v. w m w. _ w m w . _.. . (H $43913 if I Ill/IWill/l!fill/H/II/Ilml fii ' ll Tim/717: This is to certify that the dissertation entitled Causal Models of Alcoholic Family Environment Predicting Children's Behavior Problems presented by HaeeYoung Yang has been accepted towards fulfillment of the requirements for . . P h 1 Ph D degreein syc o ogy Majorp f or DMe June 25, 1992 MSU i: an Affirmauve Action/Equal Opportunity Institution 0- 12771 LIBRARY University Mlchlgan State PLACE IN REI’URN BOX to remove thie checkout from your record. TO AVOID FINES return on or before date due. DATE DUE DATE DUE DATE DUE E":. 3 I i 1 ans fir H W3 » N 22 flab! ‘49” 33 i e ‘ 7 N2 1“} a ‘t‘ 1%: - x\ P MSU Ie An Affirmative Action/Equal Opportunlty Institution tum FIND-M CA'. in CAUSAL MODELS OF ALCOHOLIC FAMILY ENVIRONMENT PREDICTING CHILDREN'S BEHAVIOR PROBLEMS BY Hae-Young Yang A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Psychology 1992 The p among Vari families is fiVe, Par the a1COho such as 0t paranting' ThUs. it s attribu risks Vere Present St functiOnin related to the leVel study was fathers, . variableS W” \a ABSTRACT CAUSAL MODELS OF ALCOHOLIC FAMILY ENVIRONMENT PREDICTING CHILDREN'S BEHAVIOR PROBLEMS BY Hae-Young Yang The present study examined the causal relationships among variables related to functioning of young alcoholic families with a biological son between the ages of three and five. Parental alcoholism is not the only risk present in the alcoholic family. There are other co-occurring risks, such as other parental psychopathology, inadequate parenting, family conflict, low SES, and social isolation. Thus, it seemed unlikely that psychopathology in COAs could be attributed solely to parental alcoholism so long as other risks were present in varying degrees and combinations. The present study assumed that the level of psychosocial functioning of alcoholics and their spouses was negatively related to the degree of disrupted family functioning and the level of their offsprings' adaptive functioning. This study was to explain the dynamic processes of how alcoholic fathers', their wives' psychosocial variables and child variables interacted to lead to children's adaptive behaviorI relation: problems support, variables results c regressic depressic problems. support, . including and their maternal g theY were regard to ihereased in eXacer p3th anal their Spc in Child; decrEasE< fathersn dept-€831 behavior. Path models were constructed to explain causal relations among parental variables such as lifetime alcohol problems, antisocial behavior, ego functioning, social support, stress, current depression, and children's variables such as temperament and behavior problems. The results of path analyses and hierarchical multiple regression analyses indicated that parents' current depression predicted significantly their children's behavior problems. Parents' ego functioning, maternal social support, and stress mediated parental psychopathology including lifetime alcohol problems and antisocial behavior and their current depression. Parents' ego resiliency and maternal social support decreased parents' depression, thus they were likely to be involved in protective processes with regard to children's adaptive functioning. Since stress increased parents' depression, it was likely to be involved in exacerbating or vulnerable processes. The results of path analysis showed the effects of alcoholic fathers on their spouses, who were supposed to mediate impact of stress in children's social environment: Alcoholic fathers' support decreased mothers' current depression whereas alcoholic fathers' current depression increased mothers' current depression. Also, it was found that children' temperament contributed significantly to children's behavior problems. I we for his c1 support, I wan members, Pl Hunter for My ap undergradu Collection Longitudin; this prgjefl Final: Stay here, broken and during Vaca ACKNOWLEDGEMENTS I want to express my gratitude to Hiram E. Fitzgerald for his contribution to this work. Without his patient support, I could not dream of finishing whole this process. I want to express my appreciation to my other committee members, Robert A. Zucker, Ellen A. Strommen, and John E. Hunter for their helpful comments and suggestions. My appreciation to all staffs, graduate students, and undergraduate students who have dedicated themselves to the collection and management of the data for the MSU Longitudinal Study. A special thanks to the families in this project that I have never met and may not meet. Finally, thanks to my family who endured my delayed stay here, old friends in New York City who took care of a broken and exhausted graduate student without any complaints during vacations, and two friends in East Lansing who offered a cookie and their sympathetic ears to me for last two years. ii LIST OF T LIST OF F ‘Ir‘~"l‘RODUC'I‘.J RE’IEW OF The C Funct State: FOFma$ mihon ..... I SUbjec PrOced HeaSurl O Mr Ms M Be Me Me Me TABLE OF CONTENTS PAGE LIST OF TABLESCOCOOOOOOOOOOOOOOOO0.0.0....0.0.0.0.... Vi LIST OF FIGURES........................... ....... .... viii *INTRODUCTION To THE STUDY............ ....... ......... 1 *REVIEW OF THE LITERATURE............................. 6 The Children at RiSROOOOOOOOOOOOOOOOOOOOOO00.... 6 The Prevalence of Alcoholism among COAs.... 6 Assessment of COAs' Psychopathology........ 9 COAs' Psychopathology...................... 10 Functioning of the Alcoholic Family............. 16 The Alcoholic Family Environment........... 16 Relationships among Parental Psychopathology............................ 19 Interactions between Male Alcoholics and Their spouse............................... 22 Statement of the Problem........................ 26 Formal Predictions.............................. 31 METHODOCOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO ..... .0. ....... 35 subjects 0 O O O I C O O O I O O C O O O O O O O O O O O O O O O O O O O O O O O O O O O 3 5 Procedures 0 O O O O O I O C O O O O I O O O O O O O O O O O O O O O O O O O O O O O O 3 9 Measures 0 I C O O I O O C O O O O O O O C O O O O O O O C O O O O O O O O O O O O O O O 40 Measuring Demographic Information.......... 40 Measuring Parent's Personality............. 40 Measuring Parent's Current Depression...... 42 Measuring Parent's Antisocial Behavior..... 43 Measuring Parent's Lifetime Drinking Behavior................................... 44 Measuring Mother's Social Support.......... 44 Measuring Parent's Stressful Events........ 46 Measuring Parenting........................ 46 Measuring Child's Adaptive Functioning..... 47 Child Behavior Problems............... 47 Child Ego Resiliency and Ego Control.. 49 Child Intellectual Ability............ 49 Child's Temperament................... 50 iii RESULTS . Hig Hie - i 20 21 *o'vowjrun 22 23 24 PAGE RESULTSCOCCOOCCCOOOIOOOOOOOOOOOOOOOOOOOOOOOOOOOO.I... 51 High Risk Families............................. 51 Parental Characteristics.................. 52 Children's Characteristics................ 55 Hierarchical Multiple Regression Analysis...... 63 Dependent Variable........................ 63 Independent Variables..................... 67 Predicting COAs' Behavior Problems........ 75 Relationships between Parents' Psychopathological Problems and Children's Behavior Problems.............. 82 PathAnaIYSiS.OOOOOOOCOOOOOIOOOOOOO0.0.0.000... 85 Model 1: Parents in Alcoholic Families.... 86 Model 2: Alcoholic Fathers................ 90 Model 3: Alcoholics' Wives. ..... .......... 92 DISCUSSION.OOOOOOOOOOCOOOOIOOOOOOOOO ....... 0.0.0.... 97 Parents in Alcoholic Families.................. 98 Children in Alcoholic Families................. 102 Predictors of COAs' Psychopathology............ 106 Path Model of Alcoholic Families............... 115 Path Models of Alcoholics and Their Wives...... 121 APPENDIX A Informed Consent Form.................... 128 APPENDIX B Statistical Tables 18 Mean Scores for Parents' Characteristics of COAs in the "Clinical Range" and COAs in the "Nonclinical Range"................ 130 19 Mean Scores for Behavior Problems, Temperament, Personality, and General Developmental Status for COAs in the "Clinical Range" and COAs in the "Nonclinical Range"............... ...... . 131 20 Correlations between Parental Variables and Positive Affective and Authoritarian Parenting.................................. 133 21 Correlations between Demographic Variables, Parental Variables and Children's Behavior Problems in Alcoholic Families............. 134 22 Path Analysis of Parental Variables Predicting COAs' Total Behavior Problems... 135 23 Path Analysis of Alcoholic Fathers' Variables Predicting COAs' Total Behavior Problems................. ......... 137 24 Path Analysis of Alcoholics' Wives' Variables Predicting COAs' Total Behavior Problems.......................... 140 iv LIST OF PAGE LIST OF REFERENCES.......................... ........ . 143 TAB 10 ll 12 13 14 TABLE 1 2 10 11 12 13 14 LIST OF TABLES Demographic Characteristics of the Study Sample.................................... Basic Risk Study Design: Michigan State University Longitudinal Study............. Mean Scores and Standard Deviations for Parental Variables for Alcoholic Families and Comparison Families..... ........... .. Mean Scores and Standard Deviations for Personality and General Developmental Status for COAs and Children in Comparison Families................. ...... Mean Scores and Standard Deviations for Children's Behavior Problems and Temperament in Alcoholic and Comparison Families...... Correlations between RYDS and Children's Personality............................... Correlations among RYDS, Children's Personality and Behavior Problems..... ..... Correlations between Temperament and Behavior Problems for Children in Alcoholic Families....... .................. Correlations among CBCL Scales for Father and Mother in Alcoholic Families.......... Interparental Agreement of Mother and Father CBCL Data.......................... Order of Entry of Independent Variables and Measurements for Hierarchical Multiple Regression Analysis....................... Observed Correlations among Parental Variables and Children's Behavior Problems in Alcoholic Families............ Multiple Regression Analysis I: Parental Variables, Children's Temperament, and Children's Behavior Problems.......... Multiple Regression Analysis II: Parental and Demographic Variables, Children's Temperament, and Children's Behavior Problems ....... ........ .......... vi PAGE 36 38 53 56 57 59 61 62 65 65 68 71 78 79 Ta' 15 16 17 Table PAGE 15 16 17 Multiple Regression Analysis III: Combined Parental Variables, Children's Temperament, and Children's Behavior Problems............................ ...... 81 Multiple Regression Analysis IV: Parental Psychopathology and Children's Behavior Problems......................... 83 Multiple Regression Analysis V: Combined Parental Psychopathology and Children's Behavior Problems...... ........ 84 vii FIG‘ LIST OF FIGURES FIGURE PAGE 1 Theoretical Model of Parental Variables Predicting the Level of COAs' Adaptive Functioning............................... 34 2 Path Model of Parental Variables Predicting Mothers' Ratings of Children's Behavior Problems.................................. 88 3 Path Model of Alcoholic Fathers' Variables Predicting Mothers' Ratings of Children's Behavior Problems......................... 91 4 Path Model of Alcoholics' Wives' Variables Predicting Mothers' Ratings of Children's Behavior Problems......................... 93 viii Chi highest problems only for disorder Many COAs childhood depressio substance Incre on Childre Children i alc0holios behaviors. same emoti The litera Children's “‘0“? Cons Clear Vhicl psychopath< variabilit) eXternal i2 j INTRODUCTION TO THE STUDY Children of alcoholics (COAs) comprise one of the highest risk groups for developing psychopathological problems in their lifetime. COAs may be at a great risk not only for the development of alcoholism, but of psychiatric disorder in general (Russell, Henderson, & Blume, 1984). Many COAs show such psychiatric and physical disorders in childhood and adolescence as conduct disorder, delinquency, depression, hyperactivity, attention deficit disorder, and substance abuse (West & Prinz, 1987; Murray, 1989). Increasing research on the effects of alcoholic parents on children provides a complex and conflicting picture of children in alcoholic families. First, not all COAs become alcoholics in adulthood and not all COAs develop maladaptive behaviors. Even COAs in the same family do not suffer the same emotional, social, and academic problems (Brown, 1986). The literature on the consequences of parental alcoholism on children's behavior indicates large individual differences among COAs between and within families. Second, it is not clear which behaviors early in the child's life predict psychopathology in later life, since there is considerable variability in COAs' psychopathology, ranging from externalizing behavior to internalizing behavior and, 1 2 furthermore, many maladapted COAs did not end up with psychiatric disorders in adulthood. One way to unravel such complexity is to perform a longitudinal study to assess developmental processes within a dynamic context in which interactions between the child and his social environment occur over time (c.f., Sameroff, 1980). The present study is a part of the MSU Longitudinal study (Zucker & Fitzgerald, 1991), that is concerned with the etiology of alcohol abuse, alcoholism, and antisocial behavior. Compared with other longitudinal studies (Zucker & Noll, 1982; Zucker, 1990) the MSU Longitudinal study recruited fairly young alcoholic families with a biological son between the ages of three and six, whose alcoholic fathers were not from clinical populations. The present study uses data from the first wave of the longitudinal study. Without considering interrelations among variables describing the functioning of alcoholic families it is impossible to infer how and why certain characteristics of either the child or the environment are causally related to developmental outcome. Rutter (1987) suggests that it is useless to label variables simply as risk or protective types, since it is the process or mechanism, not the variables, that determines the function. Thus the present study focuses on the risk or protective process/mechanism associated with the child's adaptation. There are three limitations to circumscribe the present study. First, the tC fi en is Se as en as me 01": 1c Cc a] 3 social environment in which the child is reared is limited to his family due to the child's young age (i.g., three to five years). Since most of the target children have not yet entered school, their experience of the social environment is most likely to be connected to their parents' behaviors. Second, there are two options to assess the child's social environment; to measure the child's family environment by asking subjects (e.g., parents or children) to describe their family environment, or to measure the variables associated with the quality of the child's social environment. The present study chose the second method, assessing parents"characteristics reflecting the active mechanism of the child's social environment. Third, since only male children are involved, the study can only address factors related to the etiology of alcohol abuse and alcoholism in men. It has been generally known that alcoholism is more prevalent among men than among women (Helzer & Pryzbeck, 1988). Biological vulnerability of boys (Rutter, 1984) could be one reason for higher rates alcoholism among sons of alcoholics (Cloninger, Bohman, & Sigvandsson, 1981). Zucker (1988, 1989) and his colleagues (Zucker & Fitzgerald, 1991; Zucker & Gomberg, 1986) reviewed several longitudinal studies of alcoholism and identified some common factors relating to the continuity of outcome of alcoholic process. Among the common factors, those that describe the characteristics of the social environment of pre hei PSY per ass as den P5)? of J ac PSY exa rel inc 4 prealcoholic males include inadequate and poor parenting, heightened marital conflict, and such parental psychopathological problems as alcoholism, antisocial personality, or sexual deviancy. All of these variables associated with parental psychosocial functioning may serve as risks that potentiate each other's impact on the development of children in alcoholic families. Alcoholics' and their spouses' psychopathology and poor psychosocial functioning likely facilitate the development of adverse interactions among family members (Seilhamer & Jacob, 1990). Children are not only exposed to such psychopathology but may be drawn into it. The present study examined the causal relationships among parental variables related to the child's social environment and behavior, including alcoholism, antisocial behavior, personality, depression, stress, and social support. Within this context, it is possible to assess what variable in alcoholic families is involved in either risk or protective process. Specially, the present study was interested in roles that alcoholics and their spouses play in family functioning. Do alcoholic fathers have direct effects on their children or are the effects of alcoholic fathers mediated by their wives (Jacob & Leonard, 1986)? The present study proposes an theoretical model of the COAs' social environment from which three path models are derived. The first model examines parental effects on child adaptive functioning. The second path model examines paterna examine the sec explana‘ and chii 5 paternal effects on the child, while the third path model examines maternal effects. Compared with the first model, the second and the third models test more detailed explanations of the link between parental characteristics and child characteristics. P; conclu: higher do chi; (Russel differs enviror “Coho: REVIEW OF THE LITERATURE The Children at Risk Previous reviews of alcoholic families yield two major conclusions. First, children of alcoholics (COAs) have higher rates of alcoholism and alcohol-related problems than do children whose parents are not alcoholic (nonCOAs) (Russell, 1990). Second, there are extensive individual differences in COAs' responses to risks within their social environment, particularly to parental alcoholism. T e ev ence f A co 0 ism amon COAs Many Americans grow up in alcoholic families; 22,000,000 persons aged 18 or more and 6,600,00 under age 18, or one out of every 8 Americans (Russell, Henderson, & Blume, 1984). A similar number of people diagnosed as alcohol abuse/dependence was found in the Epidemiologic Catchment Area (ECA) study that surveyed psychiatric disorder in the general population (Helzer & Pryzbeck, 1988). About 13 percent of the total sample from five cities across the United States had a diagnosis of alcohol abuse/dependence. It seems plausible to believe that there are between twenty-five and twenty-eight million children of alcoholics in American Society (Ackerman, 1983), though 6 est bet tra the alc san may it 7 estimates of the prevalence of alcoholism among COAs varied between studies. One of the problems in estimating of familial transmission of alcoholism is that the representativeness of the sample studied and the diagnostic criteria for alcoholism are different across studies (Russell, 1990; West & Prinz, 1987). Many studies use clinical samples of alcoholics in treatment so they likely inflate the risk for alcoholism in COAs (Russell, 1990). Russell argues that samples having large numbers of serious alcoholic fathers may overestimate alcoholism in their sons and underestimate it in their daughters, (see also Cloninger, Bohman, and Sigvardssons, 1981). Also, different definitions for alcoholism across studies make it impossible to compare results of different studies (West & Prinz, 1987; Russell, 1990). Generally the diagnostic criteria for alcoholism include drinking behaviors and drinking-related problems: physical dependence; loss of control, morning drinking, or another pattern of pathological use; and adverse social and occupational difficulties from drinking (West & Prinz, 1987). Secondary factors seemed to influence estimates of familial transmission of alcoholism. For example, the Epidemiologic Catchment Area study reported high comorbidity rates between alcoholism and other psychiatric diagnoses. Thus, additional psychiatric disorders may affect the incidence of alcoholism (Russell, Henderson, & Blume, 1984). se fa a] $6 an fa pa C71 "'1 8 In addition, there are other possible secondary factors