parents- alcoholic families) reported more behavior problems for their sons than did nonalcoholic mothers, even when the father was alcoholic. No significant group differences were found for father reports based on the number of alcoholic parents. Severely AD mothers and fathers both reported the highest levels of child behavior problems. For mothers, these problems were seen as significantly higher than they were for both groups of nonalcoholic mothers; for fathers, they were significantly higher than the reports of moderate AD fathers and nonalcoholic fathers. Thus, child behavior appears to be more problematic when both parents are alcoholic and/or when alcoholic parents have experienced more frequent and more severe symptomatology. This area is important and deserves more attention. In particular, additional measures of child behavior (both prosocial and antisocial) and personality/ temperament should be explored in relation to aspects of parental alcoholism. In addition, there may be specific areas of child behavior that are differentially associated with these variables (e.g., externalizing vs. internalizing). 132 This area has been explored in more detail in other work on the project data set (Cruise, 1991; Reider, 1991). Using cross-sectional path models, Reider (1991) found parental depression to be particularly salient among alcoholic families in predicting total child behavior problems and child depression. Maternal depression was also associated with child aggression. The inclusion of behavioral task measures and clinician ratings in the work of Cruise (1991) and Reider (1991) also permits us to rule out biased parental perceptions as the sole cause of the child behavior problem effects found in this study. Fenily envitonnent. The variables explored in this area were factors from the Family Environment Scale and parents' reports of the level of family crises. No differences were found for reports of either parent on family controllingness based on either number of alcoholic parents or severity of alcohol diagnosis. The alcoholic women perceived their families to be less supportive and more conflicted than did the women in neither-parent-alcoholic families. The alcoholic women also reported a higher level of family crises than the women in father-only-alcoholic families. Finally, the nonalcoholic women perceived less supportiveness in the family when they were married to alcoholics. The alcoholic men with alcoholic wives reported less support and more conflict than alcoholic men with nonalcoholic wives, and more conflict than nonalcoholic men. This is the only area, apart from level of current drinking, where the wife's alcoholism appears to impact the alcoholic husband's reports of his experiences. It is also perhaps the area where we would be most likely to expect such an effect. Psychopathology of both parents would be expected to impact the family 133 environment directly, while effects on the father's individual functioning would be less direct. The severely/moderately AD women reported less supportive environments than did the mildly AD/abuse women or the women in neither- parent-alcoholic homes. Both groups of alcoholic women reported more conflict than women in the neither-parent-alcoholic group. For fathers, the severely AD group again emerged as being markedly disparate from all of the other groups. No effects were observed for supportiveness, but the severely AD men reported significantly more conflict and family crises than did the other three groups. So overall, mothers perceived the family as less supportive when the father was alcoholic and when the mother was more severely alcoholic. Higher levels of conflict and crises were associated with maternal alcoholism of any level. For fathers, supportiveness was lower and conflict higher when the mother was alcoholic. Conflict and the level of crises were higher when the father was more severely alcoholic. Thus, both husbands and wives agree that the level of family conflict is higher when the wife is alcoholic and that having an alcoholic spouse is associated with less supportiveness. xte t of co 0 s d eve of cumulative risk. To summarize the results of the study so far, we will review the information in Table 24 regarding incremental changes in putative risk for families across the continuum of one- and two-parent-alcoholic families and with severely alcohol dependent parents. Compared to nonalcoholic families, one-parent-alcoholic families were of lower SES and income and the fathers had less education. The fathers showed greater psychopathology across all measures except self-reported current 134 depressed mood, while the mothers were more antisocial. Neither parent reported differences in child behavior problems. The mothers experienced the family as less supportive, while the fathers reported no differences in the family environment. Then, comparing alcoholic families who also have an alcoholic mother to those with only an alcoholic father, the only change for demography was that the alcoholic women had less education. The alcoholic women also showed greater alcohol-specific symptomatology, antisociality, lifetime depression, and global functioning. The alcoholic husbands of alcoholic women reported greater current drinking and higher levels of self-reported current depression. The mothers reported higher levels of behavior problems in the children. Finally, both parents agreed that the environment was more conflicted, while mothers reported more crises and fathers perceived less support. Compared to other alcoholic mothers, the moderately/severely AD women were more involved in antisocial behavior, reported more child behavior problems, and perceived their family as less supportive. The severely AD fathers, on the other hand, had less family income and evidenced greater psychopathology in the form of drinking problems, depressed mood, lifetime depression, antisociality, and impaired global functioning. They also reported less current drinking than moderately AD men. They reported more child behavior problems and higher levels of family conflict and crises. Thus, there is compelling evidence for our assumption of cumulation of risks and trouble in these families as a function of both number of alcoholic parents and severity of alcohol trouble within groups of alcoholic parents. The model posits that these risks and 135 troubles would translate into developmental risks for the children growing up in these families, and that a major source of transmission of that risk would be through the parent-child relationship. A word of caution is in order about this model, however. There is evidence for continuity between behavioral and environmental characteristics identified in adolescence, and the later development of alcoholism (see review by Zucker 6 Gomberg, 1986). Similar linkages have been established for the time-frame between childhood and adolescence (see Zucker, 1989). While this provides the basis for considering individual and family characteristics in the preschool years as potential risk factors for development into adulthood, this must be considered a hypothesis that remains to be tested (Zucker 6 Fitzgerald, in press). It also raises the important question of which contexts and experiences heighten or limit the risks present in the early environment or in the child as the child develops. It would appear that the stage was well set to discover striking differences in socialization practices as a function of divisions based on parental alcoholism characteristics. The surprising absence of such differences is addressed next. The arent-Child e tionshi The lac o ocialization ttitudes ' the a ent- ild relationship. The question of what the Child-rearing Practices Report is actually measuring needs to be considered first in understanding the results. In this regard, it is helpful to draw upon a recent, highly critical review of parent child-rearing attitude questionnaires (PCRAs; Holden 6 Edwards, 1989). The CRPR is an instrument of this type, although it was not the primary focus of the article. The reviewers 136 were critical of the lack of psychometric rigor regarding reliability and validity of these instruments