u u 4} D I." ullptrdvvv I. £D. Y.II.I. .; IIII. .E u I o I.“ fit. '1 In. nvI II J! I! II III I II. .v: I; no 4". : I I}... I r «I v n III! . II A . . t ”We, i til. I... v. 5‘ us)... I) A I ‘4‘ I n 1.0 I r fivflfiflw . “Immawmhu III I.' r . ..f..u.‘fid.. : 'I .o a) . III-I. .A. I. .J ..I /J". :3. IV: y L J”; .wur.£.ht.r . .‘II III . an“. n II: . pr ILII. p . I :IOIIII A I . til! . . 1,- e . J I .T .1...» VHF)? I LIY‘fsl...) » . II > 'I MIC CHIG TE UNI ERSITY IAHIES I sit IWM‘HI'IIHHII I 3_OOO736 LIBRARY 1 Michigan State '1 University I This is to certify that the dissertation entitled Hassles, Social Supports, Symptoms and Alcohol Involvement: Their Interaction in Young Alcoholic and Nonalcoholic Families presented by Constance M. Weil has been accepted towards fulfillment of the requirements for _P_1LD_._ degree in 25% fl/flyéa- Date.Sep1:embeI_15_._1987 MS U i: an Affirmative Action/Equal Opportunity Institution 0- 12771 g MSU RETURNING MATERIALS: Place in book drop to LIBRARIES remove this checkout from .—_. your record. FINES wiIT be charged if book is returned after the date stamped below. . -1 / "GINO-1595’ ‘7 usIIIc-r 3’ W 7 $3.9, v 4% ’41:" jgffim/ , W, M jigt 1 V. HASSLES, SOCIAL SUPPORTS, SYMPTOMS AND ALCOHOL INVOLVEMENT: THEIR INTERACTION IN YOUNG ALCOHOLIC AND NONALCOHOLIC FAMILIES BY Constance M. Weil A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Psychology 1987 ABSTRACT HASSLES, SOCIAL SUPPORTS, SYMPTOMS AND ALCOHOL INVOLVEMENT: THEIR INTERACTION IN YOUNG ALCOHOLIC AND NONALCOHOLIC FAMILIES BY Constance M. Weil This study examines the interaction between drinking behavior, daily hassles, and social support, and their impact on the emotional and behavioral adjustment of marital partners in young families where the children are at heightened risk for the development of later alcoholism. The work is a part of the Michigan State University Longitudinal Study. The goal of the larger project is to explore the nature of the intergenerational transmission of alcoholism; the current study is an initial step in this process that focuses on the relationship between alcohol involvement and adjustment of marital partners in families that vary in risk status. Self-report and interview measures of current and lifetime alcohol involvement, hassles, social support, physical illness and psychological symptoms were collected from thirty couples whose husbands were identified through the court system or neighborhhod search. Men identified via the courts had been arrested for drunk driving and had registered blood alcohol levels of 0.15% or higher when apprehended. All of these men met formal diagnostic criteria for probable or definite alcoholism, although they varied considerably in level of lifetime alcohol involvement. To increase the dispersion of risk, a second subset of families were recruited from the same census tracts as the court identified subjects and were comparable in socio-demographic characteristics, but none of the men reached alcoholic diagnostic criteria. This combined sample was utilized to insure a diversity of parent drinking and a diversity of alcohol related problems. Results of correlational and multiple regression analyses show significant relationships between level of alcohol involvement, daily hassles and adjustment. However, the nature of these relationships is different in men and women, is different for different symptom types (externalizing or internalizing), and varies depending upon whether the alcohol involvement measure is of current drinking or of cumulative lifetime involvement. To my family and friends. Without you, this work would not have grown into the dissertation that it is and I would not have grown into the person that I am. Thank you. ii ACKNOWLEDGMENTS I would like to thank my chairperson, Robert A. Zucker, Ph.D., who has been a teacher, adviser, and supporter throughout my graduate career. Continued thanks are given to Robert B. Noll, Ph.D. and John Paul McKinney, Ph.D. who have been friends as well as teachers during this process. Thanks are also given to Susan Frank, Ph.D for serving on my dissertation committee and sharing her insights about this work. iii TABLE OF CONTENTS Page LIST OF TABLESOOOOOOOOO0......OOOOOOOOOOOOOOOOOOOOOOIOOVi LIST OF FIGURESOOOOOOOO0.00.00.00.00...0.00.00.00.000Viii CHAPTER I. Review of the Literature Statement of the Problem and Rationale for the ResearChOOOOOI0.00...0....OOOOOOOOOOOOOOOOOOOOOO0.01 Review of the Relationship of Alcoholism, Stress, and Adjustment.........................................6 Behavior of the Alcoholic Husband and the Non-alcoholic Wife.................................6 Interactions Between an Alcoholic Husband and His spouse...‘0..O.I.0......00.00.000.00...0.0.0.0000009 Interaction and Adaption in Alcoholic Couples........13 Review of Stress Literature.............................24 The Dynamics of Stress...............................24 Life Stress and Symptomatology in Adults.............25 Life Stress and its Measurement in Adults............3l Social Support, Stress, and Disorder.................42 summarYOOOOOOOOOOOOOOOOOOOOO0.0.0.0.0....0.0.00.00.00.0053 Formal Predictions.0.0.0....0.000......OOOOOOOOOOOOOOOOOSS II. Method Subjects................................................59 Recruitment Procedures..................................60 Data Collection Procedures..............................61 Measures................................................62 Measurement of Stressful Life Events.................63 Measurement of Adult Symptomatology..................66 Measurement of Depression.........................67 Measurement of Antisocial Behavior................7O Measurement of Health History.....................72 Measurement of Social Support........................72 Measurement of Drinking Behavior.....................76 iv TABLE OF CONTENTS (cont'd.) III. Results Analysis................................................81 Hassles and Alcohol Involvement.........................91 Symptom Occurence and Alcohol Involvement...............95 Social Support and Alcohol Involvement..................99 The Interaction of Stress & Symptomatology in Couples..lOl Interaction of Alcohol Consumption, Stress, Support and Symptomatology......................................114 IV. Discussion Hassles and Alcohol Involvement........................l33 Symptom Occurence and Alcohol Involvement..............138 Social Support and Alcohol Involvement.................l44 The Interaction of Stress and Symptomatology in Couple.l48 Interaction of Drinking Behavior, Stress, Support and Symptomatology......................................154 Modifications and Additions to the Model of the Stress-Illness Relationship in a Drinking Couple....155 Methodological Issues and Problems.....................168 Measurement Issues.....................................170 Future Directions......................................l72 APPENDICES I. Contact and Data Collection Schedule............l74 II. Norbeck's Social Support Questionnaire and the Social Support Interview...........................l75 III. The Relationship Between Daily Uplifts and Measures of Self Drinking and Spouse Drinking......184 IVA. The Relationship Between Different Symptoms Among Women..............................................185 IVB. The Relationship Between Different Symptoms Among HenOOOOOOOO00.0.0.0...0.0.0....OOOOOOOOOOOOOOOOOO0.186 BIBLIOGRAPHY...00.0.0.0...000......0.0.00.0000000000000187 Table 8A 8B 8C 8D 10 11 LIST OF TABLES Page Significant Correlations Between Four Types of Subjective Evaluations of Life Events and Psychosocial Adjustment.........................37 Summary of Study Measures.......................64 Demographic Characteristics of the Sample.......85 The Relationship Between Hassles and Measures of Self Drinking and Spouse Drinking............92 The Relationship Between Symptomatology and Measures of Self Drinking and Spouse Drinking...96 The Relationship Between Social Support and Measures of Self Drinking and Spouse Drinking..loo The Relationship Between Husbands' and Wives' Daily HaSSIeSOOOOOOOOOOOOOOOOO...0.0.0.0000....102 The Relationship Between Husbands' and Wives' Internalizing Symptomatology...................105 The Relationship Between Husbands' Internalizing and Wives' Externalizing Symptomatology........106 The Relationship Between Husbands' Externalizing and Wives' Internalizing Symptomatology........108 The Relationship Between Husbands' and Wives' Externalizing Symptomatology...................109 The Relationship Between Symptomatology and Measures of Self Hassles and Spouse Hassles....111 Hierarchical Regressions Predicting Men's and Women's Symptomatology.........................116 Percent of the Variance Accounted for When Predicting Men's and Women's Symptomatology....123 vi 12 13 14 LIST OF TABLES (cont'd.) Summary of Independent Variables Contributing Significantly to the Prediction of Men's and Women's Symptomatology.........................126 Stepwise Regressions for Predicting Men's symptomatOIOgYOOOOOOOOOOOOOOOOOOO0.0.0.0000....127 Stepwise Regressions for Predicting Women's symptomatOIOgYOOOOOOOOOCOOOOOOOOOOOOOO00.......128 vii Figure 10 LIST OF FIGURES Page Model of the Life Stress-Illness Relationship...29 Model of the Stress-Illness Relationship in a Drinking coupleOOOOO...OOOOOOOOOOOOOOOOOOOOOIOO.54 Frequency Distribution of QFV Scores for Men andwomenOOOOOOOOOOOOOOOOOOO0.0.0.0....0..0.82 Frequency Distribution of Number of Drinking Related Problems For Men and Women..............83 Frequency Distribution of Lifetime Alcohol Problem Scores (LAPS) For Men and Women.........84 Outline of Tables on the Relationship Between Husbands' and Wives' Symptomatology............lO4 Model to Predict Internalizing Symptomatology in men.0.0...0.0...0..O0.0.0.0.0.00000000000000164 Model to Predict Externalizing Symptomatology in Hen...000......OOOOOOOOOOOOOOOOOOOOOOOOO0.0.165 Model to Predict Internalizing Symptomatology in women.OOOOOOOOOOOOOOOOOOOO0.00.00.00.0000000166 Model to Predict Externalizing Symptomatology in women...