if severe alcoholism is associated with other risks likely to increase alcoholism among COAs, such as low socioeconomic status, family conflict, divorce, violence, and social alienation (El-Guebaly & Offord 1977; Russell, 1990; Leonard & Jacob, 1988). Most studies of COAs have linked these secondary factors (c.f., Zucker's nondrinking-specific factors; Zucker & Noll, 1982; Zucker, 1986) to parental alcoholism within alcoholic families. Nevertheless, as several investigators point out, the nature of causal links among parental alcoholism and other co-occurring secondary factors, such as low SES, family disorganization, and parents' other psychiatric disorders is not clear (E1- Guebaly and Offord, 1977; West & Prinz, 1987). Although there is considerable variation in the prevalence of alcoholism among COAs, rates of alcoholism are higher among COAs than in the general population (Cotton, 1979; Russell, 1990). Certainly COAs are the highest risk group for the development of alcoholism. However, alcoholism is not the only possible outcome for children raised in alcoholic families. COAs may be at a great risk for the development of psychiatric disorders as well. Associated forms of psychopathology include conduct disorder, hyperactivity, low school performance, delinquency, substance abuse, and depression (El-Guebaly & Offord, 1977; Russell, Henderson, & Blume, 1984; West & Prinz, 1987; Murray, 1989; Windle, 1990). :1:- U“) U) I var extc del. dep: 199: not pos: inf: Prii has 9 sse s e t o OAs' Ps cho atholo There are several problems in evaluating the incidence of psychopathology in COAs. First, there is considerable variability in COAs' psychopathology, ranging from externalizing behavior (e.g., conduct disorder and delinquent behavior) to internalizing behavior (e.g., depression and somatic symptoms; West & Prinz, 1987; Windle, 1990). The interrelationship among such symptoms in COAs is not clear. Since most studies only assess a limited set of possible symptoms and depend on either one or two sources of information (e.g., parents or COAs' self-report) (West & Prinz, 1987), a general pattern of psychopathology in COAs has failed to materialize. Related to the question of etiology, it is not clear that age or developmental level is especially associated with certain types of psychopathology (West & Prinz, 1987). It is likely that as the child's developmental level and risk factors associated with the child's social environment change (Zucker & Noll, 1984; Phares & Compas, 1992), the child's psychopathological symptomatology will also change. Thus, unless one adapts a longitudinal prospective strategy, it will be impossible to determine how precursor events of early childhood are connected to alcoholism and psychopathology in adulthood. From the child' perspective, parental alcoholism is not the only risk factor present in an alcoholic family. There are other co-occurring risk factors, such as parental 5] fa v we Ea Zu 10 psychiatric illness, family disorganization, marital discord, poor parenting, low SES, and social isolation (Seilhamer & Jacob, 1990). Thus, it seems unlikely that psychopathology in COAs can be attributed solely to parental alcoholism so long as other risk factors are present in varying degrees and combinations (West & Prinz, 1987). Phares and Compas (1992) reviewed studies of developmental psychopathology and pointed out that the separate contributions of fathers and mothers to child maladjustment were ignored in analysis of data. Many studies aggregated data of fathers and of mothers, making it impossible to identify the independent contributions of fathers and mothers, or the interaction of paternal and maternal variables. Furthermore, some studies collected only one parent's data, usually mother's, thus ignoring the other parent's effect on child psychopathology. Phares and Compas argue that only by separating the data pertaining to fathers and mothers it is possible to discern under what circumstances paternal and maternal variables contribute to child psychopathology. It is likely that studies of alcoholic parents ignore the separate contributions of fathers and mothers to children maladjustment within alcoholic family environment. CO 5' o o o Zucker (1989) and his colleagues (Zucker, Noll, Drazni, Baxter, Weil, Theado, Greenberg, Charlot & Reider, 1984; Zucker & Gomberg, 1986; Zucker, 1988) reviewed several “‘11 longit factor< proces. of var‘ to pare inadequ alcohol Child v; level, ; Guebaly Goodwin, 1985; we; El-l literatu; hlPeract life, pa; hYSteria behavior Rydelius reCGived ”mg n RideliuS at Schoo He provn experiem VhEre the 11 longitudinal studies of alcoholism and identified common factors related to continuity of outcome of alcoholic process. These factors can be grouped into two categories of variables, those related to the child and those related to parents. Parent variables in prealcoholic homes include inadequate and poor parenting, heightened marital conflict, alcoholism, antisocial behaviors, and sexual deviancy. Child variables include antisocial behavior, higher activity level, poor achievement, and interpersonal deficits (E1- Guebaly & Offord, 1977; Knop, Teasdale, Schulsinger, & Goodwin, 1985; McCord, 1988; Tarter, Alterman, & Edwards, 1985; Werner, 1986). El-Guebaly and Offord (1977), in their review of the literature, concluded that there was a relationship between hyperactivity in childhood and psychiatric disorder later in life, particularly alcoholism, antisocial personality, and hysteria. In addition, hyperactivity and antisocial behavior were frequently co-occurring during childhood. Rydelius (1984) followed offspring of alcoholics who received treatment in the Swedish clinics. The fathers had a long history of alcoholism, often exceeding 10 years. Rydelius did not specify at what ages children's functioning at school was examined or how many children were examined. He provided only that half of school age children experienced considerable problems of adjustment at school where they were falling behind their classmates even though they had normal intelligence. The main reasons for their dif: conc sch: cons alcc adjL that out, were vulr adui c l 31C: 31C: 12 difficulty in school related to hyperactivity and lack of concentration. For instance, for the youngest group of schoolboys about 75 percent of the 7-9-year-old boys were considered as problem children at school. Girls from alcoholic home environments had similar problems of adjustment at school. In the follow-up study Rydelius found that the boys who reacted to their situations by "acting- out," and who were physically neglected, and whose fathers Iain—1 were especially seriously addicted to alcohol were more vulnerable to maladjustment problems and substance abuse in adulthood. From a large Danish birth cohort, Knop, Teasdale, Schulsinger, and Goodwin (1985) selected about 204 sons of alcoholics and matched controls for a longitudinal study of alcoholism. The alcoholic fathers were selected if they had been identified as alcoholic in either of two Sweden registers. At age 19 and 20 the premorbid assessment was conducted by social worker interview, psychopathological interview inquiring about adolescent COAs' drinking behaviors, and teachers completing a questionnaire on the child's intellectual, emotional and social functioning. The COAs experienced more difficulties in school than the control group; they attended more schools and more often repeated a grade than the control group. Fifty-one percent of the COAs had been referred to a school psychologist, as compared to 34 percent of the control group. Teachers evaluated the COAs as being more impulsive-restless and as an M t h I :IIM record 13 having poor verbal ability. Knop et al. have concluded that cognitive dysfunctioning may antecede alcoholism in COAs with a history of hyperactivity during childhood or with a family history of alcoholism, suggesting a possible genetic component of risks. McCord (1988) followed boys who were participating in a delinquency prevention program in order to identify developmental paths that likely lead to alcoholism. The criteria for screening alcoholic fathers were if their fathers experienced drinking-related problems in either work place or marital relation, if he received treatment for alcoholism, or if he had been convicted at least three times for public drunkenness. The COAs were selected on the basis of information about parental alcohol-related problems collected from either the primary caretaker or the COAs themselves. Among 203 families, 65 fathers (32%) were identified as alcoholics. Forty-seven percent of the COAs became alcoholics, compared to 25 percent of the nonCOAs. The COAs were more likely to have a record of juvenile delinquency and to become alcoholics or criminals in adulthood if their elementary school teachers rated them as both aggressive and shy. Unfortunately McCord did not provide other data concerning the boys' school performance. Another longitudinal study (Werner, 1986) showed that by age 10, almost a third of the COAs were in need of long-term mental health care. By age 18, 30 percent of the COAs had records of delinquency, and 41 percent of the COAs had 14 serious coping problems that caused difficulties at home, in school, at work, or in the community. Antisocial behavior, conduct disorder, or delinquency in childhood or adolescence are frequently mentioned as precursors of alcoholism in the literature of the COAs .(Zucker & Gomberg, 1986; Hesselbrock, Hesselbrock, Babor, Stabenau, Meyer, & Weidenman, 1984). Robins' (1978) classic follow-up study showed that alcoholism, drug abuse, violence, and job trouble were significantly predicted by the level of childhood antisocial behavior, such as arrests, truancy, alcohol, fighting, and dropout. Based on the fact that each childhood behavior significantly predicted a high level of antisocial behavior and each type of adult antisocial behavior was predicted by the number of childhood behaviors, Robin argued that the behaviors that lead to either childhood antisocial behavior or to adult antisocial personality operate in a similar pattern. Donovan and Jessor (1985; Donovan, Jessor, & Costa, 1988) demonstrated that problem drinking, illicit drug use, precocious sexual behavior, and delinquency are likely to form a syndrome composed of highly interrelated "problem behaviors." Thus, regardless of which problem behavior first appears in adolescent, the probability that other dysfunctional behaviors will develop increases dramatically. Brook, Whiteman, Gordon, and Cohens' (1986) study showed the substantial continuity in personality traits related to children's drug use. They studied patterns of than 59hr lung 15 continuity and change in personality attributes from childhood (5 to 10 years) to adolescence (13 to 18 years) in relation to the adolescents' drug use. The personality attributes in their study are grouped into four major dimensions; conventionality/ unconventionality (e.g., poor school behavior and delinquency), control of emotions (e.g., impulsivity), intrapsychic functioning (e.g., depressive mood and poor ego integration), and interpersonal relatedness (e.g., aggression). They found that childhood personality attributes were related to adolescent personality attributes, which in turn influenced adolescents' drug use. Evidence is accumulating to support the hypothesis that antisocial behavior, greater activity level, difficulty in achievement-related activity, and interpersonal deficits are core symptoms for development of alcoholism and psychopathology. It is likely that such variables during childhood tend to be highly correlated, particularly so when one considers the developmental history of COAs. For instance, childhood antisocial behavior associated with hyperactivity seems to be a good predictor of COAs' psychopathology in adulthood (c.f., Phares & Compas, 1992). The co-occurrence of antisocial behavior and hyperactivity in childhood is more predictive of high risk status for COAs ‘than.is the occurrence of only one of the two symptoms. Nevertheless, not much is known about the interrelationships among symptoms of psychopathology during childhood. A con; psyc rese test conc dise usef alco Fitz: famil and v and s 1990; “Sing al90h: Subjec Spouse (1984) reCoWe Soc 10d patien. alCQhO. 16 complete understanding of such interrelationship among psychopathological symptoms can only be attained by focusing research on prospective studies of family functioning and by testing models that reflect multivariate contextual conceptualizations of developmental process. Mechanistic disease based models of causality have not be especially useful in advancing our understanding of the etiology of alcoholism and its associated psychopathology (Zucker & Fitzgerald , 199 1) . Functioning of the Alcoholic Family The Alcoholic Family Environment Alcoholic families are frequently described as showing family disorganization, marital discord, adverse interaction and violence among family members, poverty, unemployment, and social isolation (West & Prinz, 1987; Seilhamer & Jacob, 1990; Leonard & Jacob, 1988). Two studies 1990; Russell, using the Family Environment Scale described similar alcoholic family environments despite differences in :ubjects between studies; one study of alcoholics and their pauses and the other study of adult COAs. Moos and Moos 1984) compared family environments among families of :covered, relapsed alcoholic patients, and ciodemographically matched controls two years after the tients completed treatment. Families of relapsed :oholics reported significantly less cohesion and ext rec act con‘ alcc retr had orie subje them dealt Baxte; in alc hostil study using - bGlues; ViolenC to'w'ard CoheSio and hig) ighEr 1 reported except t. Emmasis‘ 17 expressiveness than the other two groups, and families of recovered and relapsed alcoholics showed less emphasis on an active-recreational orientation than did families of controls. Clair and Genest (1987) asked adult offspring of alcoholic and comparison group fathers to report retrospectively on their family environments, and the COAs had lower scores on cohesion, organization, and intellectual orientation, and higher scores on conflict than did the comparison group. However, it should be noticed that the subjects were recruited from college community and most of them were female. Two unpublished papers from the MSU Longitudinal study dealt with the alcoholic family environment. Zucker, Weil, Baxter, and Noll (1984) found that both husbands and wives in alcoholic families described their relationship in more hostile terms than did comparison families. The second study (Reider, Zucker, Noll, Maguin, & Fitzgerald, 1988), using 70 alcoholic families, dealt with the relationship between perceptions of family environment and recent spousal violence. Alcoholics reporting higher levels of violence toward their wives during the past year had lower scores on organization, and active-recreational orientation cohesion, Alcoholics' wives reporting and high scores on conflict. higher levels of violence toward their husbands also reported similar perceptions of their family environment, except they reported a lower level of moral-religious concluded that violence for both emphasis. Reider et a1. alc neg alc del fat; adol alcc anti thrc Offo the faCtI addit that great The 1 0n cl fazi] Child Coot: but j of 0t antis 18 alcoholics and their spouses was associated with their more negative perceptions of the family environment. Low socioeconomic status (SES) is frequently found in alcoholic families (Russell, 1990) or families of juvenile delinquents (Rutter & Giller, 1983) . Although alcoholic fathers are common in the families of antisocial adolescents, the question is whether or not parental alcoholism has specifically contributed to the etiology of antisocial behavior or whether its main contribution is through its tendency to lower family SES (El-Guebaly & Offord, 1977). Again the same question is asked; what is the relationship between parental alcoholism and other risk factors in alcoholic families? It is unlikely that the effect of risk factors is additive, but rather that they potentiate one another so that the combined effects of risk factors together is greater than the sum of risk factors (Rutter, 1983, 1987) . The real question is how they combine to exert their effect on child development. For example, all children from families of low SES do not turn out to be antisocial children or antisocial adults (Robins, 1978) . Low SES may contribute to the level of childhood antisocial behavior, but it is more plausible that low SES moderates the effect of other risk factors in children's environment on child antisocial behavior. In the family environment of antisocial children other risk factors are consistently found, such as parents' cri par Gil eacl clee amor 19 criminality and abnormal personality, family discord, poor parenting, and weak parent-child relationship (Rutter & Giller, 1983) . These risk factors seem to correlate with each other although the causal links among them are not clear. The time has come to investigate the relationships among the risk and protective factors in alcoholic families in order to discern how risk and protective factors in the child's social environment influence the child's adjustment or maladjustment, thereby potentiating or buffering the child's risk for psychopathology. Relationships amohg Parental Psychopathology Generally studies of the COAs attempt to isolate the specific impact of parental alcoholism on children's psychopathology, but other "secondary factors" associated 'with alcoholic families likely make such attempts difficult at best. Only the presence of parental alcoholism does not account for variability in dysfunction of alcoholic family, which is partially attributed to other risk factors in family. Although alcoholic families are more dysfunctional than nonalcoholic families, there is a considerable variability within alcoholic families (Clair & Genest, 1987) . The variability in the functioning of alcoholic families may account for the observed variability in COAs' outcome (West & Prinz, 1987) . One of main reasons for the variability in the functioning of alcoholic families is that alcoholics are a heterogenous group; a substantial variability in alcoholism and co affect most c Render Ric , ( CA) across abuse/ group Pryzbe occur 5 P Rube. b ~‘v’p C39> 42 (90th percentile)" and COAs falling in the "nonclinical range (TBP raw score <= 42)." Applying this criterion to alcoholic families in the MSU Longitudinal Study, Reider (1991) found that about 30 percent of children fell in the "clinical range" on Total Behavior Problems and Externalizing Behavior. The present study confirms Reider's finding in that 30 percent of the children in alcoholic families were rated in the "clinical range." Parahrai Characteristics For all the parental variables alcoholics displayed more negative characteristics than did nonalcoholics (Table 3). However, wives of alcoholics displayed more negative characteristics than did wives of nonalcoholics on only two variables; Lifetime Alcohol Problems and lifetime Antisocial Behavior. These results are consistent with the general literature which finds, that alcoholics display more negative characteristics than nonalcoholics, whereas there are comparatively fewer differences between wives of alcoholics and wives of nonalcoholics. In addition, there were differences in demographic characteristics of alcoholic and comparison families; comparison families had higher SES [t(173)=3.68, p<.000], longer marriage [t(145)=3.11, p<.001], and more years of education than did parents in alcoholic families [for father, t(176)=4.35, p<.000; for mothers, t(179)=1.92, p<.05]. Mothers in alcoholic families ”a. nuts 0 «\u Table 3 53 Mean Scores and Standard Deviations for Parental Variables for Alcoholic Families and Comparison Families Alcoholic Comparison Families Families M §2 M §Q E Q: 2 Alcohol Problems Mothers 10.07 2.05 8.94 1.29 -3.26 190 .000 Fathers 10.63 1.84 7.25 1.67 -11.40 189 .000 Antisocial Behavior Mothers 12.69 7.55 7.44 4.77 -3.76 165 .000 Fathers 23.10 12.91 11.03 5.71 -5.31 167 .000 Depression Mothers 3.59 3.71 3.22 3.31 -.55 172 ns Fathers 3.24 3.34 1.31 1.26 -3.41 170 .000 Personality: Ego Resiliency Mothers .03 .37 .15 .37 1.80 170 ns Fathers -.11 .30 .23 .30 6.17 159 .000 Personality: Ego Undercontrol Mothers -.08 .20 -.12 .18 -1.32 170 ns Fathers .05 .20 -.12 .21 -4.42 159 .000 Stress Mothers 21.61 12.69 18.62 12.05 -1.27 157 ns Fathers 20.10 15.09 15.66 10.00 -1.64 155 .10 Social Support Mothers 15.81 2.92 16.45 1.99 1.16 153 ns Fathers 15.49 3.52 15.50 3.59 .02 151 ns 'IJ'E Ill} llc "r Ta 31: Sin Gui 54 were younger than were mothers in comparison families [t(162)=1.66, p<.10]. There was no group difference in the number of children per family. Looking more closely at the alcoholic families, the differences between alcoholic families of children in the "clinical range" and alcoholic families of children in the "nonclinical range" were especially interesting (Appendix B, Table 18). In contrast to the findings for wives in alcoholic and comparison families on all parental variables except Ego Undercontrol and Social Support, wives in alcoholic families of children in the "clinical range" displayed more negative characteristics than did wives in alcoholic families of children in the "nonclinical range" (Lifetime Alcohol Problems [t(134)=-3.52, p<.00], Antisocial Behavior [t(129)=-2.51, p<.01], Current Depression [t(129)=- 3.08, p<.00], Ego Resiliency [t(124)=2.52, p<.01], and Current Stress [t(115)=-2.28, p<.02]. Thus, wives in alcoholic families of children in the "clinical range" showed more negative characteristics than their counterparts in alcoholic families of children in the "nonclinical range." In contrast, there were no differences between alcoholic fathers of children in the "clinical range" and alcoholic fathers of children in the "nonclinical range" except Lifetime Alcohol Problem [t(134)=-2.13, p<.05] and Current Depression [t(127)=-3.55, p<.00]. 