as a group. Some of these concerns for the CRPR have been addressed in this investigation. The CRPR factors demonstrate adequate internal consistency within and beyond the sample on which the factor analyses were based, and there is considerable evidence for test-retest reliability at the level of both factors and individual items. Examination of the relationship between CRPR scores and the Conflict Tactics Scale provided some evidence of convergent validity of the CRPR, although the effect sizes were modest. Taken together there is considerable evidence for a cognitive schema underlying parenting attitudes. The issue of concurrent validity has perhaps been most vexing and raises a number of important questions about the relationships of parental attitudes to parent-child interaction. Holden and Edwards (1989) conclude that it is no surprise that attitudes and behavior are not directly linked given the other components that are present and powerful in the parent-child relationship (e.g.,characteristics of the child, characteristics of the parent, cultural and situational contextual variation). Rather than abandoning parenting attitudes as subjects of research, these reviewers urge researchers to recognize that PCRAs are measuring but one aspect of social cognition (the parenting schema as presented here) and to become less simple-minded about the connections between parenting attitudes and behavior. For example, Ajzen and Fishbein's (1980) Theory of Reasoned Action posits that behavior is ultimately determined by attitudes, but that the link is not direct. The contextual system that includes, in this case, beliefs about the child, beliefs about broader social norms with regard 137 to children and socialization, parental motivations to comply with those norms, and the individual parent's weighting of attitudes and norms, needs to be understood if behavior is to be accurately predicted. Parental affect is another important mediator between attitudes and behavior. Dix (1991) has argued that parental affect, broadly conceived, is the single most important influence on parent-child interaction. He writes: Once aroused, emotions activate and organize cognition, induce motives to maintain or change immediate circumstances, communicate to children parents' intentions and reactions, and prime action tendencies likely to be needed....1f emotion is insufficient, adaptive systems may be poorly engaged: Parents may fail to monitor and process child-related information effectively, may be unmotivated for parenting tasks, may be unprepared to display needed behaviors, and may fail to communicate intentions and reactions in ways that engage and motivate children. On the other hand, if emotion is negative and excessive, it may disrupt cognition and decision making, activate motives that conflict with responding to children's needs, induce action tendencies that overwhelm processes that normally control negative emotion, and lead to negative expressive behaviors that elicit fear, anger, or disorganization from children....Parenting in many distressed families appears to be characterized by these deficits. (Dix, 1991, p. 16) The context-free structure of instruments like the CRPR has also been criticized. The immediate context is likely to be pivotal in 138 determining the degree of coherence between generic attitudes and behavior (Fazio, 1986). Taken together, these criticisms dictate that parental responses to the CRPR need to be interpreted as reflecting a generic underlying model of socialization, a schema which should not be understood as a report of parent-child interaction directly, and which cannot be understood without considering the broader context of the parent-child relationship. However, this reconceptualization does not indicate that CRPR variables are without merit, and a better understanding of these underlying beliefs would have important clinical implications given the assumption that they do contribute to behavior. Problematic (i.e., pathogenic) beliefs would then be a serious source of developmental risk (potentially one of many) for children growing up in these families. Secialization attitudee and parental peyenopathology. Simply put, parental alcoholism and severity of parental alcoholism were unrelated to socialization attitudes on the CRPR. The only exception to this clear trend was for maternal nurturance. Alcoholic mothers reported less nurturant attitudes than nonalcoholic women who were married to alcoholic men. This was replicated when examining maternal alcoholism severity and the difference was found to be related to differences between the severely alcohol dependent women and the mothers in father- only-alcoholic families. The MANOVAs preceding both of these analyses were nonsignificant, indicating that the finding must be considered tentative, but it is consistent with a theoretical formulation that these women would be more affectionate with their sons in compensation for an impaired father-son relationship and/or a distressed marital relationship (Engfer, 1988). It is also interesting because maternal 139 nurturance was found to be associated with very few other areas in this investigation and so remains something of a puzzle. Examination of the correlational results with individual psychopathology measures also revealed a paucity of significant associations. Global maternal psychopathology was positively associated with maternal authoritarianism and negatively associated with parental agreement, and maternal antisocial behavior was negatively associated with nurturant attitudes. For fathers, current depressed mood was negatively associated with paternal nurturance. Each of these individual findings is consistent with predictions about the effect of parental psychopathology, but the pattern of findings is far less than we would expect given the literature on the effects of parental psychopathology on the parent-child relationship. Seeialization attitudes and iamily chatecterietics. There were also few significant associations between the parenting attitude variables and family characteristics. For both mothers and fathers, higher levels of reported child behavior problems were associated with lower levels of nurturance. The importance of child behavior and temperament in affecting parenting attitudes and behavior has begun to be recognized (Bell 6 Harper, 1977; Dix, Ruble, Grusec, 6 Nixon, 1986) and deserves greater attention, especially from longitudinal research which can begin to sort out causality. The only other associations to the family characteristics variables were for mothers' ratings of family supportiveness, which were related negatively to maternal authoritarianism and positively to parental agreement. One might expect to find greater coherence between these family-level variables and parent-child interactional variables, 140 either observed or self-reports of specific behaviors. This lends further support to the contention that the CRPR responses reflect a socialization schema per se, not broader family behavior or parental psychiatric disturbance. Soeialieation ettitudee ene gemograpnic eneracteristics. In contrast to parental psychopathology and family environment measures, socialization attitudes, especially parental authoritarianism, were found to be consistently and, in some cases, strongly related to demographic variables. More authoritarian attitudes were reported by mothers of lower family SES and income, less education, and greater immaturity. For fathers, these same variables, with the exception of family SES, were similarly related. Higher interparent agreement was positively associated with family income and maternal age, while paternal nurturance was associated positively with maternal education and negatively with the age of the child. Only maternal nurturance was unrelated to family demography. These findings are in keeping with the literature reviewed above regarding social class and the earlier findings of Davies et a1. (1991). There can be no question that socialization attitudes show consistent cultural and subcultural variations. Burland (1984) has gone so far as to assert that subcultures can be defined by the child-rearing practices that perpetuate them. However, the conclusion that socialization attitudes vary systematically according to cultural divisions leaves unmet the goal of understanding the sources of parenting attitudes for any given individual. It is clear from this investigation that current psychopathological characteristics of the parent and the family are associated with little, if any, of this variability. 