-OOOOOOOOOOOOOOO0.0.00.00.00.0000000167 viii Chapter 1 Statement of the Problem and Rationale for the Research The purpose of this study is to examine the relationship between drinking behavior, daily hassles, and the structures of social support and, to explore the relationship of these factors to the emotional and behavioral adjustment of marital partners. The interaction between husbands' and wives' adjustment and stress is also examined. A special focus of this work is the exploration of this interplay in a population of young families at heightened risk for the development of alcoholism among their offspring. These offspring are at risk because they have a parent who has a high enough level of alcohol consumption that he would be characterized as alcoholic (Cotton, 1979). As a result, this work focuses on a high-risk population and explores one set of factors that may impact on the intergenerational transmission of alcoholism, i.e. the interrelationships between drinking, stress, and adjustment. However, the impact of the adult stress- illness relationship 9g offspring is beyond the scope of the present study; this is taken up in other aspects of 2 the Michigan State University Longitudinal Study of which this work is a part. In the present research, stress is considered to be any type of event that upsets the equilibrium in an individual's life and that requires some sort of adaption or coping response in order to reestablish the equilibrium. Theoretically, the accumulation of a series of "unhealthy" or unsuccessful coping responses should produce mental or physical illness. The literature on alcoholism identifies multiple stressors that impact on individuals in a family with an alcoholic member. These stressors may include marital conflicts, health problems, and loss of job (Cahalan and Cisin, 1976; Ablon, 1976). Research on both stress and alcoholism has shown that each of these factors can lead to problems in adjustment, and that each may be precursors to later alcoholism. (Blane and Hewitt, 1979; Jacob, Favorini, Meisel and Anderson, 1978). Consequently, it becomes important to examine life stressors and their effects as possible consequences of, and/or etiological factors for heavier alcohol consumption and for its chronic outcome - alcoholism. The stress literature has also shown that there are a variety of factors that buffer or mediate the effects of stress. As a result it also becomes important to study these elements and their interaction with drinking. The mediators to be examined in this study are interpersonal methods of coping, in particular, the nature of the elements that compose social support networks. This coping area was selected because many of the problems in families and marriages that include problem drinkers are interpersonal in nature, and because many of these families become more and more isolated as drinking problems worsen. As a result, it is plausable that these families would be less able to use social support networks as buffers to stress. The literature review to follow will attempt to draw together work from the areas of stress, social support networks, and the characteristics of alcoholics, including their marital and family relationships. The review focuses predominantly on alcoholic drinkers vs. non- alcoholic drinkers as this is the primary emphasis of the literature. In addition, a special focus is on the interactional aspects of drinking, drinking problems and drinking behavior in these families. The importance of this interactional factor in families with alcoholic drinking will become even more apparent as we proceed. In order to give some organization to the review, some explanation of the elements that are involved in drinking behaviors and drinking problems and in social support networks is necessary. When the literature refers to alcoholic drinking, the individual is typically drinking enough alcohol and having enough problems related 4 to his drinking such that he reaches the formal diagnostic criterion for alcoholism.1 This sort of drinking would be at the high end of a continuum of alcohol consumption, that varies from alcoholic levels of consumption all the way to abstention. A parallel continuum would be one of high density drinking to low density drinking. Density, in this sense, refers to a concept that would combine alcohol consumption, drinking related problems, and the individual's prior history of drinking. It is at this point where the stage of drinking becomes important. Stage of drinking, i.e. history of alcohol use, ages of first and heaviest use, etc., potentially is related to the types of problems experienced by the individual, and the reactions of families and friends to the individual. Stage of drinking would theoretically be expected to interact with stage of family development, which in turn would also have an effect upon 1 There are a variety of diagnostic critera used to evaluate drinking and its related problems, in order to arrive at a diagnosis of alcoholism (for example, the Research Diagnostic Criteria, DSM-III criteria, Feighner Criteria, Goodwin Criteria). For example, the Feighner criteria break down drinking related problems into four categories, physical problems, loss of control of drinking, work and legal difficulties, and concern over one's own drinking, or concern of family or friends. A definite diagnosis of alcoholism is made if an individual has problems in three or more of these areas. A probable diagnosis of alcoholism is made if problems only occur n two of these areas. problems experienced by the individual. For example, some- one with older children may have fewer parental demands since the offspring are more independent. Stages of drinking and family development are typically not regularly controlled in prior research, and frequently are not ever mentioned. The present study rectifies this problem by focusing on young families with preschool age children; these families have parents who exhibit a wide range of drinking behaviors, but who especially occupy the higher end of the alcohol consumption and alcohol problems spectrum. Prior history of drinking is also to be taken into account, along with the above variables. With regard to social support, this area is also composed of a variety of elements. These include the kinds of people giving support, the types of support given (e.g. instrumental aid, emotional support, information), the quality and frequency of support, and the density of the support network, i.e. how well people in the network know each other. Again, all of the above variables will be taken into account in the present study. These structures of social support will also be reviewed in greater depth in the following literature review. A Review 2; the Relationship 9; Alcoholism, Stress, and Adjustment Research on alcoholics and their families has been plagued with methodological problems: few matched control groups, strong reliance on retrospective or self-report data, samples that include alcoholics at different stages in the unfolding process, children of different ages. They also tend to focus on male alcoholics much more frequently than females, creating a one sided view of drinking and its effects. These problems have contributed to the varied and often contradictory results in this area. However, there are some findings that consistently reappear and are presented below. Behavior pf the Alcoholic Husband and the Nonalcoholic Wife The first series of studies to be reviewed here focused on the individual's behavior, specifically, the behavior of a male alcoholic and his nonalcoholic spouse. The early research in this area concentrated on personality structures and had a psychodynamic approach. Women were described as having disturbed personalities; they had unconscious sexual fears, excessive dependency conflicts and poorly controlled aggressive impulses. Their behavior was guided by their personality and these women fulfilled their needs by marrying a male who was weakened and dependent due to his alcoholism (Futterman, 1953; McCrady, 1982). However, as additional research was done, the personality theory began losing support. The direct role alcoholism had in shaping behavior was considered. Jackson (1962), had this in mind when she began studying families involved with the Al-Anon program. In Jackson's theory, families were responding to the stress produced by living with