55 Chiigran's Characteristics As indicated in Table 4, children in comparison families had slightly better scores on the Revised Yale Developmental Schedules than did the children in alcoholic families and were rated as being more Ego Resilient than were COAs. However, there were no differences in parental reports of behavior problems (CBCL) between children in comparison families and those in alcoholic families (Table 5). Jacob and Leonard (1986) and Fitzgerald, Sullivan, Ham, Zucker, Bruckel, Schneider, and Noll (in press) reported the same results even though the two studies involved different age groups. Jacob et al. studied COAs between 10 and 18 years of age, whereas Fitzgerald et al. studied 3-year-old COAs. There were few differences in parental perceptions of their children's temperament between alcoholic and comparison families (Table 5). Parents in alcoholic families perceived their children as being more reactive than comparison parents perceived their children. Fathers in comparison families perceived their children as more attentive than did fathers in alcoholic families, and mothers in alcoholic families perceived their children as more adaptable than did comparison mothers. Compared to the findings reported for parents within alcoholic families, the differences between parents in comparison families and alcoholic families in their perceptions of their children are not impressive or striking. 56 Table 4 Mean Scores and Standard Deviations for Personality and General Developmental Status for COAs and Children in Comparison Families Alcoholic Comparison Families Families M §Q M §Q E Q: 2 Personality ER .10 .32 .24 .26 2.39 158 .01 EU .06 .19 .03 .19 -.93 158 ns RYDS Developmental Status GM 49.92 13.07 52.32 11.43 .97 157 ns FM 52.27 11.41 56.09 10.16 1.77 157 .05 AD 53.28 10.12 56.51 9.00 1.69 157 .05 L0 52.47 11.87 57.56 10.75 2.03 157 .05 PSD 53.75 10.41 55.86 9.54 1.92 157 .05 DQ 52.95 10.21 55.86 9.39 1.46 157 .05 Note. ER=Ego Resiliency. EU=Ego Undercontrol. GM=Gross FM=Fine Motor Development. LD=Language Development. DQ=Overall Developmental Motor Development. AD=Adaptability. PSD=Personal/Social Development. Quotient. 57 Table 5 Mean Scores and Standard Deviations for Children's Behavior Problems and Temperament in Alcoholic and Comparison Families Alcoholic Comparison Families Families M §Q M SD r df p CBCL Total Behavior Problems Mothers 57.30 9.63 55.60 8.35 -1.04 189 ns Fathers 56.56 10.45 54.62 10.68 -1.06 186 ns Externalizing Behavior Mothers 56.53 10.49 54.47 9.38 -1.16 189 ns Fathers 56.29 10.52 53.71 10.38 -1.40 186 ns Internalizing Behavior Mothers 55.22 8.85 54.47 8.22 -.73 189 ns Fathers 54.45 9.90 52.86 11.06 -.89 186 ns Aggression Mothers 61.28 8.17 59.84 7.68 -1.04 189 ns Fathers 61.32 8.42 59.81 7.88 -1.04 189 ns Temperament Rhythmicity Mothers 5.22 2.16 5.42 2.41 .45 160 ns Fathers 5.26 2.09 5.51 2.13 .61 160 ns 58 Table 5 (cont'd) M §Q M Reactivity Mothers 3.27 1.70 2.55 Fathers 3.41 1.42 2.90 Attention Span/Distractibility Mothers 5.19 2.94 5.87 Fathers 4.76 2.87 5.90 Adaptability/Approach-Withdrawal Mothers 4.12 1.70 3.23 Fathers 3.97 1.47 3.90 Activity Level Mothers 1.33 1.43 1.16 Fathers 1.69 1.43 1.54 lfl -2.62 -.22 -.59 -.49 160 160 160 160 160 160 160 160 .05 .05 ns .05 .01 ns ns ns Table 6 Correlations between RYDS and Children's Personality Children in Alcoholic Families Children in Comparison Families Gross Motor Fine Motor Adaptability Language Personal/Social Overall Development Quotient ER .36** .45** .55** .48** .50** .50** EU -0 19* -.22* -.24** -.24* -.22* -.25** ER .30 .52** .42* .43* .48** .44* EU -.38* -. 51** -.51** -.40* -.43* -.47* * p<.05, ** p<.01. Note. ER=Ego Resiliency. Two-tailed. EU=Ego Undercontrol. 60 As indicated in Table 6, Ego Resiliency was positively correlated with developmental status in all areas measured by the RYDS with the exception of comparison children's gross motor development. For both COAs and comparison children, Ego Undercontrol was inversely correlated with RYDS subscales and full scale DQ. However, the correlations for comparison children were substantially higher than those for COAs. Table 7 summarizes correlations for RYDS, children's Ego Resiliency and Undercontrol, and children's behavior problems. Table 8 shows correlations between parents' ratings of child temperament and behavior problems. Although there were significant correlations, most were of low order accounting for relatively little variance. Parents' ratings of children's behavior problems were generally not significantly correlated with children's personality and RYDS developmental status (Table 7). However, maternal ratings of children's Externalizing Behavior in comparison families were associated positively with children's Ego Undercontrol and negatively with children's developmental status on all subscales but Gross Motor Development. Interparent agreement about behavior problems and correlations between children's temperament and behavior problems (Table 8) will be discussed later. Results of t-tests for COAs falling in the "clinical range" and COAs falling in the "nonclinical range" showed few differences except children's behavior problems (CBCL): Children falling in the "clinical range" were rated as being 61 Table 7 Correlations among RYDS, Children's Personality and Behavior Problems Alcoholic Families Mother Father TBP EXT INT AGG TBP EXT INT AGG ER -..06 -008 -000 -.12 -000 .00 -002 .00 EU .10 .15 .03 .20* .15 .18* .15 .15 GM .04 .04 .08 -.01 .14 .12 .16 .10 FM .02 .00 .10 -.08 .09 .04 .11 .07 AD .01 -.01 .08 -.12 .09 .05 .11 .06 LAN -.02 -.03 .02 -.10 .09 .06 .11 .08 PS .02 -.01 .09 -.11 .05 .01 .08 .04 DQ -.03 -.05 .04 -.13 .09 .05 .11 .05 Comparison Families Mother Father TBP EXT INT AGG TBP EXT INT AGG ER -.24 -.34 -.17 -.18 .22 .05 .30 -.12 EU -.21 .42* -.08 .44* -.07 -.02 -.24 -.02 GM -.10 -.19 .04 -.01 .03 -.07 .10 -.02 FM -.25 -.40* -.02 -.22 .06 -.11 .21 -.16 AD -.25 -.35* -.05 -.23 -.01 -.13 .09 -.13 LAN -.28 -.36* -.13 -.19 -.01 -.10 .03 -.06 PS -.22 -.36* -.03 -.14 .09 -.07 .21 -.03 DQ -.31 -.42* -.10 -.25 -.04 -.18 .07 -.19 * p<.05, ** P<.01. Two-tailed. Note. ER=Ego Resiliency. EU=Ego Undercontrol. GM=Gross Motor Development. AD=Adaptability. PS=Personal/Social Development. FM=Fine Motor Development. LAN=Language Development. DQ=Overall Development Quotient. TBP=Children's Total Behavior Problems. EXT=Children's Externalizing Behavior. AGG=Children's Aggression. Internalizing Behavior. INT=Children's 62 Table 8 Correlations between Temperament and Behavior Problems for Children in Alcoholic Families TBP EXT INT AGG FTBP Mothers Activity .16 .15 .08 .19* .08 Attention -.21* -.23** -.18* -.35** .01 ; Adaptability .06 .12 -.03 .16 -.08 E. Rhythmicity -.24* -.19* -.20* -.19* -.05 Reactivity .21* .26** .21* .25** .13 [ Fathers Activity .24** .21* .19* .16 .25** Attention -.11 -.09 -.16 -.09 -.25** Adaptability .15 .14 .10 .21* -.14 Rhythmicity —.03 .01 -.04 -.oo -.16 Reactivity .29** .30** .19* .32** .36** * p<.05, ** p<.01. Two-tailed. Note. TBP=Children's Total Behavior Problems (Mothers' Rating). EXT=Children's Externalizing Behavior (Mothers' Rating). INT=Children's Internalizing Behavior (Mothers' Rating). AGG=Children's Aggression (Mothers' Rating). FTBP=Children's Total Behavior Problems (Fathers' Rating). 0‘ 2 Idl‘ 63 less Ego Resilient [t(119)=1.71, p<.05] and were perceived by their parents as more reactive [for mothers, t(122)=- 2.36, p<.02; for fathers, t(122)=-3.49, p<.001] and active [for mothers, t(122)=-1.75, p<.05; for fathers, t(122)=- 2.23, p<.02]. Mothers of children falling in the "clinical range" perceived their children as less rhythmic [t(122)=2.77, p<.01] and attentive [t(122)=3.37, p<.001] than did mothers of children falling in the "normal range" (Appendix B, Table 19). Hierarchical Maltiple Regression Anaiysis In this and the next section, statistical analyses only involved the alcoholic families. The purpose of these analyses was to identify predictors of the child's adaptive functioning. Several multiple regression analyses were performed to test the predicted relationships as described previously. T e e e V ' bl The Achenbach Child Behavior Checklist (CBCL) was selected as the primary measure for examining the level of children's adaptive functioning. The CBCL is a standardized instrument which, based on wide clinical and nonclinical samples, provides normative data for various age group on scales measuring children's behavior problems (Achenbach & Edelbrock, 1983). The CBCL is designed to obtain parents' idescriptions of their children's problem behaviors related 64 to psychopathological symptoms and social competence in a standardized format. The present study used scores from scales except social competence for analysis since some items for social competence dealt with children's school performance and data on school performance were not available. The CBCL provides scores for two broad-band scales (Internalizing and Externalizing) developed from eight subscales and a Total Behavior Problem score. The present study used the Total Behavior Problem (TBP) score as the primary dependent variable for multiple regression analysis and path analysis. There is a stronger association of children's clinical status with the TBP score than with other subscales in most age groups (Achenbach et al., 1983). According to Achenbach et al. the TBP score includes the problems of all the subscales, plus problems that were not included on the specific scales. Thus, the TBP score is more likely used to show clinical indications of problem behavior than either broad-band scale or any of the narrow band subscales. In the current study, there were high correlations between the TBP, the two broad-band scales, and the Aggression subscale for both mothers and fathers (Table 9). Although the strength of these intercorrelations seem to run counter to Achenbach et al.’ claim that most clinically referred children do not show equally deviant behaviors in all areas, they do show that TBP seems more reasonable to use as the primary dependent variable, with 65 Table 9 Correlations among CBCL Scales for Father and Mother in Alcoholic Families TBP EXT INT AGG Total Behavior -- .94**a .93** .84** Problems (TBP) Externalizing .93**b -- .81** .87** (EXT) Internalizing .88** .74** -- .72** (INT) Aggression .75** .84** .54** -- (AGG) ** p<.01, *** p<.001. Two-tailed. a: Fathers' correlations are listed above the diagonal. b: Mothers' correlations are listed below the diagonal. Table 10 Interparental Agreement of Mother and Father CBCL Data Alcoholic Comparison Standardization Families Families Sample Reports3 (N=131) (N=35) (N=33) Total Behavior .34** .58** .75** Problems Externalizing .39** .57** .74** Internalizing .27** .59** .75** Aggression .38** .54** .72** ** p<.01, *** p<.001. Two-tailed. a: the Achenbach Standardization Group (Boys 4-5 years; Clinically referred children; Achenbach & Edelbrock, 1983). 66 other scale scores serving as secondary dependent variables for multiple regression analysis. Reider (4991) using 90 alcoholic families in the MSU Longitudinalrstudy reported low order correlations between mothers' and fathers' scale scores, indicating that parental perceptions of their children were discrepant. The present pr- study confirms Reider's (1990) low interparent agreement i about their children's behavior (Table 10). Compared with 1 parents in alcoholic families, parents in comparison | families showed slightly higher interparent agreement. .However, in neither group did correlations between mothers' and-fathers' scale scores approach the level of agreement reported by Achenbach and Edelbrock (1983) for clinically referred children. It is likely that the high interparent agreement for clinically referred children reflects a higher . level of severity in children, with more consistency in their behavior across contexts. One possible explanation for low interparent agreement in the current study is that alcoholic fathers may have less interaction with their young children, thus they do not have much chance of observing their children's behavior, compared with their wives greater involvement in caretaking activities. Reider compared mother and father CBCL scores separately against other measures of the children's behavior. She found that there was some congruence between mothers' report and those of independent observers, even though the level of congruence was quite low. Schaughency and not sug of) per: behe age neas that 67 and Lahey (1985) compared teachers' ratings, fathers' ratings, and mothers' ratings of school children referred for treatment. They found that teachers' ratings of children's behaviors were significantly correlated with mothers' ratings, but not correlated with fathers' ratings, suggesting that the factors involved in paternal perceptions of child deviance might be different from maternal [5 perception. However, interparent agreement about children' behaviors in Schaughency et al.' study was higher than that | of the present study. There were some differences between Schaughency et al.' study and the current study; different age and risk status of subject children and different measurements for children's behavior problems. Considering that the children in the current study are of preschoolage, it is reasonable to suppose that mothers are more involved in their care and socialization. Wives of alcoholics do report that they perform more household tasks than their husbands (Moos & Moos, 1984). For these reasons, the present study used maternal ratings of children's behavior problems for analysis. Wag Hierarchical multiple regression analyses were performed on different dependent variables including mothers' report of TBP, Externalizing Behavior, and Aggressive Behavior, and fathers' report of TBP. Orders of entry of independent variables were determined by the theoretical model of the present study. Table 11 summarizes Table 11 Order of Entry of Independent Variables and Measurements for Hierarchical Multiple Regression Analysis (1) (2) (3) (4) (5) s. (6) (7) (3) Variables Mothers' Depression Fathers' Depression Mothers' Social Support Mothers' Stress & Fathers' Stress Mothers' Ego Resiliency Fathers' Ego Resiliency Fathers' Lifetime Alcohol Problems & Antisocial Behavior Mothers' Lifetime Alcohol Problems & Antisocial Behavior Child's Temperament: Reactivity, Activity, Attention Span/Distrac- tibility Measurements The Beck Depression Inventory (BDI) II II II The Norbeck Social Support a The Hassles and Uplifts .- Scales i The California Adult Q-Sort (CAQ) The Drinking & Drug History Questionnaire; The Antisocial Behavior Checklist II The DOTS-Child 69 the forced orders of independent variables and their measurements for the analysis. Independent variables that were serving as the initial variables in the causal relations were forced into the equation first; parent's Antisocial Behavior and Lifetime Alcohol Problems. However, as the two variables were highly correlated, they were entered simultaneously. If it was not possible to determine the order of two variables, they were entered simultaneously, for example were mothers' and fathers' stress, Ego Resiliency, and children's temperament variables. Even though parents' Antisocial Behavior and Lifetime Alcohol Problems were serving as the initial variables in the causal model, it was expected that such as parents' psychopathology, including depression, caused impairment in their parenting, which was directly related to children's behavior problems (Rutter, 1990). The present study planned to enter the parenting variable first, and parents' Current Depression second that was causally related to parenting. However, it did not find significant the effect of parenting variable on children's behavior problems, thus it dropped the variable from analysis, which was discussed below. The secondary analysis was conducted in reversed orders of entry of independent variables; mothers' Depression, fathers' Depression, mothers' Social Support, mothers' and fathers' Stress, mothers' and fathers' Ego Resiliency, fathers' Lifetime Alcohol Problems and Antisocial Behavior, and 70 mothers' Lifetime Alcohol Problems and Antisocial Behavior. Actually, the final results of analyses using different orders of independent variables showed the same result. Before discussing the results of hierarchical multiple regression analysis, there are several things that should be noted. First, correlations between the independent variables ranged from .00 to .60 (less 1 % of the variables were above .50), suggesting that there was no serious multicollinearity problem (Table 12). However, since almost all the independent variables were correlated, it may be difficult to isolate the specific effect of the independent variables in multiple regression analysis, despite the lack of multicollinearity problem. This issue will be discussed later. Second, the maternal social support included only spouses' and other family members' emotional and instrumental aids, since they were significantly associated with the variables in Table 11. Mothers' social support was strongly correlated with spousal support (r=.61, p<.01, two- tailed, for emotional aid; r=.60, p<.01, two-tailed, for instrumental aid), and with other family members' support (r=.47, p<.01, two-tailed, for emotional aid; r=.31, p<.01, two-tailed, for instrumental aid). However, alcoholics' social support did not show as strong association with spousal support as the maternal social support did (r=.44, p<.01, two-tailed, for spouses' emotional aid; r=.39, p<.01, two-tailed, for spouses' instrumental aid; r=.31, p<.01, Tahl Ohse Chill thPE PHASE KER hsnu mm 3151?; FLAPE PTASE YER PSTRE FTBDI HTBP E”; m hacc FTBP FEXT FIFT FAGG REACT 71 Table 12 Observed Correlations among Parental Variables and Children's Behavior Problems in Alcoholic Families MLAPSR MTASB MER MSTRESS MTBDI MSUPPORT MLAPSR -- MTASB .54** -- MER -.09 -.23* -- MSTRESS .18* .27** -.21* -- MTBDI .16 .33** -.30** .44** -- MSUPPORT -.14 -.21* .22* -.21* -.39** -- FLAPSR .15 .25** -.15 .20* .27** -.17 FTASB .11 .29** -.20* .14 .16 -.10 FER -.13 -.24* .38** -.11 -.11 .32** FSTRESS .00 .02 -.08 .22* .16 -.04 FTBDI .26** .23** -.28** .27** .45** -.29** MTBP .26** .29** -.21* .28** .38** -.23* MEXT .29** .32** -.21* .28** .36** -.21* MINT .17 .17 -.21* .26** .34** -.20* MAGG .24** .30** -.16 .26** .27** -.11 FTBP .18* .21* -.19* .16 .12 -.20* FEXT .19* .21* -.19* .16 .11 -.19* FINT .15 .18* -.15 .11 .13 -.20* FAGG .13 .25* -.15 .20* .07 -.13 REACT .13 .14 -.21* .13 .18* -.08 MATTEN -.03 -.14 .02 .07 .04 -.01 ACTIV .06 .17 -.22* .22* .21* -.11 * p<.05, ** p<.01. Two-tailed. 72 Table 12 (cont'd) FLAPSR FTASB FER FSTRESS FTBDI MLAPSR MTASB MER MSTRESS MTBDI MSUPPORT FLAPSR -- FTASB .57** -- FER -.30** -.24** -- FSTRESS .27** .18 -.20* -- FTBDI .34** .28** -.14 .60** -- MTBP .22** .21* -.10 .20* .41** MEXT .22** .22* -.14 .19* .36** MINT .14 .20* -.03 .14 .37** MAGG .20* .14 -.19* .18 .28** FTBP .29** .27** -.14 .29** .36** FEXT .30** .27** -.13 .31** .35** FINT .25** .31** -.09 .23** .34** FAGG .31** .25** -.10 .32** .38** REACT .21* .20* -.08 .13 .22** MATTEN .01 .08 .07 .08 .05 ACTIV .16 .07 .08 .18 .26** * p<.05, ** p<.01. Two-tailed. 