141 Secielizatien senemas, family onetactetistics, and parental psychopathelogy. This investigation found little evidence that socialization attitudes per se were related to parental psychopathology or family characteristics. Rather, variability in socialization attitudes was related to subcultural differences as measured by socioeconomic status, and there was some evidence of a connection to child characteristics (although the circumscribed sampling of child behavior measures limit the strength of this conclusion). The primary exception to this pattern was maternal nurturance, which was the only socialization variable which was unrelated to demography and also the only one to evidence at least a weak relationship to parental alcoholism and antisociality. Although, it must be interpreted with caution due to a nonsignificant MANOVA, there was a consistent pattern indicating that nonalcoholic women who were married to alcoholic men reported more nurturance with their sons than did the alcoholic women. This is consistent with systemic formulations that suggest the mother and son would become closer in compensation for a poor marital relationship (Engfer, 1988) and/or a poor father-son relationship. The reason(s) that maternal nurturance appears to be a unique domain are not entirely clear. At the least, there appears to be little subcultural variation (and hence a certain universality) in the amount of affection and support mothers feel they should provide. The suggestion of connections to psychopathology introduces the possibility that internal preoccupation may lead to lower levels of nurturant attitudes. This is also likely to be interactive with child 142 characteristics, as suggested by the negative relationship between nurturant attitudes and child behavior problems. It is clear that there is a great deal of variability in parental attitudes that remains to be explained. Two areas in particular stand out as possible avenues of exploration in this regard. One is characteristics and behavior of the child, as we have already discussed. The second is the parent's own early experiences in their interactions with their own parents. Psychoanalytic theorists have traditionally viewed parenting attitudes and behaviors as arising from the parent's own early experiences. Sigmund Freud (1921) and Anna Freud (1937) viewed the process in pathological terms of repetition compulsion and "identification with the aggressor", respectively. More recent psychoanalytic theorists have cast the same process more neutrally as related to the process of personality structure developed in childhood (Benedek, 1959; Terman, 1984). Of course, social learning theory based explanations (e.g., modeling) (Bandura, 1977) are also consistent with the observation of intergenerational continuity. The intergenerational transmission of parenting attitudes holds promise for understanding both the origins of parents' idiosyncratic attitudes and explaining the perpetuation of cultural differences in child-rearing attitudes. Another set of analyses out of this project (Cruise, 1991) is exploring these questions as they pertain to the transmission of aggression across the generations. These analyses provide some support for this line of causal reasoning, but an extensive review is beyond the scope of the present discussion. 143 WWW As this discussion has made clear, the primary limitation of the current study is the reliance on a single measure of the parent-child relationship; one which can only safely be presumed to tap parents' attitudes in a generic, noncontextual way, and by way of assessing the cognitive rather than the behavioral level of schematization. While this is important information, it is several steps removed from the actual parent-child interactions which make up the relationship. A complete understanding of this process would require an investigation of when and under what circumstances these parenting attitudes are or are not translated into behavior. Related to this, the use of primarily parent-report measures introduces the possibility of method variance contributing to the effects found. While consistent results were obtained when external observer ratings were available (i.e., for parental psychopathology), more diversity in the sources of information, including direct behavioral observation, would be desirable. The cross-sectional design of this study is another limitation to interpretation. It prevents the untangling of causal threads in understanding the many characteristics which were found to be associated with parental alcoholism and symptom severity. In addition, the small number of controls available relative to the number of alcoholic families reduced the statistical power of the analyses, although the sample sizes were obviously sufficient to reveal differences in all domains of interest except socialization attitudes. Finally, the inclusion of alcoholic fathers only from a convicted drunk driving population is cause for care in interpretation. As discussed above, the more antisocial subtypes of alcoholism are 144 overrepresented in this sample. Similarly, the alcoholic women in this stnniy were all married to alcoholic men, and may differ in important ways from other subgroups of alcoholic women. Additionally, although these women meet diagnostic criteria, in virtually all instances their alcoholism has been "silent" in that it has not been marked in any way i1) the public arena. Fntnte Ditectione Future work in the area of parent-child relationships in alcoholic fandlies should go beyond socialization attitudes to include measures of actual parent-child interaction from the perspectives of parents, cflaildren, and external observers. In a related vein, the role of attitudes in determining behaviors needs to be more fully explored. Self-reports of parenting behaviors need to be made more contextual, 'with enough detail included so that the parent can (1) draw on specific memories of past behavior, and (2) think through what they would actually do. More attention needs to be paid to the role of child characteristics in determining parenting behaviors. In addition, the effects of dual diagnoses within the sample of alcoholic men and women (as well as psychiatric disorders among the controls) deserves careful attention. It may also be fruitful to begin to consider differential effects of the developmental stage of the alcoholic family (as has been described by Steinglass et al., 1987) and the type of parental alcoholism (e.g., antisocial alcoholism, negative affect alcoholism; see Zucker, 1987). A substantial subset of these research questions can be addressed as the longitudinal study progresses. For the moment, these relationships remain uncharted. 