an alcoholic spouse. Denial, hostility, social isolation, and taking over the alcoholic's responsibilities were all described as wives' reactions to husbands' drinking excessively. Joan Ablon continued work with Al-Anon families and continued viewing drinking behavior as a factor in an individual's behavior. That is, Ablon (1976), studied a group of Al-Anon wives and observed five reasonably separate types of behavior in the marriages: family protectiveness, withdrawal within the marriage, attacking others, acting out, and safeguarding family interests. She noted differences in wives' behavior vis a vis the marriage while the individual characteristics of the wives were similar. The individual differences in behavior appeared to be in response to the husbands' drinking behavior with some men just beginning to drink excessively and other men simply continuing a long history of abusive drinking. Wives coped differently depending upon their husbands' drinking pattern. Ablon also described characteristics of the alcoholic fathers in this study. They were unable to take appropriate responsibility within the family. They lacked self-discipline and they were often overly dependent. Many of the alcoholic fathers were preoccupied with themselves and suffered from a sense of inadequacy. Negative attitudes toward authority as well as unrealistic and immature approaches to things were noted. Finally, these men showed limited interests and related to other people in a superficial manner (Albon, 1976). Ablon then began to consider that these drinking related behaviors might also influence the nonalcoholic spouse's behavior. A caution must be added here. These researchers have listed the above behaviors as characteristic of the alcoholic and his spouse, but such behaviors are also seen in other marriages. That is, the alcoholism may be an added complication, or stress, but it is not necessarily the sole influence on the behavior of an alcoholic and his spouse (Jacob et al., 1978). Whatever the cause, individual behaviors do influence marital and family interactions. This raises questions about the interaction of alcoholism, stress, and behavior. Does alcoholism cause more or different kinds of stress in families, or does it influence how people react to already present stressors and modify their behavior? More recent research has begun to focus more on these questions along with exploring the interactive nature of behavior in alcoholic marriages. Interactions Between An Alcoholic Husband and His 520L158: The literature on the marital relationship between an alcoholic and his wife indicates that this interaction is full of conflict, poor communication and poor resolution of the conflict. Alcoholic couples themselves report more frequent disagreements and quarrels as compared to normal population controls. These quarrels are characterized by physical or verbal abuse, silence, walking out, and moodiness (Ablon, 1976). The rate of marital violence related to alcohol use varies across studies, but when violent husbands with histories of alcohol abuse were researched, alcohol was involved in 52-74% of the cases of wife battery (Morgan, 1982). In addition, the separation and divorce rate in alcoholic families is four to eight times higher than in the normal population (McCrady, 1982). In one sample of predominantly young male alcoholics, ages 21-40, half of the marriages had been dissolved (Tahka, 1966). These statistics give clear indication that marriages containing individuals with 10 severe drinking problems are high in interpersonal conflict. Regarding more specific interactions, there is a high frequency of hostile or coercive verbal interaction and low rates of friendly acts or expressions in alcoholic marriages. Relationship-relevant messages from the alcoholic and total verbal output are also decreased in frequency (McCrady, 1982). Any attempt to settle conflicts meets with little success because of the rigid, competitive style of communication these couples have. According to Gorad's research (1971), each spouse is sending "one-up" messages to the other and tries to gain control of the situation. The alcoholic commonly uses a "responsibility-avoiding" style in an attempt to obtain control through passive, dependent-appearing ways. 0n the other hand, the wife uses an opposite, "responsibility- accepting" style to gain control with action and independence. With this competitive but complimentary manner of communicating, the alcoholic couple have difficulty in achieving joint goals (Gorad, 1971). Essentially, there is a lack of intimate positive exchanges, which also invades the alcoholic couple's physical and sexual relationship (Chiles, Stauss, and Benjamin, 1980). Further evidence for this conflictual type of marriage comes from the children of these marriages, who also see their parents in conflict and rate 11 their families significantly lower in harmony than children of nonalcoholics (Jacob et al., 1978). A prospective study done by Hore (1971a, b), attempted to explore antecedants to drinking in the marital relationship and clearly illustrated the role of interpersonal stress in drinking behavior. This research looked at the relationship between different types of environmental stressors and relapses from sobriety. Via interviews, Hore identified 52 stressful life events among 14 alcoholics (ages 26-60), over a 6-month period. About one-third of these stressful events were disturbances in interpersonal relationships, such as quarrels with spouses or lovers. Hore reported that the correlation between the number of stressful events and the frequency of relapse was not significant. However, further examination of the frequency of relapse following each type of stress revealed 16 interpersonal stress events among the 14 subjects over the 6 months. 100% of these interpersonal stresses were followed by a drinking relapse within 2 weeks, whereas the percentage of relapses following the other types of stresses was considerably lower. This study clearly identifies interpersonal stresses as a predictable antecedent to drinking. It can be hypothesized that if this drinking and conflict continues, each spouse and other family members will begin to feel more and more isolated. The alcoholic 12 gradually relinquishes responsibility as a breadwinner, supportive spouse and parent (Kaufman and Pattison, 1980). The nonalcoholic parent may feel abandoned and overwhelmed by new responsibilities (Royce, 1981). The children in these families may turn to drinking as a way to cope. In fact, studies of children and adolescents who are heavy drinkers continue to show that they come from conflictual, chaotic families (Zucker, 1976). Specifically, adolescents who are heavy drinkers describe their parents as follows: both parents are also heavy drinkers, they are more cynical and antisocial, and their child rearing is marked with open rejection or emotional distance and less parental controls. In addition, disciplinary techniques center around deprivation of privileges and property, plus social isolation (Zucker, 1976). The significance of this kind of environment is clear as Zucker goes on to describe two basic functions of the family as a group. The first is group maintenance through affectional interactions within the family. The second is a more task oriented function that involves the socialization of offspring. This is accomplished via the parent reward structure, and the parents modelling of alternative ways of behaving. Zucker then reviews these functions in families with adolescent drinkers and finds that both of these interactional tasks are carried out 13 poorly. Group maintenance breaks down because of increased parental absence, high levels of family conflict, and the evidence of open rejection between spouses and between parents and children. In the area of socialization tasks, children see the reward structure as containing little praise or positive reinforcement, along with arbitrary discipline and many inconsistencies. Modelling fares no better with parents exhibiting alienation, cynicism, and a difficulty in accepting society's rules. Also, if a parent is not held responsible for his behavior while drinking, the children may adopt this same drinking behavior. In conclusion, appropriate, supportive interactions appear to deteriorate in the family with an alcoholic. Members do not experience the supportive interaction that is believed to be necessary for healthy development. The following section further illustrates the maladaptive outcome of interactions in alcoholic couples, particularly as it is related to the alcoholic's actual drinking behavior. Interaction and Adaption in Alcoholic Couples Although drinking related problems, such as strained family interaction, job loss, and legal difficulties, contribute to the disorder in the alcoholic family and 14 require adaption, (Blane and Hewitt, 1977: Jacob et al., 1978; Zucker, 1976), the actual drinking behavior of a family member is also an event that requires adaption. Interactions and behaviors of family members can be expected to change when the alcoholic is drinking or not drinking. For example, some studies suggest that when drinking, couples produce a higher proportion of