73 Table 12 (cont'd) MTBP MEXT MINT MAGG FTBP FEXT MTBP -- MEXT .93** -- MINT .88** .71** -- MAGG .75** .84** .54** -- FTBP .34** .34** .26** .32** -- FEXT .33** .39** .23** .39** .94** -- FINT .30** .27** .27** .25** .93** .81** FAGG .29** .35** .18* .38** .84** .87** REACT .30** .34** .25** .35** .29** .37** MATTEN -.20* -.23** -.18* -.35** .01 -.03 ACTIV .25** .23* .17 .22* .21* .20* FINT FAGG REACT MATTEN ACTIV FINT -- FAGG .72** -- REACT .21* .29** -- MATTEN -.02 .02 -.35** -- ACTIV .18 .24** .11 .02 -- Note. MLAPSR=Mothers' Lifetime Alcohol Problems. MTASB=Mothers' Antisocial Behavior. MER=Mothers' Ego Resiliency. MSTRESS=Mothers' Stress. MTBDI=Mothers' Current Depression. MSUPPORT=Mothers' Social Support. FLASPR=Fathers' Lifetime Alcohol Problems. FTASB=Fathers' Antisocial Behavior. FER=Fathers' Ego Resiliency. FSTRESS=Fathers' Stress. FTBDI=Fathers' Current Depression. MBPT=Mothers' Ratings of COAs' Total Behavior Problems. MEXT=Mothers' Ratings of COAs' Externalizing Behavior. MINT=Mothers' Ratings of COAs' Internalizing Behavior. MAGG=Mothers' Ratings of COAs' Aggression. FBPT=Fathers' Ratings of COAs' Total Behavior Problems. FEXT=Fathers' Ratings of COAs' Externalizing Behavior. FINT=Fathers' Ratings of COAs' Internalizing Behavior. FAGG=Fathers' Ratings of COAs' Aggression. REACT=Parental Ratings of Child Reactivity. MATTEN=Mothers' Ratings of Child Attention Span/Distractibility. ACTIV=Parental Ratings of Child Activity Level. 74 two-tailed, for other family members' emotional aid; r=.38, p<.01, two-tailed, for other family members' instrumental aid). Alcoholic fathers' social support that was weakly correlated with child behavior problems (r=-.19, p<.05, two- tailed) was not used for analysis of the current study. Third, based on correlations observed between temperament and behavior problems scores three aspects of EP- children's temperament were selected as the independent variables; Reactivity, Activity Level, and Attention Span/Distractibility (Table 8). For Reactivity and Activity Level, parent's scores were combined into a single score, whereas only mothers' perception of Attention Span/Distractibility was used. Reactivity consisted of items that were derived from the categories of activity level, intensity, and threshold (Lerner, Palermo, Spiro, & Nesselroade, 1982). It was not significantly correlated with parents' perception of Activity Level (r=.11, ns), but was negatively correlated with maternal perception of Attention Span/Distractibility (r=-.35, p<.01, two-tailed). All these variables were correlated with the dependent variables. Furthermore, parents' perception of Reactivity and mothers' perception of Attention Span/Distractibility showed some congruence with reports of an observer. Parents' perceptions of Reactivity were positively correlated with children' Ego Undercontrol (r=.39, p<.01, two-tailed), and mothers' perception of Attention Span/ Distractibility was correlated positively with children's 75 Ego Resiliency (r=.24, p<.01, two-tailed) and negatively with children's Ego Undercontrol (r=-.26, p<.01, two- tailed). Fourth, parenting variables were dropped from multiple regression analyses and path analysis, because they were not significantly associated with children's behavior problems (see Appendix B, Table 20). Only alcoholic fathers' ‘positive affective parenting was slightly correlated with children's behavior problems, but it did not account for \nariance in multiple regression analysis if it was entered. IT! addition, only mothers' positive affective parenting in comparison families was significantly associated with maternal ratings of children's behavior problems. Finally, Ego Undercontrol was also dropped from analysis because it was not significantly associated with other variables. Ergdaficting COAs' Behavior Problems Table 13 summarizes the significant results of hierarchical multiple regression analysis on different dependent variables including mothers' ratings of Total Behavior Problems, Externalizing Behavior, and Aggressive Behavior, and fathers' ratings of Total Behavior Problems. Table 13 displays the final results of regression analysis after all the independent variables in Table 11 were forced to enter, Table 14 shows the effects of demographic Variables, whereas Table 15 displays the final results of regression analysis after entering some combined variables deVeloped from the independent variables in Table 11. 76 When all the independent variables were forced into the equation, no specific variable predicted mothers' ratings of TBP and Externalizing Behavior (Table 13). Although the specific effects of the independent variables were not found, all the independent variables combined did predict mothers' ratings of the TBP, Externalizing Behavior, and Aggression. In addition, all variables combined predicted fathers' ratings of TBP [Adj R2=.29, F=3.50(14,72), p<.000]. Mothers' ratings of Attention Span/Distractibility predicted their children's Aggressive Behavior (Table 13) , whereas fathers' Current Depression predicted their children's TBP (B=-34, T=2.32, sig.T=.02) . To examine possible effects of demographic variables on alcoholic families on children's behavior problems (see Appendix B, Table 21), Family SES, parents' Age, parents's Education, and Years of Marriage were added to the independent variables in Table 11. To control the effects of demographic variables and original independent variables, one multiple regression was conducted when demographic variables were entered first, followed by the forced entry of the independent variables, and another multiple regression was conducted when the independent variables were forced into the equation, followed by free entry of demographic variables. The results showed that controlling for either set of independent variables did not decrease the effect of the other set of variables. Demographic variables contributed significantly to 77 children's TBP, Externalizing Behavior, and Aggression (Table 14). However, Years of Marriage was responsible for an increase in variance for children's TBP and Externalizing Behavior. When adding only Years of Marriage into the equation, it accounted for a significant portion of the —.47, Adjusted R2=.34 variance of children's behaviors (R2 However, entering all for TBP and Externalizing Behavior). demographic variables except Years of Marriage into the equation did not account for additional variance (R2=.42, Adjusted R2=.25 for TBP) , beyond that accounted for by all independent variables excluding demographic variables (Table For children's Aggression, Years of Marriage, Family 13)- Education contributed SES, Parents' Age, and Parents' significantly to children's behavior. Parents' perception of Reactivity predicted children's TBP (B=-.28, T=2.25, Sig. Mothers' Education and maternal ratings of T=. 03) 0 Attention Span predicted children's Aggression: children having mothers with high education (B=.30, T=2.19, Sig. T=. O3 ) and/or perceived as less attentive by their mothers (B=-- 32, T=-2.80, Sig. T=.01) were seen as more aggressive. In addition, all the independent variables combined predicted fathers' ratings of TBP [Adjusted R2=.26, F=2 - 31 (20, 56) , p<.000], but the total amount of accounted variance decreased from .29 to .26. Given the findings of multiple regression analysis, it was eVident that there was a multicollinearity problem despite the fact that correlations between the independent 78 Table 13 Multiple Regression Analysis I: Parental Variables, Children's Temperament, and Children's Behavior Problems Adj. Standard Sig. R R2 R2 Error F df p Beta T T TBP .61 .38 .26 8.45 3.16 14,73 .007 EXT .64 .40 .29 9.12 3.54 14,73 .000 AGG .65 .42 .30 7.57 3.72 14,73 .000 FTBP .64 .41 .29 9.49 3.50 14,72 .000 AGG Attention Span/Distractibility -.25 -2.40 .02 FTBP Fathers' Depression .34 2.32 .02 Note. TBP=Total Behavior Problems (Mothers' Rating). EXT=Externalizing Behavior (Mothers' Rating). AGG=Aggression (Mothers' Rating). FTBP=Tota1 Behavior Problems (Fathers' Rating). Table 14 79 Multiple Regression Analysis II: Parental and Demographic Variables, Children's Temperament, and Children's Behavior Problems Adj. Standard Sig. R R2 Error F df p Beta T T TBP .71 .51 .33 7.10 2.98 20,57 .001 EXT .71 .51 .33 8.17 2.92 20,57 .001 AGG .71 .51 .33 7.37 2.93 20,57 .001 FTBP .67 .45 .26 10.1 2.31 20,56 .007 TBP Activity Level .28 2.25 .03 AGG Mothers' Education .30 2.19 .03 Attention Span/Distractibility -.32 -2.80 .01 FTBP Mothers' Social Support -.24 -2.05 .05 Note. TBP=Total Behavior Problems (Mothers' Rating). EXT=Externalizing Behavior (Mothers' Rating). AGG=Aggression (Mothers' Rating). Problems (Fathers' Rating). FTBP=Total Behavior 80 variables did not indicate a serious problem. To reduce the effects of multicollinearity, two variables which were theoretically related and highly correlated were combined into one variable (See correlation matrix in Table 12). Parent's Lifetime Alcohol Problem and Antisocial Activity were combined into a variable labeled as the Lifetime Difficulty. Parent's Current Stress and Depression were combined into a variable labeled Current Difficulty. Another series of hierarchical multiple regression analysis was conducted on the dependent variables (Table 15). The new combined independent variables and old variables were forced into the equation in the same way that the variables were entered for the previous regression analysis. Even though some of the correlated variables were combined, the same pattern of results was found, so combining variables did nothing to reduce multicollinearity. No single independent variable predicted mothers' ratings of TBP. The independent variables as a whole significantly predicted maternal rating of children's TBP. Mothers' ratings of Attention Span/Distractibility predicted children's Aggressive behavior and fathers' Current Difficulty (combined of Current Stress and Depression) predicted children's TBP (of fathers' rating; B=.39, T=4.08, sig T=.OO). Furthermore, parents' ratings of Reactivity predicted children' Externalizing Behavior. Once more all the equations turned out to be significant (p<.000). 81 Table 15 Multiple Regression Analysis III: Combined Parental Variables, Children's Temperament, and Children's Behavior Problems Adj. Standard Sig. R R2 R2 Error F df p Beta T T TBP .57 .33 .25 8.31 4.34 10,89 .000 EXT .60 .35 .28 8.98 4.98 10,89 .000 AGG .59 .35 .28 7.57 4.79 10,89 .000 FTBP .59 .35 .28 9.50 4.80 10,88 .000 EXT Reactivity .19 1.98 .05 AGG Attention Span/Distractibility -.28 -2.92 .00 FTBP Current Difficulty .39 4.08 .00 Note. Combined Variables were used for analysis. Parent's Lifetime Difficulty=Lifetime Alcohol Problems and Antisocial Behavior. Parent's Current Difficulty=Current Stress and Depression. TBP=Total Behavior Problems (Mothers' Rating). EXT=Externalizing Behavior (Mothers' Rating). AGG=Aggression (Mothers' Rating). FTBP=Total Behavior Problems (Fathers' Rating). 82 Relationships between Parents' Psychopathological problems and Children's Behavior Problems Having failed to find the effects of independent variables that specifically contributed to children's behavior problems, the present study examined the relationship between parents' psychopathology and child' behavior problems. This time, parents' Lifetime Alcohol Problem, Antisocial Behavior, and Current Depression were selected as independent variables, and the same dependent variables were used for a series of hierarchical multiple regression analysis (Table 16). Orders of entry of the independent variables were fathers' Lifetime Alcohol Problem and Antisocial Activity, mothers' Lifetime Alcohol Problems and Antisocial Activity, fathers' Current Depression, and mothers' Current Depression. As indicated in Table 16, mothers' Current Depression and fathers' Current Depression predicted mothers' ratings of children's TBP. Mothers' Current Depression predicted their ratings of Externalizing Behavior, while fathers' Current Depression predicted their ratings of TBP. All the equations were significant (p<.00). Next, the combined variables were used as independent variables; parent's Lifetime Difficulty (combined from Lifetime Alcohol Problem and Antisocial Activity) and Current Difficulty (combined from Current Stress and Depression) (Table 17). Mothers' Lifetime Difficulty predicted significantly their ratings of TBP, Externalizing 83 Table 16 Multiple Regression Analysis IV: Parental Psychopathology and Children's Behavior Problems Adj. Standard Sig. R R2 R2 Error F (if p Beta T T TBP .53 .25 .21 8.816.48 6,115 .000 EXT .48 .23 .19 9.88 5.76 6,115 .000 AGG .39 .15 .11.8.16 3.49 6,115 .003 FTBP .43 .19 .14 10.2 4.34 6,114 .000 TBP Mothers' Depression .23 2.44 .02 Fathers' Depression .24 2.53 .01 AGG Mothers' Depression .20 2.16 .03 FTBP Fathers' Depression .28 2.87 .00 E Note. TBP=Total Behavior Problems (Mothers' Rating). Ex'ill'=Externalizing Behavior (Mothers' Rating). AGG=Aggression (Mothers' Rating). FTBP=Tota1 Behavior PrC>k>lems (Fathers' Rating). 84 Table 17 Multiple Regression Analysis V: Combined Parental psychopathology and Children's Behavior Problems Adj. Standard Sig. R R2 R2 Error F df p Beta T T TBP .40 .16 .14 9.12 6.28 4,130 .000 EXT .42 .18 .15 9.91 7.07 4,130 .000 AGG .38 .15 .12 7.92 5.65 4,130 .000 FTBP .44 .19 .17 9.72 7.80 4,129 .000 TBP ‘ Mothers ' Lifetime Difficulty . 19 2 . 05 . 04 EXT Mothers ' Lifetime Difficulty . 22 2 . 40 . 01 AGG Mothers ' Lifetime Difficulty . 28 2 . 58 . 01 FTBP Fathers ' Current Difficulty . 28 3 . 27 . 00 Fathers ' Lifetime Difficulty . 21 2 . 48 . 01 n N9te . Combined Variables were used for analysis. Parent's Lifetime Difficulty=Lifetime Alcohol Problems and Antisocial Behavior. Parent's Current Difficulty=Current Stress and Depression. TBP=Tota1 Behavior Problems (Mothers' Rating). EXT=Externalizing Behavior (Mothers' Rating). AGG=Aggression (Mothers' Rating). FTBP=Total Behavior Prol:>].ems (Fathers' Rating). 85 Behavior, and Aggression. Fathers' Lifetime Difficulty (35.21, T=2.48, sig T=.01) and Current Difficulty (B=.28, T=3 - 27, sig T=.00) predicted significantly their ratings of TBP «- Path Analysis Path analysis is used to study the causal relations among a set of variables by using partial and multiple correlation (Hunter & Gerbing, 1982) . Hunter and Gerbing suggest that one of the major contributions of path analysis is that it decomposes relations among variables into numerical components, so separate effects and directions of relationships among variables may be observed. The present Study used Hunter and Gerbings' PACKAGE program to conduct Path analysis in prediction of COAs' Total Behavior PrOblems. Based on the theoretical model (Figure 1) and reStilts of observed correlations among variables three path mode ls with control of independent variables were used to Predict COAs' TBP. Path analysis used the same variables that were selected as independent variables for multiple regression analysis (Table 11) . Before discussing the results of path analysis there are several points to note. First, the reliability of each variable was computed based on the subjects in the current study, since many reliabilities associated with the original measurements were calculated 86 based on middle-class samples. Reliabilities of the variables are presented in Appendix B, Table 22-24. Second, the observed correlations among variables were corrected for attenuation to remove measurement errors. And for each path model, path coefficients (beta weights), the level of significance, R2, adjusted R2, and overall chi square number and degree of freedom for test of overall fit of the model were presented. MW There are two reasons why a measurement model strictly derived from a theoretical model (Figure 1) can not be tested. First, there are too many variables that are correlated with each other, so it is almost impossible to consider all causal relationships among variables in one mociel. The second reason, related to the first, is that the Present study presumes to test a recursive path model that does not have causal cycles among variables because it is based on a cross-sectional design. There were significant cOrrelations between alcoholics' and spouses' Lifetime Alcohol Problem and Antisocial Activity, and it is not clear the causal direction among these variables (Table 12) . For instance, it is not clear whether the alcoholics' Alcohol PrOblem causes their spouses' Alcohol Problem or Antisocial Activity, or both. Or it may be possible that the causal direction goes from the alcoholic's wife's problem to the alcoholic's problem. Thus, the present study developed three path models predicting maternal ratings of children's 87 behavior problems; the first one for parents' combined variables, the second for alcoholics' variables, and the third for alcoholics' spouses' variables. Based on the findings of multiple regression analysis and the theoretical model, the first path model tested the causal relationships among parental variables such as psychopathology (Lifetime Alcohol Problem and Antisocial Behavior), Ego Resiliency, Stress, and Depression by combining alcoholics' and their spouses' variables as parental variable (Appendix B, Table 22) . Parents' perception of child's Reactivity was selected for path analysis since this variable showed the strongest correlation with child's TBP (r=.30, p<.01, two-tailed), and after child Reactivity entered a path model, the addition of Parents' perception of child Activity level did not account for additional variance. Child's Attention Span/ DiS‘tractibility variable was dropped because it relied on only maternal report. The basic logic underlying the first path model is that Parents' lifetime psychopathological problems are causally related to their current difficulties including their Personality problems, stress, and depression. Parents' Problems in turn are posited to contribute directly or indirectly to their children's behavior problems (Figure 2) . As indicated in Figure 2, parents' psychopathology was positively related to parental stress and inversely related to Ego Resiliency. Their rigid personality contributed to 88 Parents' Psychopathology (Lifetime Drinking & Antisocial Behavior) .3o*** -.35*** R2=== - 30(.29)----> Stress Parents' Ego . . 2 Re5111ency <----R =.35(.35) .36*** Mothers' Social Support <----R2=.36(.35) -.45*** 1 Parents' Current Depression < ----- R?=.74(.74) .35*** ,45*** Children's Reactivity .26** 4 Children's Behavior Problems <----R?=.59(.59) I"‘igmire 2 ‘ Path Model of Parents' Variables Predicting Mothers' Ratings 0f (Zhildren's Behavior Problems Path coefficient = beta weight * p<.05, ** p<.01, *** p<.001 Varniance information: (Adjusted R?) in parentheses. Overall Chi Square=12.94, df=12, n.s. 89 low emotional and instrumental support between spouses which in turn increased the level of parents' current depression (— - 45, p<.001) . High level of stress also increased the level of parents' current depression. Parents' current depression increased the level of children's behavior problems. Parents with higher levels of depression perceived their children as more reactive, and children who were perceived to be more reactive were perceived to have more behavior problems. Parents' psychopathological problems (alcohol problem and antisocial behavior) contributed to their child's behavior problem, but indirectly via parental ego resiliency, poor social support, high stress, and current depression. Overall the model significantly fit the data (Chi Square=12.94, df=12, ns) . In conclusion, the path analysis supported the hYE>othesis, that parents' antisocial behavior and lifetime alcohol problems predicted low ego resiliency and high Stress. However, there was no support for the hypothesis that parents' low ego resiliency and high undercontrolling functioning in their personality lead to their stress that was related to their depression. Rather the path analysis Showed that parents' ego resiliency predicted their social s"ilpport, that decreased their depression. Because parenting variables were dropped from the path analysis, all the hYpotheses related to parenting variables could not be tested in the present study. However, the correlation matrix (Appendix B, Table 20) showed that parents' ego 90 resiliency and depression were associated with alcoholic fathers' positive affective parenting, suggesting the possibility that parents' personality and depression might Iggyeariict parenting. Instead of finding evidence that supported the effects of parental depression and child temperament on inadequate parenting, the path analysis zirmcigicated the direct effects of parental depression and child temperament on children's behavior problems. This path model provides a clear cut idea about causal 1:e2].aationships among variables. However, it does not explain how the alcoholic and his spouse function in relation to ‘tlleejir children's adaptive behavior. The next two path nubdieels were constructed to test causal pathways jJacitependently for father and mother effects on children's behavior. Ci : co 1' Fathers The second model was designed to test how alcoholic fa'tl'iers influenced their children's adaptive behavior (Appendix B, Table 23) . Figure 3 shows the results of the Pafltla analysis. Fathers with high antisocial behavior showed 'hisali levels of alcohol problems. The high path coefficient between the two variables (.76) was actually a reproduced COrrelation between the variables, suggesting that the two constructs functionally are the same. For alcoholics' sPouses the path coefficient number between the two variables was .74. However, further analysis indicated that the two variables were highly associated but different 91 .76*** Fathers' 3 Fathers' < ----- R2=.76( 76) Antisocial Behavior Alcohol Problems .35*** -.32*** Fathers' Stress Fathers' .32*** Mothers' R2= -49(.49) ----- > Ego Resiliency ( Ego Resiliency \ I/p. Mothers' R2=.37(.37) ----- .70*** Social Support / -,31*** -017** . -.38** 2 / -.14** It :==.83(.82)----> Fathers' Depression L V .41*** Mothers' <---R?