145 The processes and effects of the parent-child relationship need to be followed longitudinally from this very early stage of family development. This work has shown that many factors are already in place that would appear to place preschool-age children growing up in alcoholic homes at risk for developing trouble (at least with regard to parent and family characteristics). This risk appears to be greater when both parents are alcoholic and/or when the severity of the alcoholism is greater. But the ultimate answers to the questions of risk and resilience in these children lie in the cumulation of experiences over time. Longitudinal work holds the promise of completing this picture. C c s n Compared to nonalcoholic families from the same neighborhoods, young families with fathers who are convicted drunk drivers were found to have higher levels of psychopathology (both drinking-specific and nondrinking-specific), fewer resources, less parental achievement, and a less supportive environment. When families also contained an alcoholic mother, both parents' levels of functioning were significantly lower, the family was seen as more conflicted, and there were reports of greater child behavioral difficulty. Across the alcoholic families, severely alcohol dependent fathers reported higher levels of psychopathology, child behavior difficulties, and family conflict and crises. Maternal alcoholism severity was associated with more antisociality, more reported child behavior problems, and a less supportive environment. These results clearly indicate that young alcoholic families show pervasive trouble that would be expected to place the children growing 146 up in these homes at risk, especially when both parents are alcoholic and/or the severity of alcoholism is greater. It was anticipated that these deficits would be associated with marked changes in socialization attitudes, given considerable evidence in the literature that all four of these areas (parental psychopathology, lower SES, family conflict, and child behavior problems) are associated with problems in the parent- child relationship. On the contrary, consistent differences in socialization attitudes were found only related to SES and related demographic factors. This was most clear for authoritarian attitudes of both parents, but SES was also related to paternal nurturance and to the degree of agreement between parents in a given family. Only maternal nurturance was unrelated to the family's subculture. Careful consideration of the instruments used suggested that this rather surprising set of findings may be due to the assessment of parenting attitudes as cognitive schemas, even though the attitudinal statements emphasized a behavioral emphasis. The relationship of attitudes to behavior is complex and is worthy of investigation in its own right. Such investigation will be necessary before the present results can fully be understood. We may conclude, however, that socialization attitudes (i.e., the model that parents bring to child-rearing) do not vary as a function of the number or severity of alcoholic parents in a family. This would imply that group differences that are observed in parent-child behavioral interactions are driven by other characteristics of the parent, the child, their style of interactions, or the family system as a whole. 147 These finding have important clinical implications. The findings related to resources, parental functioning, family environment, and child behavior certainly indicate that problematic processes are already well-established in these families even while the children are at a relatively young age. Drunk-driving can justifiably be conceived as marking a family-level problem, and interventions should be targeted at that, the family level, rather than being focused exclusively on the cessation of drinking. Family interventions should already be underway for identified alcoholic families before the children are of school-age. While school is the more typical route of identification of family difficulties, the present study indicates that this may be too long to wait. A father's drunk driving conviction is a unique and powerful means of identifying such families at risk. These findings also indicate that child-rearing attitudes per se need not be a major component of such intervention, at least as a matter of course. Rather, parent-child inteteetinne need to be assessed on a case-by-case basis. Even in those cases where deviant or pathogenic child-rearing attitudes are present, the more effective intervention is likely to be directly into the interactions (Patterson 6 Chamberlain, 1988). Attitudes can then change as a result of learning new skills and techniques, or may need to be addressed directly when they impede clinical progress. LI ST OF REFERENCES LIST OF REFERENCES Ablon, J. (1976). Family structure and behavior in alcoholism: A review of the literature. In B. Kissin & H. Begleiter (Eds.), Ihe_hiclcgy cf alccholism (vol. 4). New York: Plenum. Achenbach, T.M., & Edelbrock, C.S. (1983). Menual fog che Chile Behavicr Checklist anc Revieeg Chile Behavic: Profile. Burlington, VT: Department of Psychiatry, University of Vermont. Alexander, C. N. (1967). Alcohol and adolescent rebellion. Social Eorces, 3;, 542-550. American Psychiatric Association (1987). Diagnostic ang sceciscicel manual of mentel disordegs (3rd ed., revised). Washington, DC: Author. Atkeson, B.M., Calhoun, K.S., Resick, P.A., & Ellis, E.M. (1982). Victims of rape: Repeated assessment of depressive symptoms. l2grnal_2f_Q2nsulLinz_end_§lisisel_£§12helegx. 29. 96-102. Baker, S.M. (1945). Social case work with inebriates. In Alcohol, §cience end ecciecy (pp. 419-436). New Haven, CT: Journal of Studies on Alcohol. Bandura, A. (1977). Sociel learging cheogy. Englewood Cliffs, NJ: Prentice-Hall. Barber, B.K. (1987). Marital quality, parental behaviors, and adolescent self-esteem. Femily Perspective, ll, 244-268. Barnes, G.M. (1977). The development of adolescent drinking behavior: An evaluative review of the impact of the socialization process within the family. Adolescence, 1;, 571-591. Barnes, G.M., Farrell, M.P., & Cairns, A. (1986). Parental socialization factors and adolescent drinking behaviors. Journal of Marriage end the Eamily, flfi, 27-36. Baxter-Hagaman, J.A. (1986). Ycung alcoholic iamilies and the ansm ssio s ' Env ronmental am' ' te act n differen2es_fr2m_thesu§Q_Lenzi£udinal_$tudx- Unpublished doctoral dissertation, Michigan State University, East Lansing. 149 150 Bech, P., Gram, L.F., Dein, E., Jacobsen, 0., Vitger, J., & Bolwig, T.G. (1975). Quantitative ratings of depressive states. cta Eeychietgice Scandinevice, fil, 161-170. Beck, A.T., & Beck, R.W. (1972). Screening depressed patients in family practice: A rapid technique. Ececgnegnece_hegicine, 52, 81-85. Beck, A.T., Steer, R.A., & Garbin, M.G. (1988). Psychometric properties of the Beck Depression Inventory: Twenty-five years of evaluation. Clinical Eeychclcgy Review, §, 77-100. Bell, R.Q., & Harper, L.V. (1977). Chile effects on adults. Hillsdale, NJ: Erlbaum. Belsky, J. (1981). Early human experience: A family perspective. Developmental Esychclogy, ll, 3-23. Belsky, J. (1984). The determinants of parenting: A process model. Chile Develcnnent, fifi, 83-96. Belsky, J., Hertzog, C., & Rovine, M. (1986). Causal analyses of multiple determinants of parenting: Empirical and methodological advances. In M. E. Lamb, A. L. Brown, & B. Rogoff (Eds.), Advences in cevelonmencel nsychclcgy (vol. 4, pp. 153-202). Hillsdale, NJ: Lawrence Erlbaum Associates. Billings, A.G., & Moos, R.H. (1983). Comparisons of children of depressed and nondepressed parents: A social-environmental perSPective. WWW. LI. 463-486. Block, J. (1957). A comparison between ipsative and normative ratings of personality. ournal o bno ma nd a s o , fig, 50-54. Block, J. (1961). Ihe Q-sort method in nereonality aseesemenc end neychiatric research. Springfield, IL: Charles C. Thomas. Block, J. (1971). Lives thrcugh cime. Berkeley, CA: Bancroft. Block, J., Block, J.H., & Reyes, S. (1988). Longitudinally foretelling drug usage in adolescence: Early childhood personality and environmental precursors. Chile_2exelcnnenc, 22, 336-355. Block, J.H. (1965). e C ld- ea n actices Re ort RP ° set 0 0 ‘v fo. ,- a-V . ., . 02 - t2, soc a : o: s -e end values. Unpublished manual, Institute of Human Development, University of California, Berkeley. Block, J.H., Block, J., & Morrison, A. (1981). Parental agreement- disagreement on child-rearing orientations and gender-related personality correlates in children. thld Qevelcnmenc, 22, 965- 974. 151 Block, J.H., Jennings, P.H., Harvey, E., & Simpson, E. (1964). Interaction between allergic potential and psychopathology in childhood asthma. Esychceometic Medicine, Zé, 307-320. Booz-Allen & Hamilton, Inc. (1974). 5 en 0 he eeds a rea22r2ee_f2r_2hildren_2f_aleeh21i2_naren£§. Rockville, MD: NIAAA. Braithwaite, R.L., & Gordon, E.W. (in press). Succese egainec the odds. Howard University Press. Brisbane, F.L. (1987). Divided feeling of black alcoholic daughters. Als2bel_Beal£h_end_fie§earsh_flerld. 12. 