interactionally relevant messages, exhibit more positive affect, and demonstrate more agreement, than when they aren't drinking. The alcoholic also increases his or her amount of verbal output and assertive and aggressive behavior when drinking (Steinglass, 1982; McCrady, 1982). In other words, the behavior that results from drinking is not always undesirable. When amount of communication, positive affect, and assertiveness is increased, drinking is reinforced. An earlier study be Steinglass (1981) also showed that some families seem to adapt quite well to the changing phases and behavior of the alcoholic. However, in this study, adaption only enabled the family to remain stable and was not necessarily positive. Specifically, Steinglass examined the interaction patterns of alcoholic families in dry (non-drinking), wet (drinking), and transitional phases. His findings suggest that in the stable wet phase, the family members tend to widely disperse throughout the house and interact only when they 15 intend to talk to someone for a purposeful reason. The verbal interactions that occur vary only moderatly in content, purpose, and affective level. On the other hand, stable dry families have high content variability in their verbal interactions and maintain a moderate amount of distance regulation. Therefore, these families experience higher rates of decision-making and greater affective display of both a positive and a negative 2 nature. When looking at families in transition, a third pattern appears. These families maintain the closest physical distance to each other and have the narrowest range of verbal communication. Steinglass (1980), describes this transitional family pattern as one of "huddling together for warmth and protection," (p. 13). Another study focused specifically on problem-solving behavior in wet and dry alcoholic families (Steinglass, 1979). The results were similar to those cited earlier. 2 This finding is opposite to the work cited earlier by McCrady (1982) and Steinglass (1982). Steinglass attempts to deal with these differences in his life history model, hypothesizing that interaction patterns will also change with life stage. For example, in his 1981 study, Steinglass also noted that the longer spouses has been married, the less their families interact, regardless of drinking stage. Therefor, drinking may be related to increases in positive affect in younger families, in part, because these families interact more overall than older families. As a result, life stage is another factor to control for in the attempt to understand interaction in alcoholic families. 16 Wet phase family members tended to behave independently when solving problems and thereby maintained their problem-solving effectiveness. By contrast, dry phase families took a uniform, interactive approach to the problems in a way that might indicate family solidarity. In sum, one way family members survive the variability and chaos in the alcoholic family is to change their patterns of behavior depending upon the current drinking phase of the alcoholic, thereby creating a "stable" family situation. Social support within the family may strengthen in transitional and dry stages and weaken in wet phases, while external social social support appears to be uniformally weak in these families regardless of drinking stage. While these adaptations promote family stability, they do not necessarily promote health. Another, methodologically rigorous study of interaction in alcoholic families and changes in behavior vis a vis drinking, was completed by 0rford, Nicholls, Oppenheimer, Egert, and Hensman (1975). 0rford's study involved 100 wives of men referred for outpatient treatment and suspected of having an active drinking problem. A factor analysis of the wives' responses to a "Coping with Drinking" questionnaire identifed ten high frequency coping factors. These factors were discord, avoidance, indulgence, competition, antidrink, assertion, l7 sexual withdrawal, fearful withdrawal, taking special action (e.g. attending Al-Anon), marital breakdown. (All of these strategies are interactional in nature with only one (Al-Anon), clearly leading to increased external social support.) Each wife was found to use a variety of these factors with all ten coping factors negatively correlated with a positive drinking outcome. However, all ten factors were positively correlated with the degree of violence and aggression wives reported experiencing at the hands of their husbands. Significant positive correlations were also found between the wife's neuroticism and both discord and assertion. In other words, coping behavior of wives in a state of "husband's drinking" is fairly poor, i.e., ineffective, in dealing with stress, thereby leading to emotional symptoms. This continues to lend support to the characterization of the alcoholic family as one in a chaotic state, with increased levels of hostility and marital disruption in the home. A study by James and Goldman (1971), also found that wives' method of, and intensity of coping changed (increased) as the husband's stage of drinking changed from social drinking to excessive drinking to alcoholismic drinking. That is, as husbands' drinking became more severe, wives tried a greater variety of coping strategies, with withdrawal from and protection of the husband being the predominant strategies. These are both 18 isolative behaviors, the first decreased social support within the family and the second decreased social support from outside the family. Schaffer and Tyler (1979) continued 0rford's line of research and administered the "Coping with Drinking" questionnaire to Al-Anon members. These authors also did a factor analysis of this scale and found some overlap and some differences between their factors and those of the 0rford group. Their nine categories of coping responses included confrontation-discord, destructive reaction, avoidance, spouse-specific reaction, seeking outside help, inaction, fearful action, taking care of husband, and financial action. Schaffer and Tyler then placed their factors in a multiple regression analysis with three criterion variables: time since last drink, AA attendance, percent of time sober in the last year. Results showed that wives who took specific action against the alcohol, not their husband, and did not nurture or care for their husband, had husbands who had better drinking outcome providing they were in the group of husbands who scored more poorly on the criterion variables. For the group of wives whose husbands scored comparitively higher on the criterion variables, (i.e. had less of a drinking problem), pretending all was well, or reacting fearfully, were more strongly associated with positive drinking outcome. Action directly against the 19 husband was negatively associated with good outcome. (Again, it is interesting to note that "seeking outside help" was not a frequently used strategy so that its effectiveness in influencing the husbands' drinking or the wives' emotional health is impossible to determine.) The authors had difficulty explaining these results. Their tentative conclusion was that engaging, interactive responses that were not directly aimed at the husband led to more positive outcome for the husband's drinking than responses that show disengagement from the husband. Again, wives style of interaction varied as husbands' drinking behavior varied. The next study to be cited in this section does not focus on identified alcoholics or alcoholism but simply looks at the interaction between drinking behavior, adaption, and stress in the form of major life events (Cronkite and Moos, 1984). Subjects included 267 married couples, randomly selected from the San Fransisco Bay area, who were assessed at two points in time concerning major life events, depression, psychosomatic symptoms, alcohol consumption (ounces per day), self-esteem, social support, and coping style. The results showed that alcohol consumption was related to coping, stress, and disorder for both men and women in a variety of ways. First, increased drinking for one spouse is related to increased drinking in the other spouse. In regard to 20 men's coping, men with a greater alcohol intake than the group mean, along with greater levels of depression and physical symptoms, were more likely to use avoidance coping, implying little use of social support. Whereas, men with wives who drank more than the group mean were more likely to use approach kinds of coping. For women, those who drank more and had more current negative life events were also more likely to use avoidance coping behaviors. However, this changed if there was a greater amount of family support: approach coping was used. For both men and women, avoidance coping at time one was associated with increases in alcohol consumption at time two. This pattern points to a destructive feedback loop between alcohol consumption and avoidance styles of coping. The study by Cronkite and Moos (1984), also shows a direct relationship between alcohol consumption and social support. Specifically, the husband's prior alcohol consumption was associated with less positive family support. This makes intuitive sense when combined with the finding that alcohol