=.66 Depression (.65) .22* Fathers' Alcohol e I .26*____) \’Children's Problems Reactivity l .21” Children's Behavior <-/--R2=. 58(. 57) Problems Figure 3 Path Model of Alcoholic Fathers' Variables Predicting Mothers' Ratings of Children's Behavior Problems Path.coefficient= beta weight * p<. 05, ** p<. 01, *** p<. 001 Variance information: (Adjusted R2) in parentheses. Overall Chi Square=24. 8, df= 26; n. s. 92 constructs . Severe alcoholic fathers had less ego resiliency and more stressful events. Maternal ego resiliency was positively related to father's ego resiliency and father's ego resiliency was positively linked with maternal social support. The maternal social support decreased current depression for fathers and mothers. Mothers' ego resiliency contributed slightly but significantly to fathers' and mothers' current depression. Alcoholics' current depression increased their spouses' current depression. Both parents' current depression contributed significantly to level of child's adaptive functioning. Alcoholic fathers' current depression contributed to Chi 1dren' behavior problems directly and indirectly through mother's depression and parents' perception of Reactivity. Fathers with high level of alcohol problems perceived their children as more reactive. The effect of this causal relation was not large (.26, p<.05), but it was the only relationship that showed a specific contribution of al-Czoholic fathers' alcohol problems to children's behavior. C)Verall, the model significantly fit the data (Chi Square=24.81, df=26, ns). LQO e; 3: Alcoholics' Wives The third model was designed to test how alcoholics' Spouses influenced children's adaptive behavior (Figure 4). Results (Appendix B, Table 24) indicated that there was a strong correlation between spouses' Alcohol Problem and 93 .74*** Mothers' Alcohol { Mothers' Antisocial Problems Behavior —.25*** R2——=—-.38(.38) -------- > Fathers' .38*** Mothers' Ego Resiliency < Ego Resiliency .37** .28*** -.22*** Mothers' \, Social -.271|r*1|rA Fathers' Depression <----R2=.52(.51) Support ' r .25*** .23** \ .37*** \a 4’ -.29*** Mothers' Stress 1 \ \V R2= - 70 ( . 69) ---> Mothers' depression IA .21** .21* Children's Reactivity .25** V R2 ~V .17* =-60(.59) -------- > Children's behavior < . __ Mothers' problem Alcohol Problems Figure 4 Path Model of Alcoholics' Wives' Variables Predicting Hethers' Ratings of Children's Behavior Problems ‘ Path coefficient = beta weight ;p<.05, ** p<.01, “at p<.001 ariance information: (Adjusted R2) in parentheses. 0Vera11 Chi Square=11.72, df=22; n.s. 94 Antisocial Activity (reproduced correlation is .74) , but comparison of correlations between Alcohol Problems, Antisocial Behavior, and parental variables did not show a similar pattern of correlations among parental variables and children's behavior problems (Table 12) . In other words, although there was considerable overlap between Alcohol Problems and Antisocial Activity, the two variables for alcoholics' spouses were not the same construct. Alcoholic spouses' Antisocial Behavior was generally correlated with parental variables whereas their Alcohol Problems were correlated with few variables like alcoholic fathers' current depression and mothers' ratings of children behavior problems. However, there was a similar pattern in Correlations between paternal Alcohol Problems, Antisocial Behavior and paternal variables. Alcoholics' spouses with high level of antisocial behavior had more drinking problems. Spouses's alcohol PrOblems increased directly alcoholics' current depression and their children's behavior problems. Spouses with high 1eVe1 of antisocial behavior had less ego resiliency and reported more stressful events in their environment. The cialisal relationships among alcoholic fathers' and their wZ'LVes' ego resilient personality, social support, and c”.11-rent depression were observed in the same way that the Second model showed. Also the causal relationships among alcoholic fathers' and their wives' current depression, Children's Reactivity and children' behavior problems 95 appeared in the same direction that the second model displayed, except the addition of spouses' alcoholic problems that directly influenced the level of child's adaptive functioning (Fitzgerald, Ham, Zucker, Bruckel, Schneider, & Noll, in press) and alcoholic fathers' current depression. ‘ Another difference between the second model and the E third model was that alcoholic fathers' current depression increased their spouses' current depression directly and ' indirectly through spouses' stress. Mothers' stress contributed significantly to their current depression. Overall, the model significantly fit to the data (Chi Square=11.72, df=22, ns). In conclusion, the path analyses supported that the parents' antisocial behavior predicted their alcohol problems. Severe alcoholic fathers had low ego resiliency and high stress whereas severe antisocial mothers had low ego resiliency and high stress. The path analyses indicated that not only mothers' ego resiliency but also fathers' ego resiliency were related to mothers' social support, which in turn decreased their depression. Alcoholic fathers' depression influenced their wives' depression directly and indirectly. There was support the hypothesis that parents' stress predicted their depression; there was a particularly strong relationship between alcoholic fathers' stress and their depression. Finally, parents' alcohol problems influenced children's behavior problems: mothers' alcohol 96 problems had a direct effect on children's behavior, whereas fathers' alcohol problems had an indirect effect on children's behavior through their reactivity. DISCUSSION The alcoholic family environment is typically characterized as presenting a negative social environment that includes conflict, disorganization, violence, and poverty (Moos & Moos, 1984; Werner, 1986; Clair & Genest, 1987; Zucker, Weil, Baxter, & Noll, 1984; Leonard & Jacob, 1988; Russell, 1990) as well as a negative genetic environment including parental antisocial behavior and criminality (Cloninger, Bohman, & Sigvardssons, 1981; Cadoret, Cain, & Grove, 1980; Frances, Timm, & Bucky, 1980; Knop, Teasdale, Schulsinger, & Goodwin, 1985) . The degree of dysfunction in the alcoholic family environment is positively associated with the level of alcoholics' psychopathology including alcoholism, antisocial behavior, other psychiatric disorders (Penick, Powell, 0thmer, Bingham, Rice, & Liese, 1984), and the level of alcoholics' spouses' psychosocial functioning (Moos & Moos, 1984; Jacob & Leonard, 1986) . These variables are considered to be major risks that are likely to increase vulnerability for Psychopathology and substance abuse in children of alcoholics. The present study assumed that level of parents' l.mpc‘:l.i.:l:'ed psychosocial functioning would predict the level of 97 98 COAs' maladaptive functioning. It was expected that parental alcoholism, antisocial behavior, and/or depression would operate as risks that are directly related to level of COAs' behavior problems, whereas social support, stress, and low SES might be secondary variables that moderate the effect of risks on COAs' behavior. The present study used two strategies of analysis, multiple regression analysis and path analysis, in order to discern early precursors of COAs' psychopathology. This study was designed to assess how variables within the alcoholic environment serve as risk or protective mechanism in relation to children's adaptive functioning early in development. Pa s 'n A cohol'c Fa i 'es Alcoholic families constitute the highest risk groups for development of alcoholism and/or antisocial personality (McCord, 1988; Robins, 1978; Zucker & Gomberg, 1986). Thus, one should be able to identify more risks within the alcoholic family environment than within the nonalcoholic family environment (Zucker, 1989). The present study assumed that parents in alcoholic families have negative characteristics that are related to more negative environments for their children. Parents in alcoholic families often have coexisting psychiatric disorders, such as antisocial personality and depression that result in difficult social consequences including intrafamilial dysfunction. Families in which both parents show Psychopathological symptoms should be more dysfunctional 99 than families in which one parent shows such problems (Moos & Moos, 1984). Parents in alcoholic families had more Lifetime Alcohol Problems and Antisocial Behavior than did parents in comparison families (c.f., Fitzgerald, Sullivan, Ham, Zucker, Bruckel, Schneider, & Noll, in press). Furthermore, alcoholic fathers had lower Ego Resiliency or the absence of Ego Resiliency with Ego Undercontrolling characteristics, and reported more Current Stress and Depression than did nonalcoholic fathers. Considering the criteria for selection of alcoholic families for the MSU Longitudinal Study, these findings seemed to support the hypothesis that alcoholic fathers have more negative characteristics associated with high risk level of their families than nonalcoholic fathers in comparison families. The high Lifetime Alcohol Problems and Antisocial Behavior of alcoholics' wives may reflect the effects of "assortative mating" (c.f., Russell, 1990). Dual parents alcohol problems and antisocial behavior are likely to increase risk for COAs because they may inherit a predisposition toward alcoholism from each parent and the alcohologenic nature of the rearing environment may be increased (Russell, 1990). The fact that parenting was not correlated with parental ratings of children's behavior problems was surprising. Parents sorted statements describing their parenting in a Q-Sort format on the Block Child-Rearing Practices Report (CRPR) (Block, 1980). The CRPR, derived 100 from empirical observations of mothers interacting with their children in structured experimental situations measured parents' child-rearing attitudes, values, and practices. Susman, Trickett, Iannotti, Hellenbeck, and Zahn-Waxler (1985) suggest that the CRPR covers four domains of socialization: expression and regulation of emotions in parent-child interaction, parental discipline and control *- strategies of child's impulses and desires, parental goals and values for their children's achievement and development Y‘ of their children's autonomy and independence. Davies, Zucker, Noll, and Fitzgerald (1991), using 126 alcoholic families participating in the MSU Longitudinal Study, used principal component extraction and found for mothers positive vs. negative affective and authoritarianism vs. child-centeredness parenting, and for fathers nurturance vs. disengaged affective and authoritarianism vs. liberal parenting. The analysis of the current study failed to provide support for the hypothesis that high level of negative affective and/or authoritarian patenting likely increased the level of COAs' behavior problems. There are several possible explanations for the failure to find a relationship between parenting and child behavior problem. First, two factors of parenting, negative affective and authoritarian parenting (Davies, et al., 1991), are found not to be significantly associated with parental variables; only alcoholic fathers' positive affective parenting was significantly correlated with 101 fathers' Lifetime Alcohol Problems, Current Depression, Stress, and Ego Resiliency (See Table 18). Mothers' Antisocial Behavior was correlated with their authoritarian parenting and their Current Depression was correlated with their positive affective parenting in alcoholic families. However, most of these correlations were of low order, and only alcoholic fathers' positive affective parenting was associated with children's behavior problems. If parents' psychopathology is causally related to their negative parenting, which in turn is causally linked to their children's behavior problems, these results are inevitable. Since the first link in the causal relationships among variables, that parents' psychopathology is related to their negative parenting, is missing, the second link should be not found. The second possible explanation is that the CRPR that measured parents' child-rearing attitudes and values is not a sensitive measure of real interactions between parent and child. It is possible that measurement of parental attitudes and values (Susman, et al., 1985) in alcoholic families with young preschool-aged children does not reflect their behavior in the context within interactions with children occur. Only positive affective parenting of mothers in comparison families and fathers in alcoholic families was significantly correlated with their children's behavior problems. When alcoholic fathers' positive affective parenting entered into path analysis, the path 102 coefficient of this variable was zero. Thus, the parenting variable was dropped from analyses of this study. A third explanation is that there is no consistent parenting in the alcoholic family. COAs have been found to be physically and/or emotionally neglected by their parents (Werner, 1986; Rydelius, 1984). It is possible that young children might easily be neglected by their parents who are preoccupied with their personal and/or social difficulties. Finally, parents in young alcoholic families may not know how to play a role as parent; they may not be able to describe their parenting behavior on such a standardized instrument as the CRPR. Children in Alcoholic Families In contrast to the results of parental characteristics, which generally corresponded with other findings in the literature, comparisons between COAs and nonCOAs were not striking. Children in comparison families showed slightly higher but significant developmental status on several areas, such as Fine Motor, Adaptability, Language, and Personal/Social. There were high correlations among children's developmental status in these areas regardless of their families' alcohol status. Children in comparison families were evaluated as more Ego Resilient than were COAs. Block and Block (1979) found that children with high scores on Ego Resiliency at age 3 were able to recoup after stress, were verbally fluent, less anxious, less likely to externalize, or to withdraw under 103 stress at age 4. However, the strength of the effects on Ego Resiliency at age 3 were reduced at age 7 when social and cognitive functioning of children were considered. Children who were ego resilient at age 3 were described as task oriented and autonomous, and independent at age 4, whereas less resilient children were seen as critical and devaluing of both self and others. Block and Block concluded that positive patterns of social interaction were associated with children's early Ego Resiliency whereas negative interpersonal behaviors were associated with children's early Undercontrol. The current study found that children in comparison families were more ego resilient than were COAs [t(158)=2.39, p<.01], and COAs falling in the "clinical range" were seen as less ego resilient than were COAs falling in the "nonclinical range" [t(119)=1.71, p<.05]. In either case children's Ego Undercontrol did not discriminate children in different groups of families. This finding suggested in general that COAs might begin to have difficulty in areas related to social competence, as early as the preschool years, moreover, these effects may be even stronger for COAs in the "clinical range." However, there was no difference in parental reports of behavior problems related to psychopathology between COAs and children in comparison families. Similar results were obtained in studies of Jacob and Leonard (1986) and Fitzgerald et al. (in press) using the same instrument as used in this study, the Achenbach Child Behavior Checklist _— ,_.____—. 104 (CBCL) (Achenbach & Edelbrock, 1983) on different age groups of children. One possible explanation for these results is that alcoholic fathers in the three studies, including the current study, were selected from nonpatient populations, thus, the level of psychopathology might not be as severe as that in studies using clinical samples (Russell, 1990; West & Prinz, 1987). If the level of functioning in the alcoholic families of the three studies is not severely disturbed, then one might expect relatively low levels of children's psychopathology. If so, this would help to explain why no significant correlations between children's behavior problems, Ego Resiliency, Ego Undercontrol, and developmental status in alcoholic families were found. It suggests that in general, children in alcoholic families function similarly to children in comparison families, although, children in comparison families did show better developmental status. On the other hand, it is difficult to accept that all preschool-aged COAs in the current study are free from psychopathological symptoms, and that they suddenly have behavior problems during childhood. Maybe most preschool- age COAs are not severe now, but some COAs should be at greater risk than others due to their severely dysfunctional families (Jacob & Leonard, 1986). There are two assumptions underlying functioning of the alcoholic families to guide this study. First, there is considerable variability of dysfunction in alcoholic families, and this variability 105 likely accounts for the observed variability in psychopathology of children of alcoholics (West & Prinz, 1987). Second, the level of psychosocial functioning of alcoholics and their spouses is positively related to the degree of disrupted family functioning and the level of their offsprings' adaptive behavior (Moos & Moos, 1984; Seilhamer & Jacob, 1990). It is reasonable to assume that the dysfunction in alcoholic families disperses on a continuum ranging from the most severely disturbed to the least severely disturbed. It is likely that the most severely disturbed alcoholic families have children with low levels of adaptive functioning and high levels of psychopathology. Applying the criterion that Achenbach and Edelbrock (1983) used to discriminate between children who were most like their referred and nonreferred sample, Jacob and Leonard (1986) found that families of severely impaired COAs falling in the "clinical range" were described as more dysfunctional including parents' social and personal difficulties. The present study used a cutoff score similar to that used by Jacob et al. for Total Behavior Problems. One-third of COAs in the current study fell in this "clinical range," confirming the findings reported by Reider (1991) for children from the same longitudinal data base. Analysis of the family environment for "clinical range" COAs showed that their mothers had more negative characteristics than mothers of COAs falling in the 106 "nonclinical range", as indicated by measures of Lifetime Alcohol Problems, Antisocial Behavior, Ego Resiliency, Current Stress, and Depression. However, alcoholic fathers of