48-50. Bronfenbrenner, U. (1961). Some familial antecedents of responsibility and leadership in adolescents. In L. Petrullo & B. M. Bass (Eds.), Leadepship eng integpersonel hehevicg. New York: Holt. Bronfenbrenner, U. (1986). Ecology of the family as a context for human development: Research perspectives. Dexelcpmengel_£§12hglggx. 22. 723-742. Bronfenbrenner, U., & Crouter, A.C. (1982). Work and family through time and space. In C. Hayes & S. Kamerman (Eds.), Eemilies the; work; ghilgmen_in_e_chenging_nc;ld. Washington, DC: National Academy of Sciences. Brook, J.S., Whiteman, M., & Gordon, A.S. (1983). Stages of drug use in adolescence: Personality, peer, and family correlates. Developmental Psychology, 12, 269-277. Brook, J.S., Whiteman, M., Gordon, A.S., & Brook, D.W. (1981). Paternal determinants of male adolescent marijuana use. Degelcpmencel Esxsbelezx. 11. 841-847. Brook, J.S., Whiteman, M., Gordon, A.S., & Brook, D.W. (1984). Paternal determinants of female adolescents' marijuana use. Developmental Zsycholcgy, ZD, 1032-1043. Brook, J.S., Whiteman, M., Gordon, A.S., & Cohen, P. (1986). Some models and mechanisms for explaining the impact of maternal and adolescent characteristics on adolescent stage of drug use. Develcpmentel Eeychclogy, 22, 460-467. Brown, P., & Tooley, J. (1989). Alcoholism in the black community. In G.W. Lawson and A.W. Lawson (Eds.), co d ubsta in epeciel pcpulations. Rockville, MD: Aspen Bushman, B.J., & Cooper, H.M. (1990). Effects of alcohol on human aggression: An integrative research review. Eeychclcgicel MD: 191! 341-354- 152 Cahalan, D., Cisin, I.H., & Crossley, H.M. (1969). Ame;icen_epinhing r t ce ' a ° a ud of ehav o a tt'tud . New Brunswick, NJ: Rutgers Center of Alcohol Studies. Carroll, B.J., Fielding, J.M., & Blashki, T.G. (1973). Depression rating scales: A critical review. W. 2.5.. 361-366. Cattell, R.B. (1966). The scree test for the number of factors. hnlcivegiece Deheviczel Reseagch, l, 245-276. Cicchetti, D., & Rizley, R. (1981). Developmental perspectives on the etiology, intergenerational transmission, and sequelae of child maltreatment. New Directicne fie; thlc Develcpmenc, ll, 31-55. Clair, D., & Genest, M. (1987). Variables associated with the adjustment of offspring of alcoholic fathers. u 8 ud e o , 3D, 345-355. Cloninger, C.R. (1987). Neurogenetic adaptive mechanisms in alcoholism. Science, gee, 410-416. Cloninger, C.R., Bohman, M., & Sigvardsson, S. (1981). Inheritance of alcohol abuse. Archives cf genepel Deychiecny, DD, 861-868. Cloninger, C.R., & Reich, T. (1983). Genetic heterogeneity in alcoholism and sociopathy. In 8.8. Kety, L.P. Rowland, R.L. Sidman, & S.W. Matthysse (Eds.), et 5 of eur a giecmceme. New York: Raven Press. Cohen, R.S., & Weissman, S.H. (1984). The parenting alliance. In R.S. Cohen, B.J. Cohler, & S.H. Weissman (Eds.), Perenthcoe; A pexchodxnemic perspeccive. New York: Guilford Press. Cohen, S., & Wills, T.A. (1985). Stress, social support, and the buffering hypothesis. Esxchclogical Dullecin, gD, 310-357. Cohler, B.J. (1984). Parenthood, psychopathology, and child care. In R.S. Cohen, B.J. Cohler, & S.H. Weissman (Eds.), 2£I2n£h22§;_8 psychodxnemic pecspeccive. New York: Guilford Press. Cork, R.M. (1969). c ‘ of c w elcohclic perencs. Toronto: Addiction Research Foundation. Cotton, N. (1979). The familial incidence of alcoholism: A review. Journal cf Scueies cn Alcchcl, 52, 89-116. Crook, T., Raskin, A., & Eliot, J. (1981). Parent-child relationships and adult depression. Child Developmenc, e2, 950-957. Cruise, K. (1991). essiv v hree enerat onS' e t o e e at r n be wee a e ts the c 1d e e ch dren's chi d 153 xcnng_elcchclic_femiliee. Unpublished dissertation, Michigan State University, East Lansing. Davies, W.H. (1989). e elat on i etwee arent life troub es a d - ea atter s o a o o a es. Unpublished master's thesis, Michigan State University, East Lansing. Davies, W.H., Noll, R.B., & Bukowski, W.M. (1991, August). ghilghccg :_g ‘7 ild o: ' ° VO‘! ‘ o :_ e“: th - Paper presented at the meeting of the American Psychological Association, San Francisco. Davies, W.H., Noll, R.B., DeStefano, L., Bukowski, W.M., & Kulkarni, R. (1991). Differences in the child rearing practices of parents of children with cancer and controls: The perspectives of parents and professionals. Journal of Dediatzic Eeychclogy, lg, 295-306. Davies, W.H., Zucker, R.A., Noll, R.B., & Fitzgerald, H.E. (1989, June). arental s cho at 010 and chi d- ea n ract ces n ou elccholic familiee. Paper presented at the meeting of the Research Society on Alcoholism, Beaver Creek, CO. Davies, W.H., Zucker, R.A., Noll, R.B., & Fitzgerald, H.E. (1991). Eeglx o _ali : 01 ..ac _ces , : Cogo am_ eS' I - ‘_a to i- ._ d- ea in atte t e o r en psychcpathology. Manuscript submitted for publication. Deal, J.E., Halverson, C.F., Jr., & Wampler, R.S. (1989). Parental agreement on child-rearing orientations: Relations to parental, marital, family, and child characteristics. thlc_Degelcpmenc, QD, 1025-1034. Demone, H.W. (1973). The nonuse and abuse of alcohol by the male adolescent. In M.E. Chafetz (Ed.), Epoceedingc oi che Secong 523W. Washington. DC: DREW. Deutsch, C., DiCicco, L., & Mills, D. J. (1982). Services for children of alcoholic parents. In Alco hcl end health monoggeph no , 03; evention ntervent on a t ea t Co cern Rockville, MD: NIAAA. Dix, T. (1991). The affective organization of parenting: Adaptive and maladaptive processes. Esxchological Dulletin, llD, 3-25. Dix, T., Ruble, D.N., Grusec, J.E., & Nixon, S. (1986). Social cognition in parents: Inferential and affective reactions to children of three age levels. Child Developmenc, 21, 879-894. Donovan, J.E., & Jessor, R. (1978). Structure of problem behavior in adolescence and young adulthood. Jonpnal cf Concnlcing ane 21W. 21. 890-904. 154 Donovan, J.E., Jessor, R., & Jessor, S.L. (1983). Problem drinking in adolescence and young adulthood: A follow-up study. lcngnel_cf coh l 44 109-137. Downey, C., & Coyne, J.C. (1990). Children of depressed parents: An integrative review .Esxchelegical_§211etin. IQQ. 50-76. Drake, D.E., & Vaillant, G.E. (1988). Predicting alcoholism and personality disorder in a 33-year longitudinal study of children of alcoholics. British Jcnpnal of Adciction, DD, 799-807. Egeland, B., Jacobvitz, D., & Sroufe, L.A. (1988). Breaking the cycle of abuse. ghilc Development, ca, 1080-1088. Elder, G.H., Jr. (1981). History and the life course. In D. Bertoux (Ed.), Dicgpaphy anc eociecy; The life DISEQII epppcech in the eccial sciencee. Beverly Hills, CA: Sage. Engfer, A. (1988). The interrelatedness of marriage and the mother-child relationship. In R.A. Hinde & J. Stevenson-Hinde (Eds.), Relationshipe within families; Mncuel infiluencee. Oxford: Clarendon. Featherman, D.L., & Hauser, R.M. (1977). Commonalities in social stratification and assumptions about status mobility in the United States. In R.M. Hauser & D.L. Featherman (Eds.), Ihe_p;cce§§_2£ ssratifieaLign;_IrenQ§_anQ_enelz§is. New York: Academic Press. Feighner, J., Robins, L., Guze, 8., Woodruff, R., Winokur, G., & Munoz, R. (1972). Diagnostic criteria for use in psychiatric research. Anchives oi Genegal Dsychiatgy, an, 57-63. Fellios, P.G. (1989). Alcoholism in women: Causes, treatment, and prevention. In G.W. Lawson & A.W. Lawson (Eds.), o a d enbstance ehnee in epecial populatione. Rockville, MD: Aspen. Fox, R. (1962). Children in the alcoholic family. In W.C. Bier (Ed.), Problems in addiction; Alcohol and crug addiction. New York: Fordham University Press. French, A.P. (1983). Parental depression and child-family development: A systems approach. In R.L. Morrison (Ed.), Children of cepgessed parents; gisk, identificacion, and intepgention. New York: Grune & Stratton. Gjerde, P.F. (1988). Parental concordance on child-rearing and the interactive emphases of parents: Sex-differentiated relationships during the preschool years. Develcpmentel Esxchology, 23, 700-706. Goodwin, D. (1979). Alcoholism and heredity. Archives cf general Psychiatry. lg. 57-61. 155 Goodwin, D.W., Schulsinger, F., Hermansen, L., Guze, S.B., & Winokur, G. (1973). Alcohol problems in adoptees raised apart from alcoholic biological parents. c ves Ge er s chiat , 2D, 238-243. Goodman, S.H. (1987). Emory University Project on Children of Disturbed Parents. WM. 1.3. 