consumption is related to avoidance coping. Avoidance versus approach or engaging coping behaviors would lead to isolation from family and friends. Studies specifically focusing on social support and diagnosed alcoholics and drug abusers support this 21 connection between substance abuse and isolation. Rhoads' research (1983) on heroin abusers showed that increases in social support were significantly correlated with continuation in treatment, decrease in heroin and other drug use, and decreases in anxiety and depression. Another study, by Joan Jackson (1954), compared male alcoholics' number of attempts to stop drinking with and without support of an institutional program and/or A.A.. 0f the 26 men surveyed who successfully stopped drinking, 23 had received some kind of support. A third study by Quinn (cited in Cobb, 1976; personal communication), found that escapist drinking but no other forms of drinking was significantly higher among men who experience high job stress and are not supported by their supervisors. That is, heavy drinking and alcoholism are related to weak social support networks. In sum, the family with an alcoholic member can be characterized as a family under stress. Much of this stress is interactional in nature and ranges from the alcoholic taking a negative attitude toward authority (Ablon, 1976), spouse abuse (Morgan, 1982), divorce (McCrady, 1982), spouses sending "one-up" messages to each other (Gorad, 1971), to parents open rejection of their children and use of harsh discipline techniques (Zucker, 1976). These conflictual families have also been described as attempting to react to their situations in 22 numerous ways. Family members may distance themselves from each other during an alcoholic's stable wet phases of drinking or draw closer together when the alcoholic is moving between a dry and wet phase of drinking (Steinglass, 1981). Other coping techniques by wives of alcoholics include indulgence, competition, assertion and withdrawal (0rford et al., 1975). Seeking outside help is rarely used although, social support appears to have positive affects on drinking and adaption (Jackson, 1954; Cobb, 1976). Logically, some of these reactions are more or less successful in leading to adjustment. Discord and assertion are related to wives' neuroticism (0rford et al., 1975). Indulgence or protection of the husband is related to increases in the husband's drinking (Schaffer and Tyler, 1979). Avoidence is related to greater levels of depression and physical symptoms (Cronkite and Moos, 1984). In other words, the model of an alcoholic family as a family under stress is a bleak one. The system is rigid (Steinglass, 1982) and conflictual (Morgan, 1982; McCrady, 1982). The chance for some type of maladjustment developing out of this system appears high and can include abuse, divorce, depression, increased drinking, and the development of alcoholic offspring (Zucker, 1976). Although the preponderance of destructive interactions associated with drinking is obvious even from these few studies, more detailed research is needed to 23 understand the factors contributing to this process. Ideally, such research would be longitudinal in nature, sampling stress and adjustment in different situations and at different stages in the family's development, including the different phases of drinking. In addition, the effects of stress and drinking behavior, and their interaction, on both husbands and wives adjustment should be examined. The current study examines the relationship between drinking behavior, daily hassles, and elements of social support, and explores the relationship of these factors to the emotional and behavioral adjustment of marital partners. The interaction between husbands' and wives' adjustment and stress is also examined. Again, a special focus of this work is the exploration of this interplay in families who are quite high in levels of alcohol consumption, and among which many would be characterized as "alcoholic". The current level of drinking in these families as well as their lifetime level of alcohol involvement is also taken into account. This study is called for as a clear next step, based on the show of evidence that social interactions are problematic in alcoholic and heavy drinking families, and that they are related to life stresses, increased drinking, and emotional disorders. Review 2; Stress Literature The following sections well go into more detail on stress, symptomatology, and social support as they relate to the general population and populations experiencing psychopathology other than alcoholism. This review is included in order to validate the importance of looking at these variables in relation to drinking behavior, and to highlight the important factors and difficulties in this area of research. The Dynamics 2; Stress Again, a life stressor, or stress, temporarily upsets the balance of an individual's life and he/she needs to exhibit some kind of coping response in order to reestablish that equilibrium. Theoretically, an "unhealthy" or unsuccessful coping response will result in some kind of mental or physical symptom (Rabkin and Streuning, 1976). The effects of life stress on behavior and ongoing functioning can take many forms. A quick survey of the Journal 9; Human Stress shows a connection between stress and return to heroin use, hypertension, infectious disease, depression, cancer, heart disease, accidents, athletic injuries, diabetes, and ulcers, to 24 25 name a few. Obviously these connections are going to vary for children and adults, although stress has an impact for both groups. Life Stress and Symptomatology in Adults One of the most thorough studies of life stress and symptomatology was done by Brown and Harris in 1978. They set out to study the etiology of depression in women and analyzed the effects of different types of life events in great detail. Their samples included depressed patients and general population controls. Only two types events occured at a higher rate in the depressed sample. The first type of event involved moderate or marked long term threat to the individual where "long term threat" implies that the consequence of the event follows a week or longer after its occurence. The second type of life event that occured more frequently in the depressed patients involved the isolation of the individual or a difficulty with that individual's intimate relationship. These "severe" events usually involved some kind of loss or disappointment and were most likely to cause depression. Brown et al. also examined the additivity of events, their long and short term effects, and their causal vs. triggering potential. The results are as follows: in 26 comparing number of events experienced by depressed patients and normals, 60% of depressed patients and 75% of normals experienced one severe event in a 38 week period. Twenty-one percent and 8%, respectively, experienced three or more severe events. The difference between depressed patients and normals, when comparing proportions experiencing three or more events, was significant Significantly more depressed patients also experienced three or more nonsevere events compared to controls. Therefore, an increased number of events in a certain time period is more likely to produce maladjustment. By examining the number of weeks that an event occurred before depression onset, one can determine the short or long term effect of that event. Results showed that 30% of the severe events occured within nine weeks of onset while 4.6% of these events occured 37-45 weeks before onset. That is, most severe events quickly led to depression. This usually happened within nine weeks and in almost all instances, within six months. The less severe events did not appear to have the same long term effects. This phenomenon is described in more detail below. Finally,'Brown and Harris wondered if their severe events were having a triggering as compared to a formative effect on depression. To determine this, the researchers computed a Time Brought Forward (be) equation which 27 estimates how long it would have been before the onset of depression if an event had not occurred. The longer this time, the more likely the event played a formative role because the depression was hastened to such a large degree. The results of this analysis showed severe events playing a formative role in depression, with the average be equal to 2.13 years. Minor events played a triggering role, with a be averaging 10 weeks. The role of events in the onset of schizophrenia differed somewhat. That is, only events in the three week period right before onset were capable of influencing schizophrenia. Thus, events play a triggering role in this disorder. We may conclude that not only do different types of events play different roles in the onset of disorder, similar events may lead to different disorders (depression vs. schizophrenia), based on another moderator variable. Minor events clearly function differently than do severe events, but there may be a connection between these two types of circumstances. Specifically, a minor event may cause a woman to reassess the impact of an earlier major event. This may lead to hopelessness and depression. For example, going