children in the "clinical range" only scored higher in Lifetime Alcohol Problems and Current Depression than alcoholic fathers of children in the "nonclinical range." Jacob and Leonard (1986) found that the alcoholic families showed more alcohol-related difficulties, alcoholic fathers showed higher depression, and fathers and mothers had higher scores on the MMPI than alcoholic families who did not report severely impaired children. If it is true that alcoholics' spouses mediate the detrimental effects of alcoholism on their children (c.f., Fisher, Kokes, Cole, Perkins, & Wynje, 1987), mothers of children in the "clinical range" might not be providing the mediational buffering effect. Indeed, they may actually potentiate the negative effects of alcoholic father, thereby exacerbating the level of risk in the alcoholic families (c.f., Rutter, 1983, 1987; Sameroff & Seifer, 1983, 1987; Masten & Garmezy, 1985; Sroufe, 1983) Pgedighors of COAs' Psyghopgthology The above results reinforce the general finding that alcoholic parents provide poor nurturing environments for their children. Many studies examining the familial transmission of alcoholism focused on the causal relation between parental alcoholism and children's psychopathology. However, the conflicted findings from previous studies 107 suggest that alcoholics are a heterogenous group that should be classified into different subtypes according to their coexisting psychiatric symptoms, in as much as different alcoholic subtypes have different family history, levels of education and employment, social difficulties, and marital relationships (Penick, Powell, 0thmer, Bingham, Rice, & Liese, 1984). Heterogeneity among alcoholic families is likely to be associated with varying kinds and degrees of psychopathology and related social difficulties, including varying levels of psychopathology among spouses (Moos & Moos, 1984; Seilhamer & Jacob, 1990). In the present study measurement of COAs' psychopathology relied on the Achenbach Child Behavior Checklist (CBCL). This instrument is a standardized instrument that provides sex and age norms, and is designed to identify problem behaviors (Achenbach & Edelbrock, 1983). In addition, the CBCL assesses children's social competence, but since few subjects in the current study were enrolled in school, this part of the CBCL was not used. During wave 2 of the current study when all children will be in school, social competence can be assessed. It is expected that social competence will associated with more positive features that contribute to resilient characteristics (Garmezy, 1981; Werner & Smith, 1982, Masten & Garmezy, 1985; Sroufe, 1983). Orders of entry of independent variables for regression analysis (Table 11) were determined theoretically. Parents' 108 Lifetime Alcohol Problem, Antisocial Behavior, and Ego Resiliency were expected to reflect personality and behavioral characteristics established over the lifetime. Even though the California Adult Q-Sort does not demarcate present and past behaviors, it is assumed that personality reflects cumulative lifetime experiences rather than current experiences. On the other hand Social Support, Current Stress, and Depression are assumed to reflect current states. In the theoretical model it was assumed that parents' lifetime difficulties increased their current difficulties. Only one child characteristic, temperament, was used as the independent variable. It is possible that the relationships between children's temperament and their behavior problems can be attributed partially to parents' ratings. However, the observed correlations among these variables were not large enough to suspect this kind of question, and did not show any consistent pattern of parental ratings on both scales. For multiple regression analysis, four dependent variables were used for comparison; mothers' ratings of Total Behavior Problems (TBP), Externalizing Behavior, and Aggressive Behavior, and fathers' ratings of TBP. All the multiple regression analyses were significant, but there were few variables that remained as significant after all the independent variables entered in each equation. Two variables predicted independently children's behavior 109 problems; mothers' perception of Attention Span/Distractibility predicted their ratings of children's Aggressive Behavior, whereas alcoholic fathers' Current Depression predicted their ratings of children's TBP. Children who were perceived by their mothers to have high Attention Span were rated as less aggressive. Children's Attention Span/Distractibility may be a reliable predictor of COAs' behavior problems, especially related to either hyperactivity or conduct disorder (Windle, 1990; West & Prinz, 1987), since it is a single variable which shows low but significant positive correlation with children's Ego Resiliency and negative correlation with children's Ego Undercontrol. Other studies have showed that children of alcoholics have adjustment problems due to their lack of concentration (Rydelius, 1984) and impulsivity (Knop, Teasdale, Schulsinger, & Goodwin, 1985) and tend to have poor achievement at school. In addition, in the present study maternal ratings of Aggression were weakly but significantly associated with children's Ego Undercontrol. In comparison families a similar but stronger relation between the two variables was observed. Demographic variables are believed to be related to parental psychopathology. Particularly, low SES is frequently found in alcoholic families (Russell, 1990). In the current study correlations between parental variables and demographic variables were not very impressive except Years of Marriage. Multiple regression analyses suggested 110 that demographic variables such as family SES, parents' Age and Eduction, and Years of Marriage did not contribute to children's different behavior problems in the same way. Some of the variance for mothers' ratings of children's TBP and Externalizing Behavior was accounted for by Years of Marriage but not by family SES and parents' Age and Education. But all demographic variables contributed to maternal ratings of children's Aggression. In addition, mothers' education and maternal ratings of Attention Span independently predicted children's Aggression. The observed correlations among variables are not high enough to indicate serious problems with multicollinearity. Since the independent variables are unlikely to be totally uncorrelated (Shcroeder, Sjoquist, & Stephan, 1987), one possible solution for multicollinearity is to combine the related variables into one variable (Lewis-Beck, 1986). Thus, the variables that were theoretically related and highly correlated were combined. Lifetime Alcohol Problem and Antisocial Activity were combined and labeled Lifetime Difficulty. Current Stress and Depression were combined and labeled Current Difficulty. Combining variables had little effect on the regression equations. The variance explained by children's TBP, Externalizing Behavior, and Aggression was similar to that explained by the original independent variables. Children's Reactivity predicted children's Externalizing behavior and children's Attention Span/ Distractibility predicted their Aggression. Fathers' 111 Current Difficulty predicted their ratings of children's TBP. It was found that parents' perception of children's Reactivity was significantly related to alcoholic fathers' Current Depression; alcoholic fathers with high levels of Current Depression perceived their children as more reactive. This finding suggests that parents' emotional state influenced their ratings of children's actual behavior (Shaughency & Lahey, 1985). The present study found a weak relation between mothers' Current Depression and their rating of children's Reactivity and a strong relationship between parents' Current Depression and their ratings of children's TBP. Schaughency and Lahey (1985) reported a relationship between maternal depression and maternal perception of children's externalizing behavior problems, but they also found that maternal ratings and teachers' ratings of children's behavior were strongly correlated. Parental perception of children's temperament is believed to influence parent-child interactions (Belsky, 1984; Rutter, 1984, 1987; Tarter, Alterman, & Edwards, 1985; Werner, 1986; Werner & Smith, 1982). Werner (1986) found that more of the resilient COAs who did not develop serious mental and coping problems by age 18 had received a great deal of attention from their caretakers during the first year of life and were perceived as "cuddly and affectionate." Children with difficult temperament were more likely than other children to be the target of parental 112 hostility, criticism, and irritability (Rutter, 1987). Temperamentally difficult children were found to be less adaptable to change and more vulnerable to adversity than were temperamentally easy children (Hetherington, 1989). In a longitudinal study of the effects of divorce on children's adjustment, Hetherington compared parents' retrospective report of children's temperament and temperament ratings made by nurses during the first 2 years of life with later children's behavior. The nurses' ratings of infant temperament predicted later children's behavior better than mothers' reports of infant temperament did, whereas the mothers' reports of infant temperament predicted their own later behavior toward their children, and their report was weakly but significantly correlated with later children's behavior. Hetherington studied the role of children's temperament on their adjustment to their parents' divorce. The results seemed to generally support Rutter's (1987) claim that temperamentally difficult children were more likely to be both the elicitor and the target of parents' criticism. However, Hetherington found effects of maternal variables on interactions between children with difficult temperament and maternal personality and stress. Unless temperamental difficulty was compounded by such maternal variables, there were no differences in mothers' negative responses to their children's temperamental difficulty, compared with children having easy temperament. These findings suggest that temperament of children in alcoholic 113 families has the potential to significantly influence the quality of parent-child interaction. A series of multiple regression analyses were conducted for only parents' psychopathological problems such as Lifetime Alcohol Problem, Antisocial Behavior, and Current Depression. The amount of variance accounting for children's behavior problems appeared to decrease, suggesting that variables such as maternal Social Support, parents' Current Stress and Ego Resiliency contributed significantly to children's Externalizing Behavior and Aggression. However, there was not much difference in the amount of variance accounted for by mothers' ratings of children's TBP when entering all the variables and when entering only parents' psychopathological problems. This result suggested that parents' psychopathology was more related to children's TBP. Parents' Current Depression was a significant predictor of children's behavior problems. Alcoholic fathers' and their spouses' Current Depression predicted children's TBP. Alcoholics' spouses' Current Depression predicted children's Externalizing behavior whereas Alcoholics' Current Depression predicted children's TBP of fathers' rating. This finding indicated that parents' current depression was a strong predictor of children's behavior problems. One possible explanation for this relationship (Rutter, 1990) is that parents' current depression is more likely to be involved in adverse interactions with other family members (Colletta, 1983; Moos 114 & Moos, 1984; Phares & Compas, 1992). Families in which alcoholic fathers had more drinking problems and complained of more depression and physical symptoms reported more family arguments, less cohesion and expressiveness, and showed less agreement about their family environment between spouses (Moos & Moos, 1984). Jacob and Leonard (1986) found that alcoholic fathers of children falling in the "clinical range" reported higher depression than did alcoholic fathers of children in the "nonclinical range." T-test results of the present study indicated that alcoholic fathers showed higher depression than did fathers in comparison families and alcoholic fathers and their wives in alcoholic families of children in the "clinical range" reported higher depression than those in alcoholic families of children in the "nonclinical range." Again only the combined variables Lifetime Difficulty and Current Difficulty entered for multiple regression analysis. All the regression analysis were significant. The amount of variance for children' behavior problems decreased. Mothers' Lifetime Difficulty predicted children's TBP, Externalizing Behavior, and Aggression. This finding seemed interesting because alcoholics' spouses had more Lifetime Alcohol Problem and Antisocial Behavior than did nonalcoholic' spouses, but there were no differences in other parental variables between the two groups of spouses. This result indicated the effect of alcoholics' spouses' psychopathology on children's adaptive 115 functioning. Alcoholic fathers' Lifetime and Current Difficulty predicted their ratings of children's TBP. Davies, Zucker, Noll, and Fitzgerald (1989) found that the characteristics of the father-child relationship were independent of the mothers' psychological functioning, using the sample of the MSU Longitudinal Study. Inspection of a correlation matrix from the present study actually suggests similar conclusions; alcoholic fathers' psychosocial functioning was correlated with both mothers' ratings and fathers' ratings of children's behavior problems, but maternal psychosocial functioning was not significantly associated with fathers' ratings of children's behavior problems. This finding suggests that alcoholic fathers' psychosocial variables are likely to influence their marital relationship, which, in turn, influences mother-child interactions. Path analysis showed direct and indirect effects of alcoholic fathers on children's behavior. Path hodel of Alcoholic Families Three path models derived from a theoretical model were examined. The first path model tested how parental variables, alcoholics' and alcoholics' wives combined, lead to children's behavior problems. There are several reasons why parental variables should be combined. The results of multiple regression analysis did not find specific contribution of parents' psychopathology to children' behavior problems except parents' depression, suggesting some problems with multicollinearity. To improve this 116 problem parental variables that were theoretically related and strongly correlated, were combined into a single variables. Using the combined variables with other parental variables in multiple regression analysis did not provide better results. Especially there were fairly strong associations between alcoholics' and their spouses' Lifetime Alcohol Problem and Antisocial Behavior. However, whether alcoholics' psychopathology influenced their spouses' problems or alcoholics' spouses' psychopathology influenced alcoholics' problems can not be determined. Nor can it be determined whether women with alcohol problems have a tendency to marry alcoholics (c.f., Russell, 1990). However, the first model using combined parental variables was not contaminated by the issue of direction of effects. The problem with the first model was that it did not provide a way to test how alcoholic' and spousal variables interact within the alcoholic family environment. The second and third path models dealt with this problem, that is, these models assessed how alcoholic' and spousal variables causally influenced their children's behavior problems. Parents' Alcohol Problems and Antisocial Behavior reflect their impaired psychosocial functioning (Penick, Powell, 0thmer, Bingham, Rice, & Liese, 1984; Hesselbrock, Hesselbrock, Babor, Stabenau, Meyer, & Weidenman, 1984; Reider, Zucker, Noll, Maguin, & Fitzgerald, 1986). Alcoholic parents tended to be less well educated, have low level occupations, difficulty with interpersonal 117 relationships including marital relationship, and trouble with the law (Penick, et al., 1984; Werner, 1986; McCord, 1988). Thus, it was almost inevitable that parents with a high level of alcohol problems and antisocial behavior reported high levels of stress in their social environment. It is likely that parents with a long history of alcohol problems and antisocial behavior have dysfunctional personality (Zucker & Gomberg, 1986; Rutter & Giller, 1983). Block's two constructs of personality (Block & Block, 1979), Ego Resiliency and Ego Undercontrol, did not deal with abnormal personality in clinical setting. However, the two constructs seem to show a general level of a person's adaptive ability within a social context. The ego undercontroller is expected to have a low level of threshold for response, to manifest desires and impulses directly into behavior, to want the immediate gratification of desires, and to be easily distracted. The ego resilient person is expected to accommodate well to the environment, to have coping or problem solving ability, and to recover well under and after stress. The description of the ego undercontroller seems to fit well the image of an alcoholic. Block, Block, and Keyes (1988) found in their longitudinal study that the presence of ego undercontrol and the absence of ego resiliency were likely to be positively related to adolescents' drug use. In the present study, observers rated alcoholic fathers to be more ego undercontrolled and less ego resilient. 