411-423. Gove, W.R., & Crutchfield, R.D. (1982). The family and juvenile delinquency. Sociologicel Qnepteply, 3;, 301-319. Hall, R.L., Hesselbrock, V.M., & Stabenau, J.R. (1983). Familial distribution of alcohol use: II. Assortative mating of alcoholic probands. Dehavior Genetics, lg, 373-382. Halpern, R. (1990). Poverty and early childhood parenting: Toward a framework for intervention. WW. DD, 6-18. Hamburg, D. (1985). educ the asu lties o e e' eve t ve crientacion. President's Essay, Annual Report of the Carnegie Corporation, New York. Hamilton, M. (1960). A rating scale for depression. Journal of Neuzology, Neurosurgegy, anc Esychiecpy, 2;, 56-62. Hamilton, M. (1969). Standardized assessment and recording of depressive symptoms. WM. 7.2. 201-205. Hammen, C.L. (1980). Depression in college students: Beyond the Beck Depression Inventory. WW Mica. aa. 126-128. Handel, G. (1970). Sociological aspects of parenthood. In E.J. Anthony & T. Benedek (Eds.), Derenthood; lts psychclogy end psychopathology. Boston: Little, Brown, & Company. Harder, D., Kokes, R., Fisher, L., & Strauss, J. (1980). Child competence and psychiatric risk: IV. Relationships of parent diagnostic classifications and parent psychopathology severity to child functioning. Journal of Nervous and Menta D e s , ice, 343-347. Harper, F.D. (1983). Alcohol use and alcoholism among Black Americans: A review. In T.D. Watts & R. Wright (Eds.), Dlack alcoholism; Towapd a comprehensive understanding. Springfield, IL: Charles C. Thomas. Harper, F., & Dawkins, M. (1976). Alcohol and blacks: Survey of periodical literature. British gcncnel cf Addiction, ll, 327-334. Harrington, D.M., Block, J.R., & Block, J. (1987). Testing aspects of Carl Rogers's theory of creative environments: Child-rearing antecedents of creative potential in young adolescents. Jou of Bensonalicy end §ocial Esychclogy, e2, 851-856. 156 Hecht, M. (1973). Children of alcoholics are children at risk. Amenicen Jcnrnal cf hugging, 1, 1764-1767. Helzer, J.E., & Pryzbeck, T.R. (1988). The co-occurrence of alcoholism with other psychiatric disorders in the general population and its impact on treatment. i22rnal_2£_§£sdie§_en_A122hel. 32. 219-224. Hirsch, B.J., Moos, R.H., & Reischl, T.M. (1985). Psychosocial adjustment of adolescent children of a depressed, arthritic, or normal Parent. J22rnal_ef_Ahnermal_£§xch212£x. 23. 154-164. Holden, G.W., & Edwards, L.A. (1989). Parental attitudes toward child rearing: Instruments, issues, and implications. Eeychclcgicel figlletin. 199. 29-53- Holland, B.S., & Copenhaver, M.D. (1988). Improved Bonferroni-type multiple testing procedures. Esychological Dulletin, lDA, 145- 149. Holm, S. (1979). A simple sequentially rejective multiple test procedure. S and n v' ' Jou a o Statis ' s, D, 65-70. Holmes, S.J., & Robins, L.N. (1988). The role of parental disciplinary practices in the development of depression and alcoholism. Esychiatgy, Dl, 24-36. Hops, H., Biglan, A., Sherman, L., Arthur, J., Friedman, L., & Osteen, V. (1987). Home observations of family interactions of depressed women. Joupnal cf gonsnlcing anc Qlinical gsycholcgy, fig, 341-346. Jackson, J.K. (1962). Alcoholism and the family. In D.J. Pittman & C.R. Snyder (Eds.), Society, cnlcnpe, end dzinhing peggezné (pp. 472- 492). New York: John Wiley & Sons. Jacob, T., Favorini, A., Meisel, S.S., & Anderson, C.M. (1978). The alcoholic's spouse, children and family interactions: Substantive findings and methodological issues. Journel of Stuciee on Alcchcl, 12, 1231-1251. Jacob, T., Krahn, C.L., & Leonard, K. (1991). Parent-child interactions in families with alcoholic fathers. gcp;nel_e£_§cnenlcing_eng Clinical_£sxehelezx. :2. 176-181. Jacob, T., & Leonard, K. (1986). Psychosocial functioning in children of alcoholic fathers, depressed fathers and control fathers. lengnel of Studies on Alcohol, A1, 373-380. Jacob, T., & Leonard, K. (1991). Experimental drinking procedures in the study of alcoholics and their families: A consideration of ethical issues. J2srnal_2f_Q2ns2lting_ans_§linical_2sxchelezx. 52. 249- 255. 157 Jacob, T., & Seilhamer, R.A. (1987). Alcoholism and family interaction. In T. Jacob (Ed ). Eaailx_inrera2tien_an§_2§xsheeatheleg1i Iheogiee, mechode, end fincings. New York: Plenum. Jensen, G. (1972). Parents, peers, and delinquent action: A test of the differential association perspective. American Doupnal of Sociology. Zfi. 562-575. Jessor, S.L., & Jessor, R. (1974). Maternal ideology and adolescent problem behavior. Developmencal Eeychclcgy, lD, 246-254. Jones, M.C. (1968). Personality correlates and antecedents of drinking patterns in adult males. ou a Con n a d C nica Eeychology, D2, 2-12. Jones, M.C. (1971). Personality antecedents and correlates of drinking patterns in women. ourna Consu in a d Clin ca s 010 , DD, 61-69. Joereskog, K.G., & Soerbom, D. (1984). LIDRDD Vl neep'e guide. Mooresville, IN: Scientific Software. Kazdin, A.E. (1987). Treatment of antisocial behavior in children: Current status and future directions. Esychological Dullecin, l02, 187-203. Kendall, P.G., & Beutler, L.E. (1991). Editorial comment on Jacob et al., ”Parent-child interactions in families with alcoholic fathers”. Journal of Donsnlcing end Qlinicel Esychclcgy, Dg, 183. Kidd, K.K., & Weissman, M.M. (1978). Why we do not yet understand the genetics of affective disorders. In J.0. Cole, A.F. Schatzberg, & S.D. Frazier (Eds.), Deppession; Biology, psychodynemice, enc tgeatment. New York: Plenum Press. Kochanska, G. (1990). Maternal beliefs as long-term predictors of mother-child interaction and report. Dhilc Development, D1, 1934- 1943. Kochanska, G., Kucynski, L., & Radke-Yarrow, M. (1989). Correspondence between mothers' self-reported and observed child-rearing practices. Child Development, DD, 56-63. Kochanska, G., Kucynski, L., Radke-Yarrow, M., & Welsh, J.D. (1987). Resolution of control episodes between well and affectively ill mothers and their young children. Doupnal oi Abnogmal Dhilg Esychology, lD, 441-456. Kohn, M.L., Naoi, A., Schoenbach, C., & Schooler, C. (1990). Position in the class structure and psychological functioning in the United States, Japan. and Poland. Amerisan.lesrnal_2f_fieeielezx. 25. 964- 1008. 158 Koocher, G.P. (1991). Questionable methods in alcoholism research. l2nrnal_2f_Q2nanlnina_enn_§lininal_£§xehnleax. 52. 246-248. Krauthamer, C. (1974). The personality of alcoholic mothers and their children: A study of their relationship to birth-order, mother- child attitude, and socioeconomic status. Dissertation Abetrecce International. 15. 51983. Krauthamer, C. (1979). Maternal attitudes of alcoholic and nonalcoholic uPper middle class women. 1nnernani2nal_Jnnrnal_2f_£he_annietien§. 13, 639-644. Kronenberger, W.C., & Thompson, R.J., Jr. (1990). Dimensions of family functioning in families with chronically ill children: A higher order factor analysis of the Family Environment Scale. Dcn;nel_cf glinical Child Eeychclcgy, lg, 380-388. Kuyler, P.L., Rosenthal, L., Igel, G., Dunner, D.L., & Fieve, R.R. (1980). Psychopathology among children of manic depressive patients. Biological Deychiaczy, lD, 589-597. Lee, C.M., & Gotlib, I.H. (1989). Maternal depression and child adjustment: A longitudinal analysis. Dcurnal of Abnczmal mm, 2g, 78-85- Lin, C.C., & Fu, V.R. (1990). A comparison of child-rearing practices among Chinese, immigrant Chinese, and Caucasian-American parents. shiln_nexelennent. él. 429-433. Loeber, R. & Stouthamer-Loeber, M. (1986). Family factors as correlates and predictors of juvenile conduct problems and delinquency. In M. Tonry & N. Morris (Eds.),D;ime_ene_Dnecice (vol. 7). Chicago: University of Chicago Press. Luthar, S.S., & Zigler, E. (1991). Vulnerability and competence: A review of research on resilience in childhood. Ame;icen_Dcn;nel_c§ Qrthsnszshiatrx. 91. 6-22. McCord, J. (1988). Identifying developmental paradigms leading to alcoholism. J2nrnnl_ef_§£ndiss_2n_nlconnl. $1. 