to a friend's husband's funeral may trigger a repressed memory of her own husband who died five years earlier (Brown and Harris, 1978). All of the above findings complicate the relationship between life stress and illness. First, an event may or 28 may not be stressful to a person, depending on his individual characteristics and those of his environment. Second, events themselves may vary in effect depending on their severity of threat, their term of effect, and whether or not they occur alone or in close proximity to other events. These conclusions are based on the Brown and Harris research (1978). This study also points out the importance of looking at the effects of both major life events and daily hassles on people's lives. A possible model for these relationships is outlined in Figure l. The following studies help to clarify the complex relationship between stress and symptomatology. More characteristics of stress are examined as well as other types of maladjustment. Like Brown and Harris, E. S. Paykel also conducted a study on depression, but focused completely on the specific types of life events that preceded the depression (Paykel, 1974). He found that depressives reported three times as many major life events as did controls. Suicide attemptors reported more events than depressives who in turn reported more events than did schizophrenics. There were also significant differences among these groups in the category or type of event they experienced. That is, depressives had significantly more exits from their social field than did controls. Depressives also experienced Life Event Cognitive Factors: Assessment of the Situation, Adult's Denial Coping Individual's with the Characteristics: Situation Age T Stage of Development Personality ‘ Past History of the Social Coping Skills Individual Support, Other Material Resources Consequences for the Individual Figure 1; Model of the Life Stress-Illness Relationship 29 30 significantly more negative events, health, and law- related events. In comparing both experimental groups, depressives reported experiencing more exits, negative events, and finance and health related events compared to the schizophrenic group. According to Paykel, These findings do give some indication of specific relationships. Only certain kinds of events precede depression. It is not just a question of magnitude of life change: the direction of the change and its desirability are also important. However, the link between event and disorder is far from exact. (Paykel, 1974; p. 138) Rahe's (1979) study comparing life stress in schizophrenics and neurotics, and found similar results. Both groups reported an increase in the number of life events prior to symptomatic onset, with neurotics reporting twice as many life changes during the three months prior to onset. Sarason, Johnson, and Siegel (1979) expanded the types of maladjustment being examined but limited the type of events they focused on. They looked at the relationship between adjustment and positive, negative, and total amount of change. Results showed one significant correlation with positive change: positive change was significantly related to extraverted expression. Significant correlations with negative change included greater trait and state anxiety, social nonconformity, neuroticism, current depression, and an increased external locus of control. The only significant 31 correlations for total change scores were with the two anxiety measures (Sarason et al., 1978). Many more studies relating stress and disorder have been done in recent years; however, they have aimed at clarifying this relationship more than confirming it. For this reason, the more current research is reviewed in later sections. At this point, the most general conclusion to draw from all of these studies is that different types of stressors influence different types of maladjustment. The specific relationships are not clear and need to be focused on more directly. One such relationship may be the interaction between an individual's perception of stress and pathology. The fairly consistent relationship between an external locus of control and increased pathology may indicate that individuals who see stressful events as out of their control may be more vulnerable than those individuals who believe they have control over those same events. This external locus of control is also found in alcoholic families (reviewed later), making it another area to explore when comparing alcoholic and normal control families. Life Stress and its Measurement in Adults The research on life stress in adulthood is quite extensive; yet, the type of research varies. In the 32 adult literature there is much controversy on how to measure life stress and fewer studies, proportionally, on the specific psychological effects of life stress. The most frequently cited measure of life events is the Social Readjustment Rating Scale (SRRS), by Holmes and Rahe (1967). Holmes and Rahe had 194 subjects rate 43 life events on the amount of readjustment required for each event. Readjustment scores were obtained by taking the mean rating for each event and dividing by ten. Because this scale is non-interval and has a skewed distribution of arithmetic means, the geometric mean was considered the best measure of central tendency. This measure discounts the extreme score but takes into account the distribution of scores (Masuda and Holmes, 1967). There is strong agreement on the rank order of life events according to their readjustment score (the readjustment score - the number of life change units for an event), and the magnitudes of the mean readjustment scores for these events, as evidenced by high Pearson product moment correlations of these scores between various groups in the above sample of 194 subjects. Correlations between groups divided by sex, marital status, age, generation, education, social class, and religion were all above .90. The lowest correlation was .82 between whites and blacks. All correlations were statistically significant (Holmes and Rahe, 1967). 33 A second study by Casey, Masuda, and Holmes (1967), examined the reliability of recall of life events. Fifty-four subjects were asked, at two points in time, to complete the SRRS for the years 1957, 1960, and 1963. (Time one, in 1964, and time two, in 1965, occured at nine month intervals.) Each time, total life change scores were computed by summing the life change units of each major life event reported. Scores that differed by more than 40 life change units between time one and time two were considered discrepant. This meant that subjects had reported significantly more or less life events as occuring at time two as compared to time one. (Individual items were also examined for consistency in reporting at these two points.) Correlations between time one and time two, for the recall of events in 1957, 1960, and 1963, were .669, .638, and .744 respectively. Clearly, the more recent years were recalled more reliably: though all three correlations were significant. However, despite the statistical significance, there were discrepancies in reporting. Thirty-two of 44 subjects had no discrepancies in their reports of major life events between two of the three years, and had no consistent change in direction (i.e. more or less life events), for the discrepant years. More specifically, the items with higher readjustment ratings were recalled more 34 consistently. Also, items that had qualifiers such as major/minor and more/less in the question, were prone to subjective interpretation and may have caused change in recall over time. In fact, there was a significant difference in recall consistency between items with qualifiers and those without. However, the saliency of the items and not the presence of qualifying words affected the consistency of recall the most. Items with double questions were also recalled less consistently. Finally, the amount of life stress at the time of recall did not influence the magnitude of scores recalled (Casey et al., 1967). In sum, there is a strong reliability in the recall of life events. This conclusion is given additional support by Brown and Harris' study of life events. Their results showed that the reported rate of events did not significantly drop off in the year before the events interview. That is, there was not a significant increase in the number of events reported in the more recent months (Brown and Harris, 1978). Though recall over time is fairly consistent, one identified bias in the initial reporting of events is the personal characteristic of denial. Those higher on the denial measure did report fewer life events that those low on denial (Cobb, 1974). The main controversy in measuring life stress centers around the question, what is stressful about an event? 