118 However, since parents' ego undercontrol was not significantly correlated with other parental variables and child variables, only parents' ego resiliency was included in the analysis. Parents with a high level of lifetime alcohol problem and antisocial behavior were associated with low level of ego resiliency. According to Block and Block's (1979) definition of Ego Resiliency, an ego resilient person is seen to be an adaptable and functional individual within the social context. In the current study low levels of ego resiliency is assumed to reflect low levels of adaptive functioning, rather than their specific maladaptive behaviors. Personality is considered to be a general psychological resource that people draw upon to help them withstand stress in their environment (Pearlin & Schooler, 1978). Parents' high lifetime alcohol problem and antisocial behavior are expected to show defects in such a general psychological resource, that is likely to serve as a risk mechanism for children's development in the alcoholic family environment. Parents' personality is believed to influence both the marital relationship and parenting behavior (Belsky, 1984). Alcoholics and their spouses who used avoidance coping were more associated with family arguments and the extent of disagreement on joint tasks, whereas alcoholics' spouses who used active coping strategies agreed more closely with their husbands on the joint tasks (Moos & Moos, 1984). The presence of marital support influenced parenting behavior of 119 women of high risk group, who were reared in institutions, indicating protective processes associated with marital support (Rutter, 1987). The presence of social support suggests that there should be at least one significant intimate relationship with another person (Crockenberg, 1988; Sullivan, Maguin, Fitzgerald, Zucker, & Noll, 1990). Sullivan et al. (1990), using alcoholic families participating in the MSU Study, found that alcoholics' spouses felt less liked and loved by their partner and felt that they could less depend on their husbands in a crisis than did spouses in comparison families. The current study, combined emotional and instrumental aids of spouses and other family members, did not find any significant difference in social support between spouses in alcoholic families and in comparison families. It infers that families in which alcoholics' wives receive emotional and instrumental support from their husbands will have less marital conflict. The social support for mothers is strongly related to her psychological well-being, which in turn is associated with her behaviors directed toward children (Belsky, 1984; Colleta, 1981; Powell, 1981). The results of the present study indicated that alcoholics' spouses' social support decreased their current depression and their husbands' current depression. Since alcoholics' spouses' social support is especially associated with their husbands' support, it is likely that low levels of parental depression results from their 120 supportive marital relationship. Support in marital relationship decreased parents' current depression whereas parents' stress increased their current depression. Depression is one of the most frequent co-morbid symptoms that alcoholics have reported (Russell, Henderson, & Blume, 1984; Penick, 0thmer, Bingham, Rice, & Liese, 1984). Forty-eight percent of a sample of antisocial alcoholics were positive for depression and/or mania in Penick et al's study. Moos and Moos (1984) found that alcoholic families in which both spouses complained of depression and physical symptoms had more arguments and less cohesion than those in which only one spouse had such complaints. The question is why parents' depression is related to children's behavior problems. The present study hypothesized that parents' depression increased parents' negative affective and authoritarian parenting, which in turn increased children' maladjustment. But since the results of the study did not find any causal relationship between parenting and children's behavior problems. Thus there is no such causal link in the path model. The results of another study (Susman, Trickett, Iannotti, Hollenbeck, & Zahn-Waxler, 1985) indicated that depressed mothers expressed inconsistency in discipline and hostility in rearing situation. Longfellow and Belle (1984) explained that depressed mothers were less responsive, less nurturing, and more punitive to their young children. It is hardly 121 expected that depressed mothers serve as a buffering system for children to reduce the impact of stress in environment. Children's temperament influenced children's adaptive functioning. Parents' current depression was associated with their perception of children's Reactivity, but also alcoholic fathers' lifetime alcohol problems were related to children's Reactivity. Parents' depression, especially fathers' depression, might influence their perception of children's temperament, but parent's perception of Reactivity showed some congruence with an observer's rating of children's behavior, Ego Undercontrol. Path Models of Alcoholics and Thei; Wives The second model of alcoholics and the third model of alcoholics' spouses are discussed together. Parents' antisocial behavior predicted parents' alcohol problems (Zucker & Gomberg, 1986; McCord, 1988). The high correlation between the Lifetime Alcohol Problem and Antisocial Behavior suggested that the two instruments measured the same constructs. However, even though there was considerable overlap between two variables (c.f., Robins, 1978), they were believed to be different constructs, since there were alcoholic probands without antisocial characteristics (Penick, Powell, 0thmer, Bingham, Rice, & Liese, 1984) and probands with antisocial personality without alcoholism (Hesselbrock, Hesselbrock, Babor, Stabenau, Meyer, & Weidenman, 1984). There were two findings to support that antisocial behavior and alcohol 122 problems were highly associated but different constructs in the current study. Antisocial alcoholics had different sociodemographic background, family history of psychiatric disorder, and drinking history, compared with other alcoholic subtypes (Hesselbrock, Hesselbrock, Babor, Stabenau, Meyer, & Weidenman, 1984; Penick Powell, 0thmer, Bingham, Rice, & Liese, 1984). Parents' antisocial behavior . .5. a. W Uri-.17 was significantly correlated with most demographic variables, whereas their lifetime alcoholic problems was not. Parents' antisocial behavior was associated with low family SES, low parents' education, and younger age. Path analyses showed that alcoholics' Lifetime Alcohol Problems and their spouses' Antisocial Behavior - not their Lifetime Alcohol Problems - operated in a similar way in the model. Alcoholics' alcohol abuse and related difficulties predicted low Ego Resiliency and high stress, whereas their spouses' antisocial characteristics predicted low Ego Resiliency and high stress. It should be noted that more mothers of COAs in the "clinical range" showed an absence of Ego Resiliency compared with mothers of COAs in the "nonclinical range," whereas there was no difference in Ego Resiliency between alcoholic fathers of COAs in the "clinical range" and COAs in the "nonclinical range." Alcoholics' spouses with ego resilient characteristics obtained more support in their marital relationship and their relationship with other family members. Alcoholics' spouses' ego resilient 123 characteristics predicted weakly but significantly low level of their husbands' current depression, indicating the importance of alcoholics' spouses' adaptive ability in the alcoholic family environment. The influence of alcoholics' wives' adaptive functioning on alcoholics' behavior including their support and depression, contrasts with the effect of alcoholics' depression on their wives' depression. These findings indicate that there are bidirectional interactions between alcoholics and their wives in the alcoholic family environment. Mothers' social support decreased their current depression (Colletta, 1981). Mothers' social support was negatively associated with alcoholic fathers' current depression, but spouses' social support did not have a direct effect on their alcoholic husbands' depression. Alcoholic families in which spouses obtained their husbands' support probably have less family discord and marital conflict. Alcoholics in families that reported less marital conflict might be less depressed than alcoholics in families experiencing conflict among family members. Social support decreased alcoholics' spouses' current depression, whereas their husbands' current depression increased directly and indirectly through spouses' stress and spouses' current depression. Sources of alcoholics' spouses' current depression derive from their social support, stress, and their husbands' depression, however sources of alcoholics' current depression are their stress 124 and maternal social support. In the second model and the third model there was a considerable difference in the amount of variance accounting for alcoholics' current depression; .82 in the second model including fathers' stress variable, which results from their drinking problems and .51 in the third model without their stress variable, in which alcoholics' spouses' drinking problems influence significantly their husbands' current depression. Thus, spouses' drinking problems do not seem to be a source of stress to alcoholics. However, for alcoholics' spouses there is little difference in variance accounted for spouses' current depression; .65 in the second model without their stress variable and .69 in the third model including their stress. The results suggested that alcoholics' wives' depression was significantly influenced by their husbands' depression, however, alcoholics' depression was associated with their stress. Thus, alcoholics' spouses likely mediate between alcoholics' depression related to their personal and social difficulty and children's adaptive functioning. Alcoholics' and their spouses' current depression predicted their children's behavior problems as compared with other parental variables. Considering only parental psychopathology, hierarchical multiple regression analysis showed that only parents' depression predicted children's TBP. Path analyses showed that alcoholics' current depression is the most powerful and significant variable to lead to children's behavior problems. Path analysis showed 125 that alcoholics' current depression resulted from their personal and social difficulties, related to their alcohol abuse. Regarding effects of parental depression on children Rutter (1990) questions whether impaired parenting stems from parental depression, or from other risk factors leading to depression. As already indicated there is a missing link between parents' depression and children's behavior problems. Only some speculations are possible right now. Alcoholic family environments in which parents report more alcohol problems, antisocial behavior, and depression may be characterized as having more conflict and adverse interactions among family members, and as being less nurturing (Seilhamer & Jacob, 1990; Moos & Moos, 1984). Such variables associated with the negative alcoholic family environment seem to serve as risks in relation to COAs' adaptive functioning. Parents' ego resilient characteristics, stress and maternal social support seem to serve as mediators of other parental variables such as parents' alcohol problems, antisocial behavior, and depression. The maternal social support and parents' ego resiliency are likely to be involved in protective processes with respect to children' adaptation by decreasing levels of parents' current depression, whereas stress is likely to be engaged in exacerbating processes by increasing levels of parents' current depression. Demographic variables such as family SES, parents' age and education, and years of marriage are likely to moderate the 126 effects of risk in alcoholic families. Children's temperament influences children's adaptive behavior by involving parent-child interactions: COAs' Reactivity and Attention Span/Distractibility are related to their psychopathology. The findings of the present study showed a complex context of alcoholic family environment in which parental and children variables interact with each other to lead to COAs' psychopathology. Measurement of alcoholic fathers' problems (e.g., alcoholism) does not explain the process of how alcoholic family environment as a system is involved in COAs' adaptive functioning. These results indicate the importance of multimeasurement of alcoholic family environment (Zucker, 1989; Zucker & Noll, 1982; Phares & Compas, 1992) that has multiple risks and of focus on the process of how variables observed within the alcoholic family environment function in relation to children's behavior. The basic focus of previous alcoholic studies is on why COAs become alcoholics rather than why some COAs do not become alcoholics in adulthood. It is important to understand why and how resilient COAs acquire sufficient competence to resist negative factors in their adverse environments whereas other COAs are victimized by such factors. Many people interested in the etiology of alcoholism realize that study of adaptive COAs can provide important information about normal as well as abnormal 127 development (West & Prinz, 1987), especially for development of intervention programs for COAs. The present study showed that COAs acquired lower scores on RYDS developmental status and were rated as less ego resilient by an observer than did children in comparison families, whereas COAs falling in the "nonclinical range" were seen as more ego resilient. In future studies it will be interesting to determine whether these children turn out to be resilient COAs or vulnerable COAs, and how variables in the alcoholic family environment influence the processes affecting the child's development. APPENDIX A Informed Consent Form 128 RESEARCH PARTICIPATION INFORMED CONSENT PORN (T-l) We freely consent to take part in a scientific study of child development and family functioning being conducted by Drs.Robert A.Zucker and Hiram E.Fitzgerald (from the Department of Psychology at Michigan State University) and Dr.Robert B.Noll (from the Children's Hospital Medical Center at the University of Cincinnati). The study has been explained to us and we understand the explantion that has been give and what our partcipation and what the participation of our child(ren) will involve. We understand that we will be asked to fill out questionnaires, and take part in interviews, and that our child(ren) will do the same. We have been informed that participation in this phase of the project involves taking part in 9 sessions with a time commitment of approximately 11 hours for each adult and approximately 7 hours for our child. We understand that we are free to discontinue our participation in the study at any time without penalty; we also are aware that our involvement or lack of involvement in the study, or in different subparts of it, will not affect our obtaining treatment that is offered by your staff in any other capacity. We understand that the results of the study will be treated in strict confidence and that we and our child(ren) will remain anonymous. Within these restrictions, results of the study will be made availabe to I us periodically throughout the course of the project and for I a minimum of 3 years after the projects has concluded. Also | within the restrictions noted above, we understand that general results of the research will appear in professional journals and will be presented at scientific meetings. We understand that our participation in the study does not guarantee any beneficial results to us or to members of our family. 129 SIGNATURES Informed Consent: Mother Father Assent: Child Date Witness Family code number 4/91 Fwd APPENDIX B Statistical Tables Table 18 Mean Scores for Parents' 130 Characteristics of COAs in the "Clinical Range" and COAs in the "Nonclinical Range" Alcoholic Families (Clinical Range) Alcoholic Families (Normal Range) M §2 Alcoholic Problems Mothers 11.21 2.42 Fathers 11.25 1.91 Antisocial Behavior Mothers 15.55 9.56 Fathers 25.75 11.84 Depression Mothers 5.16 4.46 Fathers 4.97 4.05 Personality: Ego Resiliency Mothers -.11 .33 Fathers -.15 .29 Personality: Ego Undercontr Mothers -.05 .20 Fathers .05 .19 Stress Mothers 26.31 12.92 Fathers 22.10 15.25 Social Support Mothers 15.46 2.93 Fathers 14.70 4.15 l3 9.79 10.46 11.71 22.51 .08 -.11 01 -.09 .05 20.09 19.35 16.10 15.86 E -3.52 -2.13 -2.51 -1.07 -3.08 -3.55 2.52 .71 -1.02 .17 134 134 129 129 129 127 124 117 124 117 115 115 115 107 .00 .05 .01 ns .00 .00 .01 ns ns ns .02 ns ns ns 131 Table 19 Mean Scores for Behavior Problems, Temperament, Personality, and General Developmental Status for COAs in'the "Clinical Range" and COAs in the "Nonclinical Range" Alcoholic Alcoholic Families Families (Clinical Range) (Normal Range) M §D M £2 E Q: E CBCL Total Behavior Problems Mothers 69.42 5.75 53.54 7.19 -11.94 146 .00 Fathers 61.26 9.55 55.01 10.34 -3.17 143 .02 Externalizing Behavior Mothers 68.40 6.46 52.85 8.60 -9.86 146 .00 Fathers 61.09 10.50 54.71 10.31 -3.22 143 .00 Internalizing Behavior Mothers 64.80 7.44 52.26 6.95 -9.17 146 .00 Fathers 58.31 8.55 53.15 10.03 -2.75 143 .01 Aggression Mothers 71.43 8.37 58.14 4.92 -11.63 146 .00 Fathers 64.71 9.98 60.25 7.63 -2.78 143 .01 Tempergment Rhythmicity Mothers 4.32 2.41 5.54 2.01 2.77 122 .01 Fathers 4.94 2.10 5.32 2.12 .88 122 ns 132 Table 19 (cont'd) M §2 M £2 E Qfi 2 Reactivity Mothers 3.87 1.46 3.04 1.76 -2.36 122 .02 Fathers 4.13 1.09 3.15 1.43 -3.49 122 .00 Attention Span/Distractibility Mothers 3.71 3.06 5.71 2.80 3.37 122 .00 Fathers 4.10 2.99 4.94 2.89 1.39 122 ns Adaptibility Mothers 4.39 1.61 4.09 1.66 -.88 122 ns Fathers 4.35 1.14 3.89 1.54 -1.53 122 ns Activity Level Mothers 1.71 1.42 1.19 1.42 -1.75 122 .05 Fathers 2.19 1.25 1.55 1.44 -2.23 122 .02 Begsonality ER .02 .33 .13 .31 1.71 119 .05 EU .09 .21 .06 .19 -.71 119 ns RYDS Developmental Status GM 48.34 12.89 49.38 12.49 .39 117 ns PM 50.93 11.59 52.00 10.95 .45 117 ns AD 51.89 10.55 53.28 9,61 .66 117 ns LAD 50.11 13.51 52.69 11.00 1.04 117 ns PS 52.61 11.58 53.64 9.68 .47 117 ns DQ 51.07 11.14 53.01 9.40 .92 117 ns Note. ER=Ego Resiliency. EU=Ego Undercontrol. GM=Gross Motor Development. FM=Fine Motor Development. AD=Adaptibility. LAN=Language Development. DQ=Overall Development PS=Persona1/Social Development. Quotient. 133 Table 20 Correlations between Parental Variables and Positive Affective and Authoritarian Parenting Alcoholic Families Comparison Families Affective Authoritarian Affective Authoritarian M F M F M F M F MLAPSR -.14 .02 .10 -.07 -.01 .05 -.15 .30 MTASB -.16 -.10 .20* .06 -.05 -.08 .18 .10 MER .14 .25** -.18 -.14 .24 -.11 .02 .19 MSTRESS .02 -.01 .11 -.05 -.09 .08 -.14 -.03 MTBDI -.21* -.20* .13 .08 -.02 -.21 -.21 -.08 FLAPSR -.06 -.19* .04 -.16 -.07 -.06 -.06 .12 FTASB -.02 -.05 .16 -.07 -.26 .16 .16 -.33 FER .07 .25** -.26** -.00 .02 .22 .22 -.21 FSTRESS -.04 -.24** .28** -.13 -.20 -.2 -.22 .05 FTBDI -.01 -.36** .24** .12 -.08 -.15 -.15 .10 MTBP -.06 -.22* .08 .08 -.42* .23 .23 -.18 MEXT -.09 -.20* .06 .06 -.33 .11 .11 -.33 MINT -.00 -.23* .15 .11 -.40* .30 .30 .02 MAGG -.18 -.21* .05 .03 .40 .05 .05 -.23 FTBP -.07 -.14 .03 .03 -.24 -.06 -.06 -.12 FEXT -.07 -.14 .06 .00 -.24 -.01 -.01 -.08 FINT -.07 -.13 -.03 .08 -.19 .00 .00 -.04 FAGG -.07 -.20* .10 -.03 -.18 -.03 -.03 -.10 * p<.05, ** p<.01. Two-tailed. Note. M=Mothers. F=Fathers. MLAPSR=Mothers' Lifetime Alcohol Problems. MTASB=Mothers' Antisocial Behavior. MER=Mothers' Ego Resiliency. MSTRESS=Mothers' Stress. MTBDI=Mothers' Depression. FLAPSR=Fathers' Lifetime Alcohol Problems. FTASB=Fathers' Antisocial Behavior. FER=Fathers' Ego Resiliency. FSTRESS=Fathers' Stress. FTBDI=Fathers' Depression. MTBP=Mothers' Ratings of Total Behavior Problems. MEXT=Mothers' Ratings of Externalizing Behavior. MINT=Mothers' Ratings of Internalizing Behavior. MAGG=Mothers' Ratings of Aggression. FTBP=Fathers' Ratings of Total Behavior Problems. FEXT=Fathers' Ratings of Externalizing Behavior. FINT=Fathers' Ratings of Internalizing Behavior. FAGG=Fathers' Ratings of Aggression. 134 Table 21 Correlations among Demographic Variables, Parental Variables and Children's Behavior Problems in Alcoholic Families FYSES Mother Father Mother Father Years Age Age Educat- Educat- Marri- MLAPSR -.02 .08 .00 -.06 -.08 -.30** MTASB -.32** -.18* -.14 -.37** -.12 -.27** MER .33** .17 .09 .28** .18* .22* MSTRESS -.06 -.15 -.06 -.01 -.14 -.25* MTBDI -.17 -.02 -.13 -.11 -.27** -.11 SUPPORT .01 .01 -.01 -.01 -.08 .11 FLAPSR -.18* -.03 -.05 -.06 -.06 -.14 FTASB -.34** -.29** -.19* -.28** -.19* -.20* FER .18 .13 -.14 .09 .10 .22* FSTRESS -.13 -.03 .06 -.17 -.08 .19 FTBDI -.22* -.13 -.05 -.26** -.09 -.31** MTBP -.22* -.18* -.13 -.14 -.25** -.31** MEXT -.23** -.18* -.16 -.10 -.24** -.27** MINT -.25 -.23* -.12 -.17 -.31 -.37** MAGG -.13 -.14 -.15 -.05 -.22** -.37** FTBP -.23** -.10 -.07 -.09 -.07 -.23* FEXT -.24** -.07 -.07 -.08 -.13 -.19* FINT -.21* -.14 -.11 -.10 -.06 -.24* FAGG -.26** -.11 -.09 -.09 -.12 -.20* * p<.05, ** p<.01. Two-tailed. MLAPSR=Mothers' Lifetime Alcohol Problems. MTASB=Mothers' Antisocial Behavior. MER=Mothers' Ego Resiliency. MSTRESS=Mothers' Stress. MTBDI=Mothers' Depression. FLAPSR=Fathers' Lifetime Alcohol Problems. FTASB=Fathers' Antisocial Behavior. FER=Fathers' Ego Resiliency. FSTRESS=Fathers' Stress. FTBDI=Fathers' Depression. MTBP=Mothers' Ratings of Total Behavior Problems. MEXT=Mothers' Ratings of Externalizing Behavior. MINT=Mothers' Ratings of Internalizing Behavior. MAGG=Mothers' Ratings of Aggression. FTBP=Fathers' Ratings of Total Behavior Problems. FEXT=Fathers' Ratings of Externalizing Behavior. FINT=Fathers' Ratings of Internalizing Behavior. FAGG=Fathers' Ratings of Aggression. Note. 135 Table 22 Path Analysis of Parental Variables Predicting COAs' Total Behavior Problems Correlations (Corrected for attenuation) 1 2 3 4 5 6 7 1) Psychopathology -- 2) Ego Resiliency -35 -- 3) Stress 30 ~23 -- 4) Social Support -17 36 -13 -- 5) Depression 45 -36 59 -52 -- 6) Child's Reacti- 21 -2s 23 -13 35 -- 7) Child's TBP 26 -20 30 -27 54 42 -- Standard Errors for Path Coefficientsa and Path Coefficientsb 1 2 3 4 5 6 7 1) Psychopathology -- 5a 6 2) Ego Resiliency -35b -- 6 3) Stress 30 -- 6 4) Social Support 36 -- 7 5) Depression 53 -45 -- 9 7 6) Child's Reacti- 35 -- 9 7) Child's TBP 45 26 -- Reproduced Correlations 1 2 3 4 5 6 7 1) Psychopathology -- 2) Ego Resiliency -35 -— 3) Stress 30 -23 -- 4) Social Support -13 36 -8 -- 5) Depression 22 -28 57 -49 -- 6) Child's Reacti- 8 -10 20 -17 35 -- 7) Child's TBP 12 -15 31 -27 54 42 -- .11 136 Table 22 (cont'd) Errors: (Actual - Reproduced) 1 2 3 4 5 6 7 1) Psychopathology -- 2) Ego Resiliency 0 -- 3) Stress 0 0 -- 4) Social Support -4 0 -5 -- 5) Depression 23 -8 2 -3 -- 6) Child's Reacti- 13 -15 3 4 0 -- 7) Child's TBP 14 -5 -1 0 0 0 -- The sum of squared errors in the lower triangle is .132 The reliabilities of the variables in the model are 1) Psychopathology 1.00 2) Ego Resiliency 1.00 3) Stress 0.94 4) Social Support 0.75 5) Depression 0.79 6) Child's Reacti- 0.53 7) Child's TBP 0.94 Table 23 137 Path Analysis of Alcoholic Fathers' Variables Predicting COAs' Total Behavior Problems Correlations (Corrected for attenuation) 1 2 3 1 -- 2 76 -- 3 -25 -38 -- 4 19 35 -19 5 33 48 -17 6 20 38 -12 7 23 29 -10 8 -12 -25 37 9 -21 -19 38 10 29 39 -11 Path Coefficients 1 2 3 1 2 76 3 -32 4 35 5 6. 