357-362. McCord, J. (1990). Crime in moral and social contexts--The American Society of Criminology, 1989 Presidential Address. C i o , 2D, 1-26. McCord, W., & McCord, J. (1960). Origins of elcoholism. Stanford, CA: Stanford University Press. McCord, W., & McCord, J. (1962). A longitudinal study of the personality of alcoholics. In D. J. Pittman & C. R. Snyder (Eds.), Dociecy, 2nlnnre1_nnn_nrinkinz_na££erns (pp. 413-430). New York: John Wiley & Sons. 159 McLoyd, V.C. (1990). The impact of economic hardship on black families and children: Psychological distress, parenting, and socioemotional development. Chile Developmenr, Dl, 311-246. Merikangas, R.R., Weissman, M.M., Prusoff, B.A., Pauls, D.L., & Leckman, J.F. (1985). Depressives with secondary alcoholism: Psychiatric disorders in offspring. Journal of Dtudies on Alcohol, DD, 199- 204. Midanik, L. (1983). Familial alcoholism and problem drinking in a national drinking practices survey. Aegicriye_Deheyicre, D, 133- 141. Miller, D., & Jang, M. (1977). Children of alcoholics: A 20-year longitudinal study. Sccial Worh Eeeeerch enD Abscraccs, lD, 23-29. Monnelly, E.P., Hartl, E.M., & Elderkin, R. (1983). Constitutional factors predictive of alcoholism in a follow-up of delinquent boys. Jonrnal of Studies cn Alcohol, AD, 530-537. Moos, R.R., & Moos, B.S. (1976). A typology of family environments. Eamily Erocess, lD, 357-371. Moos. R.B.. & Moos. B.S. (1981). W. Palo Alto, CA: Consulting Psychologists Press. Mueller, G.W., & Parcel, T.L. (1981). Measures of socioeconomic status: Alternatives and recommendations. Chile Developmenc, D2, 13-30. Neff, J.A. (1986). Alcohol consumption and psychological distress among U.S. Anglos, Hispanics, and blacks. Alcchcl_enD_Alcchcliem, Zl, 111-119. Neisser, U. (1976). chnirlgn_engnreeliry. San Francisco: Freeman. Nock, S.L., & Rossi, P.H. (1979). Household types and social standing. Dcciel Eorcee, DZ, 1325-1345. Noll, R.B., & Zucker, R.A. (1980). e st tu of a av in different populations. Unpublished paper, Department of Psychology, Michigan State University, East Lansing. Noll, R.B., & Zucker, R.A. (1983). eve 0 me ta ind ro a elcoholic vulnerability erndy. Paper presented at the meeting of the American Psychological Association, Anaheim, CA. Noll, R.B., Zucker, R.A., Fitzgerald, H.E., & Curtis, W.J. (1990). Deere-11'! -' ! °l__!°_ 0 11;. ' 0 -.- 1' 0.- =. o 0 c ahe Ihe_eerly_chilghccn_yeere. Man script submitted for publication. Noll, R.B., Zucker, R.A., & Greenberg, 6.8. (1990). Identification of alcohol by smell among preschoolers: Evidence for early 160 socialization about drugs occurring in the home. thld Meant. t1. 1520-1527. O'Gorman, P.A. (1971). el - o e t cus- -control a d e ce t WWW prchleme. Unpublished doctoral dissertation, Fordham University, New York. Ojala, C. (1988). osoc d t tera o s e e n a en d 1d en n lco o ic am e ° stud ea manifiestetions cfi risk. Unpublished doctoral dissertation, Pacific Graduate School of Psychology, Menlo Park, CA. Oliver, J.M., May, M.J., & Handal, P.J. (1988). Factor structure of the Family Environment Scale: Factors derived from subscales. Jcnrnel cf Clinical Esychology, AD, 723-727. Olson, D.H., Sprenkle, D.H., & Russell, C.S. (1979). Circumplex model of marital and family systems: I. Cohesion and adaptability dimensions, family types, and clinical applications. femily grocess, lD, 3-28. 4 Patterson, G.R. (1982). Coercive family procese. Eugene, OR: Castalia. Patterson, G.R., & Chamberlain, P. (1988). Treatment process: A problem at three levels. In L.C. Wynne (Ed.), Ihe state of rhe ert in ami era re ea ' C trov es nd ec at 8 (pp. 189-223). New York: Family Process Press. Peele, S. (1986). The implications and limitations of genetic models of alcoholism and other addictions. Jcnrnel_clernDie§_cn_Alcchcl, AZ, 63-73. Pennick, E.C., Powell, B.J., Liskow, B.I., Jackson, J.0., & Nickle, E.J. (1988). The stability of coexisting psychiatric syndromes in alcoholic men after one year. Journal cf Dtudiee on Alcohcl, Ag, 395-405. Pittman, F.S.,III. (1985). Children of the rich. Family Erocese, DD, 461-472. Regier, D.A., Farmer, M.E., Rae, D.S., Locke, B.Z., Keith, S.J., Judd, L.L., & Goodwin, F.K. (1990). Comorbidity of mental disorders with alcohol and other drug use. Jonrnal ci rhe American heeicel Association, ZDD, 2511-2518. Reider, E.E. (1987). coho volv a v e ce’ t o emcng alcchclic ene nonelccholic ramilies. Unpublished masters thesis, Michigan State University, East Lansing. Reider. E.E. (1991). WWW ramily ccnflict, and child behavior problems in young alcoholic 161 femiliee. Unpublished dissertation, Michigan State University, East Lansing. Reider, E.E., Zucker, R.A., Maguin, E.T., Noll, R.B., & Fitzgerald, H.E. (1989, August). Alcchcl involvement end viclence tcwerge chileren emcng_high_riek_femilie§. Paper presented at the meeting of the American Psychological Association, New Orleans. Reider, E.E., Zucker, R.A., Noll, R.B., Maguin, E.T., & Fitzgerald, H.E. (1988. August). AltehnLinxellemtntJndianilllitltntLimuh rick eample. Paper presented at the meeting of the American Psychological Association, Atlanta. Richards, T.M. (1979). Working with children of an alcoholic mother. Re e r Wo , D, 22-25. Ricks, D., & Berry, J.C. (1970). Family and symptom patterns that precede schizophrenia. In M. Roff & D.F. Ricks (Eds.), Lire sto esea at o . Minneapolis: University of Minnesota Press. Rivers, P.C. (1989). Substance abuse and psychopathology: The special population of the dual diagnosis patient. In G.W. Lawson & A.W. Lawson (Eds . ). MW popularions. Rockville, MD: Aspen. Roberts, C.C., Block, J.M., & Block, J. (1984). Continuity and change in parents' child-rearing practices. Dhllg_Deyelcpmenr, DD, 586-597. Robins, L.N., Bates, W.M., & O'Neal, P. (1962). Adult drinking patterns of former problem children. In D.J. Pittman & C.R. Snyder (Eds.), Dociety, culcnre, anc drinking petterng (PP. 395-412). New York: John Wiley & Sons. Robins, L.N., Helzer, J.H., Croughan, J., & Ratcliff, K.S. (1981). The National Institute of Mental Health Diagnostic Interview Schedule: Its history, characteristics, and validity. Archiye§_c£_generel Esychietry, DD, 381-389. Robins, L., Orvaschel, B., Anthony, J., Blazer, D., Burnam, M.A., & Burke, J. (1985). The Diagnostic Interview Schedule. In W.W. Eaton & L.C. Kessler (Eds.), idemio ’c e1 e hod i s c iat ° Ihe NIMH epidemiologic catchment area program. New York: Academic Press. Russell, M. (1990). Prevalence of alcoholism among children of alcoholics. In M. Windle & J.S. Searles (Eds.), Chileren of elcohclice; Crirical perepecrivee. New York: Guilford. Russell, M., Henderson, C., & Blume, S.B. (1985). Children cf altoholits1_A_retitw_2£_the_literatnrt. New York: Children of Alcoholics Foundation. 162 Sameroff, A., & Chandler, M.J. (1975). Reproductive risk and the continuum of caretaking casualty. In F.D. Horowitz (Ed.), Deview WW (vol. 4. pp- 197-244). Chicaso= University of Chicago Press. Sameroff, A., Seifer, R., Barocas, R., Zax, M., & Greenspan, S. (1987). Intelligence quotient scores of 4 year old children: Socioenvironmental risk factors. Pediatrics, 1g, 343-350. Sameroff, A., Seifer, R., Zax, M., & Barocas, R. (1987). Early indicators of developmental risk: Rochester Longitudinal Study. Dchizophrenie Dnllecin, lD, 383-394. Schiller, J. (1978). a eme t a d e 0 un o ' s a t e o ou 1d e . Unpublished doctoral dissertation, University of California, Berkeley. Schneider-Rosen, K., Braunwald, K., Carlson, V., & Cicchetti, D. (1985). Current perspectives in attachment theory: Illustration from the study of maltreated infants. Mono ra th oc e Research in Dhilc Development, DD (Serial No. 209, pp. 194-210). Schuckit, M.A., Goodwin, D., & Winokur, G. (1972). A study of alcoholism in half siblings. ou o , lgD, 1132-1136. Sclare, A.B. (1970). The female alcoholic. Dritieh Jonrnel cf Aedicticn, DD, 99-107. Seagull, E.A., Seagull, A.A., Tirosh, E., Adam, D., Jaffe, M., Sagi, A., Scharf, M., Zaruches, L., & Sternberg, K. (1988, April). ChilD e r n act es abus ve and non-abus ve sr e ents. Paper presented at the National Symposium on Child Victimization, Anaheim, CA. Searles, J.S., & Windle, M. (1990). Introduction and overview: Salient issues in the children of alcoholics literature. In M. Windle & J.S. Searles (Eds.), ldre alcoholic3° C t ca perspectives. New York: Guilford. Sears, R.R., Maccoby, E.E., & Levin, H. (1957). Patterns of child rearing. Evanston, IL: Row, Peterson, & Company. Seilhamer, R.A. (1987). Parterns of consumption and parent-child reletionships in iemilies of alcoholics. Unpublished doctoral dissertation, University of Pittsburgh. Seilhamer, R.A., & Jacob, T. (1990). Family factors and adjustment of children of alcoholics. In M. Windle & J.S. Searles (Eds.), Dhileren er alcchclice; Critical perspectives (pp. 168-186). New York: Guilford. 