35 Holmes et al., (1967), originally believed the critical element was the amount of readjustment needed after an event occured. Others believed it was the desirability or undesirability of the change, whether or not the event involved affiliative needs, if the event had short or long term effects, the degree of threat involved, or the number of events experienced in a certain period of time (Suls, 1981; Rahe, 1979; Cobb, 1974; Burchfield et al., 1982: Sarason et al., 1978; Brown and Harris, 1978). These researchers have objectively rated the desirability and impact of certain life events; yet, there are those that believe it is primarily the individual's subjective interpretation of an event that leads to its stressfullness. Fontana, Hughes, Marcus, and Dowds, (1979), have based their research on the above hypothesis and measured individual perceptions of 1) the desirability or undesirability of an event, 2) the amount of readjustment required for an event, 3) the degree of anticipation prior to an event, and 4) the amount of control over an event. Their results show that the subjectively desirable events were also those that required little adjustment, could be anticipated, and fell under the control of the individual. That is, the four types of evaluations were linked in some way. Also, each of the four types of subjective evaluations were significantly correlated with 36 different areas of psychosocial adjustment. For example, a high level of perceived undesirability of life events was associated with a high level of psychological disturbance. The specific significant relationships Fontana et al. found are presented in Table 1. These results suggest that certain types of subjective evaluations predict poor psychosocial adjustment in certain domains. All of the studies reviewed in this section show a significant relationship between their own measures of stress and various types of illness. However, Rahe's comment in 1979 still applies now: "The superior utility of any life change scaling method over a simple counting of the number of events, per unit time, has yet to be demonstrated convincingly." (Rahe, 1979; p. 4). This statement is not meant to negate the value of these different classifications, but implies that they may be better understood as moderator variables in the stress- illness relationship. In sum, the most common instrument in the field for measuring life stressors has been the Holmes and Rahe Social Readjustment Rating Scale (SRRS, 1976). This self- report instrument measures the frequency of both positive and negative major life events. Over the years, the SRRS has received a great deal of criticism and modification. Despite Sarason et al.'s (1978) addition of subjective Table 1 Significant Correlations Between Four Types 2; Subjective Evaluations g; Life Events and Psychosocial Adjustment Perceived 1 Subjective Evaluation g; Life Events Perceived Perceived Perceived Desirability Adjustment Anticipation Control Psychosocial Adjustment Extent of: Psychological Disturbance Interpersonal Involvement Alcohol Abuse Outside Social Participation Employment 1 -.35** .33** .29* .48** -.26* -.28* -.41** Subjective evaluations involve the individual's perception of the desirability of the event, the amount of readjustment it would require, how much the event was anticipated, and the degree of control he/she felt over the event. Note: Adopted from Fontana et al., 1979. Note: These are Pearson r's: *p .10, **p .05 37 38 weightings of life events and other modifications, and the plethora of research on this instrument, the "proper" way of measuring life stresses, particularly as a predictor of physical and emotional symptomatology continues to be debated (Kessler and Eaton, 1985). Partly based on these inconclusive results, Kanner, Coyne, Schaefer, and Lazarus, (1981), felt that measurement of major life events was not helpful in predicting health outcomes. These authors felt that events 9; issues that required day 39 day attention were more likely to affect health. Based on their conceptualization, Kanner and his colleagues developed the Hassles and Uplifts Scales - measuring the daily transactions with the environment. Kanner et al. consider this a measure of daily stress; others have labeled this instrument as a measure of chronic ongoing stress (Kessler et al., 1985). Research on this scale has examined the relationship of daily stress to major life events, and has compared the predictive power of both of these measures for health outcomes. In relation to life events, hassle intensity scores were not significantly related to life events while frequency of hassles were. More specifically, life events for men were positively correlated to hassles and negatively related to uplifts. Life events for women were positively related to both hassles and uplifts. 39 The next step in the research was to determine if life events or hassles were better predictors of symptoms. A step-wise regression analysis showed that hassles were superior to life events in predicting symptomatology. In addition, when comparing hassles and uplifts, hassles were again better at prediction of symptoms for both men and women. A second study by DeLongis, Coyne, Dakof, Folkman and Lazarus, (1982), compared the relationships of major life events and daily hassles to somatic health. Again, hassles and major life events shared similar portions of the variance in predicting health status but when the effects of the life events were removed, hassles and health stayed significantly related (r=.27, p<.01 and r=.35, p<.01). Critiques of this instrument come from Dohrenwend, Dohrenwend, Dodson, and Shrout, (1984). These researchers had 371 clinical psychologists rate each of the hassles items as likely to be symptoms of psychological disorder. They pointed out that all but 12 of the 117 items on the hassles scale received ratings of "almost certainly a symptom," "likely to be a symptom," or "as likely as not to be a symptom of psychological disorder." The authors also noted that the undesirable events on the Holmes and Rahe scale have similar problems. Dohrenwend et al. went on to write that, "because undesirable events tend to show 40 the strongest positive correlations with psychological distress, this confound is a serious matter," (Dohrenwend et al., 1984; p. 228). In addition, it was noted that the life events questionnaire used in DeLongis et al.'s and Manner et al.'s studies did not have this confounding problem; therefore, it will show less correlation with psychological symptoms than the hassles scale. Dohrenwend noted that "... it seems premature to conclude as do Kanner and his colleagues that daily hassles proved a more direct and broader estimate of stress in life than major events," (Dohrenwend et al., 1984; p. 228). In the studies cited above by Kanner at al. and DeLongis et al., the authors anticipated this problem of confounding and explored it's possible effect on their data. They found five somatic health related items and fifteen overlap items with the Hopkins Symptom Checklist on the Hassles Scale. They removed these items and found that the new and old scores still had a correlation of .99. With regard to this issue of confounding, Lazarus (1984), acknowledges not having thought through "all of the implications of format and wording." He goes on to state that the aim had been "to cover the broad ground of relatively minor psychological difficulties of living as sensed by the person rather than to create pure and objective stimulus and response categories," (Lazarus, 41 1984; p. 376). In this theoretical article, Lazarus procedes to build his case for the subjectiveness with which individuals mark items as hassles, that hassles are reflections of how problematic the environment is, and that hassles are reflections of the individual's personal structure and coping responses. In sum, Lazarus is arguing that it is this subjectiveness of hassles that makes them stressful, not simply the fact that they occured. He also emphasizes the transactional nature of hassles and health, pointing out that confounding occurs inevitably because of this as well. Additional support for the use of daily hassles as an indicator of stressors, and as a factor in health status, comes from research by Burks and Martin (1985). They attempted to eliminate any items that might be seen as responses to stress (versus stressors), from their scale of "Everyday Problems." Burks and Martin also found stronger correlations between everyday problems and symptoms (r-.42) than between major life events and symptoms (r=.21). A step-wise regression analysis also showed everyday problems as more powerful predictors of total symptoms than life events. It appears than, that the controversy of how to measure stress will continue. Yet, despite its structural flaws, Lazarus and his colleagueas have developed an instrument to measure daily hassles that appears to get at 42 the concept of stress from a different angle. This instrument is based on a theory of stress as threat versus stress as change, and includes personal factors as determinants of stress as well as environmental factors. The Lazarus instrument was chosen for the current study because earlier data from this project using the Holmes and Rahe SRRS and using subjective weightings of major events found no significant differences in stress but did in levels of symptomatology. Past literature shows a relationship between these two variables, therefore we questioned our lack of results along these lines. It could be that our sample experienced similar economic and geographic conditions which may have caused similar major life events. It is hoped