7 8 37 9 10 26 69 19 21 18 70 5 6 57 -- 48 44 -38 -50 -32 -35 34 28 5 6 41 27 21 22 7 -17 -22 42 8 -38 -31 32 -17 -14 10 10 27 138 Table 23 (cont'd) l-‘KDmQC‘UI-hUNl-J HomQOSUI-bUNb-J HOQQGUfiUNH Standard Errors for Path Coefficients 1 2 3 4 5 6 7 6 .7 7 5 8 9 9 6 13 12 Reproduced Correlations 1 2 3 4 5 6 7 76 -- -25 -38 -- 27 35 -13 -- 26 33 -30 73 -- 16 20 -29 32 58 -- 17 22 -18 33 47 42 -- -9 -14 37 -5 -44 -51 -26 -21 -19 38 -7 -27 -30 -16 25 33 --16 25 30 18 39 Errors: (Actual - Reproduced) 1. 2 3 4’ 5 (5 7 0 __ 0 () -- -8 0 -6 -- 7 15 13 -4 -- 4 18' 17 -13 -1 -- 6 7 8 -12 l 2 -- -3 -11 0 11 6 1 9 0 0 0 -1 -5 -5 -6 4 3 5 -¥7 4 10 3 14 -13 10 10 10 139 Table 23 (cont'd) The sum of squared errors in the lower triangle is .281 The reliabilities of the variables in the model are 1) Fathers' Lifetime Alcohol Problems 0.90 2) Fathers' Antisocial Behavior 0.63 3) Fathers' Ego Resiliency 1.00 4) Fathers' Stress 0.96 5) Fathers' Depression 0.79 6) Mothers' Depression 0.80 7) Child's TBP 0.94 8) Social Support 0.75 9) Mothers' Ego Resiliency 1.00 10) Child's Reactivity 0.53 140 Table 24 Path Analysis of Alcoholic Wives' Variables Predicting COAs' Total Behavior Problems Correlations (Corrected for attenuation) 1 2 23 4 5 6 7 £3 9 10 1 —. 2 -25 -- 3 -27 25 -- 4 41 -35 -50 -- 5 28 ~32 -38 57 -- 6 30 -21 -25 47 31 -- 7 21 -29 -13 28 34 18 -- 8 33 -22 -27 44 48 21 42 -- 9 -26 38 37 -12 -17 -11 -11 -10 -- 10 74 -11 -20 22 36 23 22 35 -16 -- Path Coefficients 1 2 3 4 5 6 7 8 9 10 1 2 -25 3 37 4 -29 37 28 5 -22 -27 28 6 23 25 7 34 8 21 21 25 17 9 38 10 74 141 Table 24 (cont'd) HOGQGU‘bUNI-J O meQmU'I-bUNi-J H‘OGQOU‘I-RUNH Standard Errors for Path Coefficients 1 2 3 4 5 6 7 6 8 8 7 7 8 7 7 9 9 10 9 5 6 Reproduced Correlations 1 2 3 4 5 6 7 -25 -- -27 25 -- 29 -23 -47 -- 34 -32 -38 57 -- 31 -14 -16 45 32 -- 11 -11 -13 20 34 11 -- 29 -16 -25 43 48 24 39 -26 38 37 -22 -23 -12 -8 74 -11 --20 27 36 26 12 Errors: (Actual - Reproduced) 1 2 3 4 5 6 7 o -- 0 0 -- 12 -12 -3 -- -6 0 0 0 -- -1 -7 ‘ -9 2 -1 -- 10 -18 0 8 0 7 -- 4 -6 -2 1 0 -3 3 0 0 0 10 6 1 -3 0 0 0 -5 0 10 10 8 9 3 8 9 -14 -_ 34 -16 8 9 4 -_ 1 0 10 10 10 142 Table 24 (cont'd) The sum of squared errors in the lower triangle is .139 The reliabilities of the variables in the model are 1) Mothers' Antisocial Behavior 0.83 2) Mothers' Ego Resiliency 1.00 3) Social Support 0.75 4) Mothers' Depression 0.80 5) Fathers' Depression 0.79 6) Mothers' Stress 0.96 7) Child's Reactivity 0.53 8) Child's TBP 0.94 9) Fathers' Ego Resiliency 1.00 10) Mothers' Lifetime Alcohol Problems 0.65 LIST OF REFERENCES LIST OF REFERENCES Achenbach, T.M., & Edelbrock, C. (1983). Manual for the chilg behavior checklist and revised child behavior profiles. Burlington: Univeristy of Vermont. Ackerman, R.J. (1983). Children of alcoholics: A guidbooh for educators, therapists, and parents. 2nd Ed., Holmes Beach, Fla.: Learning Publications. Beck, A.T., & Beck, R.W. (1972). Screening depressed patients in family practice: A rapid technique. gospgpadpate hedicine, pg, 81-85. Beck, A.T., Steer, R.A., & Garbin, M.G. (1988). Psychometric properties of the Beck Depression: Twenty-five years of evaluation. Clinical Psychology Review, §, 77-100. Belsky, J. (1984). The determinants of parenting: A process model. Child Development, pp, 83-96. Block, J. (1961/1978). The Q-sort method in personality gssessmehh hpd psychiatric pesearch. Springfield, Illinois: Charles C. Thomas. Block, J.H., & Block, J. (1979). The role of ego-control and ego- resiliency in the organization of behavior. In W.A. Collins (Ed.), Minne§ota Sypposia on Child Psychology. Vol.13, Hillsdale, N.J.: Lawrence Erlbaum Associates, 39-101. Block, J.H. (1980). The Child-rearing Eractices Report (CRPR): A get pf Q items for the desciption of parental socilizahion attitudes and values. Unpublished manuscript, Univeristy of California, Institute of Human Development, Berkeley. Block, J.H., Block, J., & Morrison, A. (1981). Parental agreement-disagreement on child-rearing orientation and gener-related personality correlates in children. Child Development, pg, 965-974. 143 144 Block, J., Block, J.H., & Keyes, S. (1988). Longitudinally foretelling drug usage in adolescence: Early childhood personality and environmental precursors. Child Qevelopment, p2, 336-355. Brook, J.S., Whiteman, M., Gordon, A.S., & Cohen, P.C. (1986). Dynamics of childhood and adolescent personality traits and adolescent drug use. Developmental Psychology, lg, 403-414. Brown, S. (1986). Children with an alcoholic parent. In N.J. Estes & M.E. Heinemann (Eds.), Alcoholism: Development, conseguences, and intervention, 3rd ed., St.Louis: The C.V. Mosby Company, 207-220. Cadoret, R.J., Cain, C.A., & Grove, W.M. (1980). Development of alcoholism in adoptees raised apart from alcoholi biological relatives. Archives of General Psychiatry, g1, 561-563. Clair, D., & Genest, M. (1987). Variables associated with the adjustment of offspring of alcoholic fathers. Journal of Studies on Alcohol, pp, 345-355. Cloninger, C.R., Bohman, M., & Sigvardsson, S. (1981). Inheritance of alcohol abuse: Cross-fostering analysis of adopted men. r hives of Gene a s ch'at , lg, 861-868. Colletta, N.D. (1981). Adolescent mothers's vulnerability to stress. Journal of Nervous and Mentgl Disease, 169, 50- 54. Colletta, N.D. (1983). At risk for depression: A study of young mothers. Journal of Genetic Psychology, 14;, 301- 310. Cotton, N.S. (1979). The familial incidence of alcoholism: A review. Journal of Studies on Alcohol, 59, 89-116. Cronckenberg, S. (1988) Social Support and Parenting. In H.E.Fitzgerald, B.M.Cester, & M.W.Yogman (Eds.), Theory and research in behavioral pediatrics, Vol.4. N.Y.: Plenum Press, 141-174. Davies, W.H., Zucker, R.A., Noll, R.B., & Fitzgerald, H.E. (in press). Early socialization practices in alcoholic families: The relationship of child-rearing patterns to demographics and parental psychopathology. ‘Donovan, J.E., & Jessor, R. (1985). Structure of problem behavior in adolescence and young adulthood. Journal of Qopsulhihg ahd Clinical Psychology, pg, 890-904. 145 Donovan, J.E., & Costa, F.M. (1988). Syndrome of Problem Behavior in Adolescence: A replication. Journal of Consulting and Clinical Psychology, pp, 762-765. El-Guebaly, N., & Offord, D.R. (1977). The offspring of alcoholics: A critical reviews. American Journal of Psychiatry, 134, 357-365. Fisher, L., Kokes, R.F., Cole, R.B., Perkins, P.M., & Wynne, L.C. (1987). Competent children at risk: A study of well-functioning offspring of disturbed parents. In E.J. Anthony & B.J. Cohler (Eds.), The invulnerable child, N.Y.: The Guilford Press, 211-228. Fitzgerald, H.E., Sullivan, L.A., Gover, M.R., Maguin, E.T., Zucker, R.A. (1990). Parents differ in their temperment ratings of children at risk for alcoholism. Paper presented at the American Psychological Meetings, Dallas, TX. Fitzgerald, H.E., Sullivan, L.A., Ham, Zucker, R.A., Bruckel, S. Schneider, A.M., & Noll, R.B. (in press). Predictors of behavioral problems in three-year-old sons of alcoholics: Early evidence for the onset of risk. Frances, R.J., Timm, S., & Bucky, S. (1980). Studies of familial and nonfamilial alcoholism. Achieves of General Psychiatry, l1, 564-566. Funder, D.C., Block, J.H., & Block, J. (1983). Delay of gratification: Some longitudinal personality correlates. Joprnal of Personality and Social Psychology, 43, 1198-1213. Garmezy, N. (1981). Children under stress: Perspectives on antecedents and correlates of vulnerability and resistance to psychopathology. In A.I. Rabin, J. Aronoff, Aj.M. Barclay, & R.A. Zucker (Eds.), Fupther explorations in personality, N.Y.: Wiley, 196-269. Garmezy, N. (1983). Stressors of Childhood. In N.Garmezy & M.Rutter (Eds.), Stress, coping, ang development in chilgren. Baltimore: The Johns Hopkins University, 43- 84. Garmezy, N., Masten, A., & Tellegen, A. (1984). The study of stress and competence in children: A building block for developmental psychology. Child Development, pp, 97- 111. Helzer, J.E., & Pryzbeck, T.R. (1988). The co-occurrence of alcoholism with other psychiatric disorders in the 146 general population and its impact on treatment. Journal pf Shudies oh Alcohol, 52, 219-224. Hesslbrock, M.N., Hesselbrock, V.M., Babor, T.F., Stabenau, J.R., Meyer, R.B., & Weidenman, M. (1984). Antisocial behavior, psychopathology and problem drinking in the natural history of alcoholism. In D.W. Goodwin, K.T. Van Dusen, & S.A. Mednick (Eds.) Longitudinal research oh alcoholism. Boston: Kluwer-Nijhoff, 197-214. Hetherington, E.M. (1984). Stress and coping in children and families. In A.B., Doyle, D.Gold, & D.S. Moskowitz (Eds.), Children in families under stress. San Francisco: Jossey-Bass, 7-33. Hetherington, E.M. (1989). Coping with family transitions: Winners, losers, and survivors. Child Development, pp, 1-14. Hunter, J.E., & Gerbing, D.W. (1982). Unidimensional measurement, secondary order factor analysis, and causal models. Research in Organizational Behaviop, Vol.4, 267-320. Jacob, T., & Leonard, K. (1986). Psychosocial functioning in children of alcoholic fathers, depressd fathers, and control fathers, Jourhal of Sthdies oh Algohol, 51, 373-380. Kanner, A., Coyne, J., Schaefer, C., & Lazarus, R. (1981). Comparison of two modes of stress measurement: Daily hassles and uplifts versus major life events. Jgprnal of Behavioral Medicine, A, 1-39. Knop, J., Teasdale, T.W., Schulsinger, F., & Goodwin, D.W. (1985). A prospective study of young men at high risk for alcoholism: School behavior and achievement. Journal of Studeis on Alcohol, Ag, 273-278. Leonard, K.E., & Jacob, T. (1988). Alcohol, alcoholism, and family violence. In V.B.Van Hasselt, R.L.Morrison, A.S.Bellack, & M.Hersen (Eds.), Handbook of Family Viglence. N.Y.: Plenum Press, 383-406. Lewis-Beck, M.S. (1986). Applied regression: An ihhroguction. Sage University Paper series on Quantitative Applications in the Social Sciences, Beverly Hills and London: Sage Publications. Lewis, M., Feiring, C., McGuffog, C., & Jaskir, J. (1984). Predicting psychopathology in six-year-olds from early social relations. Child pevelopmenp, pp, 123-136. 147 Longfellow, C., 8 Belle, D. (1984). Stressful environments and their impact on children. In J.E. Humphrey (Ed.), Stress, in childhood, N.Y.: Ams Press, inc., 63-78. Masten, A.S., 8 Garmezy, N. (1985). Risk, vulnerability, and protective factors in developmental psychopathology. In B.B. Lahey, 8 A.E. Kazdin (Eds.), Advances in clinical child psychology, Vol.8, N.Y.: Plenum Press, 1-51. McCord, J. (1988). Indentifying developmental paradigms leading to alcoholism. Journal of Studies on Alcohol, 52, 357-362. Moos, R.H., 8 Moos, B.S. (1984). The process of recovery from alcoholism: III comparing functioning in families of alcoholics and matched control families. Journal of Studies on Alcohol, Ad, 111-118. Murray, J.B. (1989). Psychologists and children of alcoholic parents. Psyghological Reports, pg, 859-879. Noll, R.B., 8 Kulkarni, R. (1989). Cognitive and motor development of infants coping with cancer: Longitudinal observations. Infant Mental Health Jo rnal, lg, 252- 262. Norbeck, J.S., Lindsey, A.M., 8 Carrieri, V.L. (1981). The development of an instrument to measure social support. Nursing Research, lg, 264-269. Pearlin, L.I., 8 Schooler, C. (1978). The structure of coping. Journal of Health and Social hehaviog, l2, 2- 21. Penick, E.C., Powell, B.J., 0thmer, E., Bingham, S.F., Rice, A.S., 8 Liece, B.S. (1984). Subtyping alcoholics by coexisting psychiatric syndromes: Course, family history, outcome. In D.W. Goodwin, K.T. Van Dusen, 8 S.A. Mednick (Eds.) Longitudinal research on alcoholism. Boston: Kluwer-Nijhoff, 167-195. Phares, V., 8 Compas, B.B.:(1992). The role of fathers in child and adolescent psychopathology: Make room for daddy. Psychological Eulletin, 111, 387-412. Powell, D.R. (1980). Personal social network as a focus for primary prevention of child maltreatment. Infanp Mental Health Journal, l, 232-239. Provence, S. (1964). Revised Yale developmental schedule: A composite of Gesell examination withlgelectpd itepg from Merrill-Palmer Stanford-Binet and Hetzer Woolf. Unpublished manuscript, Yale Univeristy, Yale Child 148 Study Center, New Haven. Reider, E.E., Zucker, R.A., Noll, R.B., Maguin, G., 8 Fitzgerald, H.E. (1988). Alcohol involvement and family violence in a high risk sample: I. Spousal violence. Presented at the American Psychological Association Meetings, Atlanta, GA. Reider, E.E. (1991). Relationships between parental psychopathology, family conflict, and child behavior problems in young alcoholic families. Unpublished doctoral dissertation, Michigan State University, East Lansing, MI. Robins, L.N. (1978). Sturdy childhood predictors of adult antisocial behavior: Replications from longitudinal studies. Psychological Medicine, A, 611-622. Robins, L.N., Helzer, J.H., Croughan, J.,8 Ratcliff, K.S. (1980). The NIMH Diagnostic Interview Schedule: It'g history, characteristics and validity. St. Louis: Washington Univeristy School of Medicine. Russell, M., Henderson, C., 8 Blume, S.B. (1984). Childpen of alcoholics: A review of the literature. N.Y.: Children of Alcoholics Foundation. Russell, M. (1990). Prevalence of alcoholism among children of alcoholics. In M. Windle 8 J.S. Searles (Eds.), Children of alcoholics: Critical perspectives. N.Y.: The Guilford Press, 9-38. Rutter, M. (1983). Stress, coping, and development: Some issues and some questions. In N.Garmezy 8 M.Rutter (Eds.), Stress, coping, and development in children. Baltimore: The Johns Hopkins University, 1-42. Rutter, M., 8 Giller, H. (1983). Juvenile delinguency: Trends and perspectives. Harmandsworth, Middlesex, England: Penguin. Rutter, M. (1984). The family, the child, and the school. In M.D. Levin 8 P. Satz (Eds.), Middle Childhood: Development and dysfunction. Baltimore: University Park Press, 293-343. Rutter, M. (1987). Psychosocial resilience and protective mechanisms. American Journal of Orthopsychiatry, 51, 316-331. Rutter, M. (1990). Commentary: Some focus and process considerations regarding effects of parental depression on children. Developmental Psychology, Ad, 60-67. 149 Rydelius, P. (1984). Children of alcoholic fathers - a longitudinal prospective study. In D.W. Goodwin, K.T. Van Dusen, 8 S.A. Mednick (Eds.), Longitudinal research on alcoholism. Boston: Kluwer-Nijhoff, 27-37. Sameroff, A.J. (1980). Issues in early reproductive and caretaking risk: Review and current status. In D.B. Sawin, S.R. Hawkins, L.O. Walker, 8 J.N. Penticuff (Eds.), Exceptional infant (Vol.4), Psychosocidl risks in infant-environment transactions. N.Y.: Brunner/Mazel, 343-359. Sameroff, A.J., 8 Seifer, R. (1983). Familial risk and child competenc. Child Developmenh, pg, 1254-1268. Schroeder, L.D., Sjoquist, D.L., 8 Stephan, P.E. (1987). Understanding regression analysis. Sage University Paper series on Quantitative Applications in the Social Sciences, Beverly Hills: Sage Pubns. Seifer, R., 8 Sameroff, A.J. (1987). Multiple determinants of risk and invulnerability. In E.J. Anthony 8 B.J. Cohler (Eds.), Invulnerable child, N.Y.: The Guilford Press, 51-69. Seilhamer, R.A., 8 Jacob, T. (1990). Family factors and adjustment of children of alcoholics. In M. Windle 8 J.S. Searles (Eds.), Children of alcoholics: Critical perspectives. N.Y.: The Guilford Press, 168-186. Selzer, M.L. (1975). A self-administered Short Michigan Alcoholism Screening Test (SMAST). Joughal of Stpdies on Alcohol, lg, 117-126. Shedler, J., 8 Block, J. (1990). Adolescent drug use and psychological health. American Psychologist, Ad, 612- 630. Shaughency, E.A., 8 Lahey, B.B. (1985). Mothers' and fathers' perceptions of child deviance: Roles of child behavior, parental depression, and marital satisfaction. Journal of Cohsulting and Clinical Psychology, dd, 718-723. Sroufe, L.A. (1983). Infant-caregiver attachment and patterns of adaptation in preschool: The roots of maladaptation and competence. In M. Perlmutter (Ed.), The Minnessota Sypposia on Child Psychology, vol.16. Hillsdale N.J.: Lawrence Erlbaum Associates, 41-83. Stevens, 6., 8 Featherman, D.L. (1981). A revised socioeconomic index of occupational status. Sgcial Science Research, lg, 364-395. 150 Sullivan, L.A., Maguin, E.T., Fitzgerald, H.E., Zucker, R.A., 8 Noll, R.B. (1990). Social support and dlcgholism: An examination of spousal support. Paper presented at the Eastern Psychological Association meetings. Philadelphia, PA. Susman, E.J., Trickett, P.K., Iannotti, R.J., Hollenbeck, B.B., 8 Zahn-Waxler, C. (1985). Child-rearing patterns in depressed, abusive, and normal mothers. American dpprhdl pf Orthopsychiatry, dd, 237-251. Tarter, R.B., Alterman, A.I., 8 Edwards, K.L. (1985). Vulnerability to alcoholism in men: A behavior-genetic perspective. Journal of Studies on Alcohol, Ad, 329- 356. Werner, E.E., 8 Smith, R.S. (1982). Vulnerability but ihvipcible: A longitudinal study of resilient children dhd youth. N.Y.: McGraw-Hill. Werner, E.E. (1986). Resilient offspring of alcoholics: A longitudinal study from birth to age 18. Journal of Shudies op Alcohol, 51, 34-40. West, M.D., 8 Prinz, R.J. (1987). Parental alcoholism and childhood psychopathology. Psychological Bulletin, 102, 204-218. Windle, M. (1990). Temperament and personality attributes of children of alcoholics. In M. Windle 8 J.S. Searles (Eds.), Children of alcoholics: Chitical perspegtives. N.Y.: The Guilford Press, 129-167. Zucker, R.A., 8 Barron, F.H. (1973). Parental behaviors associated with problem drinking and antisocial behavior among adolescent males. In M.E. Chafetz (Ed.), Proceedings of the Firtst Annual Alcoholism Cohference of the udtional Institute on Alcohol Abuse and Alcoholism. Washington, DC: NIAAA, 276-296. Zucker, R.A., 8 Noll, R.B. (1980). Assessmehp of dhpisocial behaviop: Development pf an instggment. Unpublished paper, Deparment of Psychology, Michigan State Univeristy, East Lansing, MI. Zucker, R.A., 8 Noll, R.B. (1982). Precursors and developmental influences on drinking and alcoholism: etiology from a longitudinal perspective. In N.I.A.A.A. (Ed.) Alcohol dhd Health Monographs, No.1: Algohol consumption and related problems. Rockville, MD, National Institute on Alcohol Abuse and Alcoholism, 289-327. 151 Zucker, R.A., 8 Baxter, J.A., Noll, R.B., Theado, D.P., 8 Weil, C.M. (1982). An alcoholic risk study: Design and early health relateg findings. Presented at the American Psychological Association Meetings, Washington, D.C. Zucker, R.A., Noll, R.B., Draznin, T.H., Baxter, J.A., Weil, C.M., Theado, D.P., Greenberg, G.S., Charlot, C., 8 Reider, E. (1984). The ecology of alcoholic families: Conceptual framework for the Michigan State University longistudinal study. Presented at the National Council on Alcoholism, National Alcoholism Form Meetings, Detroit, MI. Zucker, R.A., Weil, C.W., Baxter, J.A., (1984). Differences in interpersonal and individual psychopathology in young families at high risk for the development of alcoholism. Presented at the Society of Life History Research in Psychopathology Meetings, Johns Hopkins University, Baltimore, MD. Zucker, R.A., 8 Gomberg, E.S.L. (1986). Etiology of a alcoholism reconsidered: The case for a biopsychosocial process. American Ps cholo ist, Al, 783-793. Zucker, R.A. (1987). The four alcoholism: a developmental account of the etiologic process. In P.C. Rivers (Ed.), ebraska S osium on Motivation 1986: V0 . gg. Alcohol and Additive Bepaviors. Lincoln, NE: University of Nebraska Press, 27-83. Zucker, R.A. (1988). Developmental factors and gisk and the beginnings of alcohol related problems. Presented at 17th International Medical Advisory Conference. Zucker, R.A. (1989). Is risk for alcoholism predictable? A probabilistic approach to a developmental problem. Drugs and Society, A, 69-93. Zucker, R.A., Fitzgerald, H.E., 8 Noll, R.B. (1990). Drinking and drug history: Revised edition (version 4). Unpublished instrument, Department of Psychology, Michigan State University, East Lansing, MI. Zucker, R.A. (1991). Scaling the developmental momentum of alcoholic process 1991via the Lifetime Alcohol Problems Scores. Alcohol and Alcoholism. Zucker, R.A., 8 Fitzgerald, H.E. (1991). Early developmental factors and risk for alcohol problems. Alcohol Health and Reseach World, l_. Zucker, R.A. 8 Fitzgerald, H.E. (1991). Risk and coping in 152 children of alcoholics: Years 6 to 10 of the Michigan State University Longitudinal Study. Grant proposal submitted to the National Institute on Alcohol Abuse and Alcoholism.