163 Seilhamer, R.A., Jacob, T., & Dunn, N.J. (1990). The impact cf elcchol 0 en - am es elcchclice. Manuscript submitted for publication. Selzer, M.L. (1975). A self-administered Short Michigan Alcoholism Screening Test (SMAST). Jcnrnal 2f Stngies on Alcohol, DD, 117- 126. Shedler, J., & Block, J. (1990). Adolescent drug use and psychological health: A longitudinal inquiry. Americen_£eychclcgier, AD, 612- 630. Sloboda, S.B. (1974). The children of alcoholics: A neglected problem. BMW. 1:. 605-606. Steinglass, P., Bennett, L.A., Wolin, S.J., & Reiss, D. (1987). The alccholic femily. New York: Basic. Stevens, 6., & Featherman, D.L. (1981). A revised socioeconomic index of occupational status. Docial Science Research, lD, 364-395. Stoneman, 2., Brody, G.R., & Burke, M. (1989). Marital quality, depression, and inconsistent parenting: Relationship with observed mother-child conflict. Americen_Jcnrnel_c£ W. 22. 105-117- Straus, M.A., Gelles, R.J., & Steinmetz, S.K. (1980). DehinD_clceeD WM- New York: Anchor PreSS/ Doubleday. Susman, E.J., Trickett, P.K., Iannotti, R.J., Hollenbeck, D.E., & Zahn- Waxler, C. (1985). Child-rearing patterns in depressed, abusive, and normal mothers. Ame a Jo a 0 c at , DD, 237- 251. Swiecicki, A. (1969). Adult adjustment of children from alcoholic and nonalcoholic families: A ten-year follow-up study. Erchl, Alhzmu, War§,, l1, 1-7. Abstract in Dnarterly Journal of Drngies on Alcohol, 1970, Dl, 1041. Tarter, R.E., Hegedus, A.M., Goldstein, C., Shelly, C., & Alterman, A.I. (1984). Adolescent sons of alcoholics: Neuropsychological and personality characteristics. A o ' C ic d e 'me tal Research, D, 216-222. Trickett, P.K., & Susman, E.J. (1988). Parental perceptions of child- rearing practices in physically abusive and nonabusive families. Developmentel Eeycholcgy, 2A, 270-276. 0.5. Bureau of the Census. (1990). Dretiecicel ehstracr of rhe D,S,; lggD (110th ed.). Washington, DC: Author 164 Vaillant, G.E., 6 Milofsky, E.S. (1982). The etiology of alcoholism: A prospective viewpoint. Americen_£eychclcgicr, Dl, 494-503. Vaughn, B.E., Block, J.H., 6 Block, J. (1988). Parental agreement on child rearing during early childhood and the psychological characteristics of adolescents. Dhild Development, Dg, 1020-1033. Watts, T.D., 6 Wright, R. (1983). Introduction. In T.D. Watts 6 R. Wright (Eds.), Dlech elcchcliem; Inware e compreheneive unnererenging. Springfield, IL: Charles C. Thomas. Wechsler, H., 6 Thum, D. (1973). Alcohol and drug use among teenagers: A questionnaire study. In M. E. Chafetz (Ed.), Ercceeginge_cfi_rhe W. Washington. DC: DHEW. Weil, C.M. (1987). assles soc a u or tom co nvolvement' The r teraction n oun alcoholic nd nonalco o c femiliee. Unpublished doctoral dissertation, Michigan State University, East Lansing. Welner, 2., Welner, A., McCrary, M. D., 6 Leonard, M. (1977). Psychopathology in children of inpatients with depression: A controlled study. Jonrnel of Neryous and Mental Diseeee, lDA, 408- 413. Werner, E.E. (1986). Resilient offspring of alcoholics: A longitudinal study from birth to age 18. W. 9.1. 34- 40. West, M.0., 6 Prinz, R.J. (1987). Parental alcoholism and childhood psychopathology- W. 19.2. 204-218. Wexler, M. (1991). Comments on Shedler and Block [Letter]. Americen Eeychologicr, DD, 165. Whisman, M.A., Strosahl, K., Fruzzetti, A.E., Schmaling, K.B., Jacobson, N.S., 6 Miller, D.M. (1989). A structured interview version of the Hamilton Rating Scale for Depression: Reliability and validity. Esychclogicel Aeeessmenr, 1, 238-241. Williams, C.N. (1987). Child care practices in alcoholic families: Findings from a neighborhood detoxification program. Alcchcl WM. 74-77. 94. Williams, C.N., 6 Klerman, L.V. (1984). Female alcohol abuse: Its effects on the family. In S.C. Wilsnack 6 L.J. Beckman (Eds.), o w e ' t ce e uences nd inreryenrlcn. New York: Guilford Press. Wilson, C., 6 Orford, J. (1978). Children of alcoholics: Report of a preliminary study and comments on the literature. Jgnrnel_2f W21. 3.2. 121-142. 165 Wolf, R. (1966). The measurement of environments. In A. Anastasi (Ed.), MW (pp. 491-503). Washington. DC: American Council on Education. Zigler, E.F., Lamb, M.E., 6 Child, I.L. (1982). Dccializarion ene ve o t (2nd ed.). New York: Oxford University Press. Zucker, R.A. (1976). Parental influences on the drinking patterns of their children. In M. Greenblatt 6 M.A. Schuckit (Eds.), W. New York: Grune & Stratton. Zucker, R.A. (1979). Developmental aspects of drinking through the young adult years. In H.T. Blane 6 M.E. Chafetz (Eds.), Icix.ch_,___a_lcchcl_L enc social poligx (PP. 91-146). New York: Plenum. Zucker, R.A. (1987). The four alcoholisms: A developmental account of the etiologic process. In P.C. Rivers (Ed.), co 0 nd addictive behavior; hebraske Symposinm cn hotivetion, l986 (pp. 27-83). Lincoln, NE: University of Nebraska Press. Zucker, R.A. (1989). Is risk for alcoholism predictable?: A probabilistic approach to a developmental problem. Drugs and fittlttl. 1. 69-94- Zucker, R.A. (1991). Scaling the developmental momentum of alcoholic process via the Lifetime Alcohol Problems Scores. Alcchcl_eng Alcoholism, Duppl, l, 505-510. Zucker, R.A. (in press). Alcohol involvement over the life span: A developmental perspective on theory and course. In L.S. Gaines 6 P.H. Brooks (Eds.), Alcohcl srudiee; A lifecpen perepective. New York: Springer. Zucker, R.A., 6 Barron, F.H. (1973). Parental behaviors associated with problem drinking and antisocial behavior among adolescent males. In M.E. Chafetz (Ed.), fieeeerch cn alcoholiem; i, Dlinicel prcblems ane epeciel populations (pp. 276-296). Washington, DC: NIAAA. Zucker, R.A., 6 Davies, W.H. (1989). e Rev se a tit -F e enc - Veriability Index; Detionele eng formnlae. Unpublished paper, Department of Psychology, Michigan State University, East Lansing. Zucker, R.A., 6 Devoe, 0.1. (1975). Life history characteristics associated with problem drinking and antisocial behavior in adolescent girls: A comparison with male findings. In R.D. Wirt, G. Winokur, 6 M. Roff (Eds.), Life hiecory reeearch in peychcperhplcgy_lyclr_el. Minneapolis: University of Minnesota Press. 166 Zucker, R.A., 6 Fillmore, K.M. (1968). motivational factore and problem 12W. Paper presented at the 28th International Congress on Alcohol and Alcoholism, Washington, DC. Zucker, R.A., 6 Fitzgerald, H.E. (in press). Early developmental factors and risk for alcohol problems. Alcohol Health and geeearch Wcrlg. Zucker, R.A., Fitzgerald, H.E., 6 Noll, R.B. (1990). Drinhing and drng h1§£2£1i_3221§2§_221£129 (version 4). Unpublished instrument, Department of Psychology, Michigan State University, East Lansing. Zucker, R.A., 6 Gomberg, E.S.L. (l986). Etiology of alcoholism reconsidered: The case for a biopsychosocial process. e can 2mm. .41. 783-793. Zucker, R.A., 6 Noll, R.B. (1980). ssessme t o antiso be v 0 ° Develo me t an ns ume t. Unpublished paper, Department of Psychology, Michigan State University, East Lansing. Zucker, R.A., 6 Noll, R.B. (1982). Precursors and developmental influences on drinking and alcoholism: Etiology from a longitudinal perspective. In Alcchol and health Monograph No, l; c l o s t a e ted ob s. Rockville, MD: NIAAA. Zucker, R-A., Noll, R., Draznin, T., Baxter, J., Weil, C., Theado, D., Greenberg, C., Charlot, C., 6 Reider, E. (1984). Ihe ecolcgy ci a o o 'c e ° 0 e t al mewo k f0 he ch an tate Dnivereity Lengicudinal Stney. Paper presented at the NCA National Alcoholism Forum, Detroit, MI. Zucker, R.A., Noll, R.B., 6 Fitzgerald, H.E. (l986). Drenr_prcpceel; Dish end ccping in chileren cf elcchclics. Unpublished manuscript, Michigan State University, East Lansing, MI. Zucker, R.A., Weil, C.M., Baxter, J.A., 6 Noll, R.B. (1984). Dififerencee n e e al nd ind v dual 5 c 0 at 010 on i s t s r the evelo ent o loo 0 . Paper presented at the Society for Life History Research in Psychopathology meetings, Baltimore, MD. "I7'11llflllfillll‘llfllls