that the Hassles Scale will get at stressful situations that may differ in this sample. Hassles may get at more personal characteristics of stress, i.e. stress caused by more individual characteristics versus environmental or national economic conditions. Social Support, Stress, and Disorder There are a great many other potential moderators in the stress-illness relationship, including an individual's locus of control, personality constructs, a variety of demographic variables such as SES and age, cognitive 43 appraisal, etc.. However, the present research is more modest in aim, and focuses specifically on on the relationship of social support networks to stress and illness. Research on social support has increased rapidly in the past ten years, and continues to receive empirical support for its place in the relationship between stress and symptomatology. Research in this area has also become more refined as a result of the quantity of work and interest in social support. Recent studies are more precise about what kinds of social support they measure, the structure of the network, (i.e. size, density, etc.), and the function of the network, (i.e. instrumental, informational, emotional, direct aid), (Brown, 1972). These studies also control confounding factors such as level of education and income, preexisting level of disorder and preexisting level of stress. All of the variables mentioned above have been shown to affect the amount of stress and disorder that an individual may experience, and they should be factors in study design (Broadhead, Kaplan, James, Wagner, Schoenback, Grimson, Heyden, Tibblin, and Gehlback, 1983; Holahon and Moos, 1981; Monroe and Steiner, 1986; Cronkite and Moos, 1984; Monroe et al., 1983). Support for such carefully controlled designs can be found in one study where education, race, prior depression, physical 44 symptoms, and alcohol consumption accounted for six to eleven percent of the variance in stress levels, symptom levels and support levels (Cronkite and Moos, 1984). This work clearly illustrated that disorder is not simply a function of stress and level of social support. Other findings showed alcohol consumption, depression and physical symptoms were fairly stable over time, regardless of stress. And, higher levels of education were associated with higher alcohol consumption for men, while nonwhite ethnicity was associated with lower alcohol consumption for women. In sum, alcohol use and other symptoms can be determined by race, education, gender, and prior levels of functioning. Therefore, research that examines the connection between stress and disorder needs to control for these additional factors. Monroe and Steiner (1986), have also emphasized the importance of careful methodology in research on social support. These authors discuss measurement redundancy, method limitations, and conceptual considerations in their comprehensive review along with the need to control for preexisting disorder in subjects. Regarding this last point, Monroe and Steiner state that the underlying assumption of the majority of research is that we are looking at acute disorder when examining the affects of stress, not chronic disorder. Hypotheses based on acute disorders assume no prior level of disturbance and that 45 the stress and social support measured then on, influence the acute disorder. However, these authors cite past studies that have shown as much as 25% of the population exhibit chronic patterns of disorder over many years. As a result, this field needs studies with longitudinal designs that measure changes in stress, support, and disorder over time. This will help control for the influence of chronic disorder on both stress and current illness. Monroe and Steiner go on to discuss problems in confounding of measures and causal effects. In discussing measurement redundancy, these authors point out that some of the criteria for depression overlap with factors influencing social support, e.g. loss of interest in social activities. In addition, stressful events such as death of a friend, and fights with spouse are also measured as losses in social support. There is an artificial/methodological correlation built into disorder, support, and stress before the chosen sample is even interviewed. Finally, the potential importance of personality and its impact on social support was discussed. The hypothesis was that an individual's personality may determine the kind of social support network he or she has. For example, an introverted person may prefer a small number of special friends instead of a large number 46 of social acquaintances. Personality may also influence utilization of networks, e.g., an individual with strong independence characteristics may be less likely to turn to friends for help. In summary, there are many of methodological concerns in this area of research, yet, even the well controlled studies continue to show links between stress, disorder and support. Cohen and Wills (1985), completed a comprehensive review of this literature through 1983, and found evidence for both main effects of social support on health and buffering effects, with social support acting on the stress which would lower level of symptomatology. In a careful analysis of these studies, Cohen and Wills conclude that it is the degree of an individual's integration into a social network, (i.e. the structural component of social support), that has a main or direct effect on health. An explanation for this may be that belonging to a network maintains feelings of stability, self-worth, and well-being irrespective of stress level. On the other hand, the authors conclude that it is the functional aspects of social support, (i.e. instrumental and emotional support), that buffers stress, thereby reducing its influence on disorder. Based on these conclusions, social support appears to be composed of a variety of factors that function differently and somewhat independently. Therefore, it seems more important to 47 determine which components of social support are helpful in which situations, rather than trying to determine which factor of social support is most important. Because of their conclusions, Cohen and Wills recommend longitudinal research with data collection at a minimum of two points in time, and factorial designs that include at least two levels of stress and social support. This would help to sort out buffering and main effects. A regression analysis with a cross-product term of stress and social support, or an analysis of covariance that partials out confounding variables should also be completed. In addition, they emphasize the need to include different groups and different types of stressors in the design. The studies reviewed below support the above conclusions and begin to illustrate that different aspects of social support have different effects for different individuals on certain disorders. To begin with, research by Cohen and Wills (1985) notes that women derive satisfaction from talking with intimate friends while men obtain satisfaction from companionship activities and instrumental task accomplishment. As a result, women may profit less from social integration than men. For further support, Cohen et al. (1985), quotes studies by Kessler and Essex, 1982, and Fleming, 1982. Their results showed 48 that social integration had a buffering effect on parental strain, but not on economic or homeowner strain, and the availability of confiding relationships had a buffering effect on psychological symptoms but not on physiological symptoms. Unfortunately, every study does not produce results that are quite as clear cut. For example, Wilcox's (1981) study on 320 community residents measured both quantity and quality of social support as well as major life events, different symptoms, and various mood states. He found no direct effects in regard to social support, symptoms or mood states though his results showed a significant positive relationship between life events and symptom level. Findings supported the buffering hypothesis, with higher levels of support correlated with lower symptom levels, and lower levels of support associated with higher symptom levels under high levels of stress. Contrary to Wilcox's findings, Anashensel and Stone (1982), found support for a direct effect of social support on functioning, specifically on depression. Lower levels of support were significantly related to increased levels of depression. Interaction between stress and social support did not improve the prediction of depression. 49 Reviewing the above two studies again illustrates the need to look at discrete/specific types of support and symptomatology instead of using global measures of these factors. Otherwise, we will continue to have "contradictory" results across studies. In actuality, these different results may be confirming that different types of support function differently and have different affects on well-being. A more carefully controlled study was completed by Monroe et a1. (1983). These researchers conducted a prospective study on university students during final exams with pre and post measures of symptoms (including anxiety, physical health, and depression), life events and social support. Two types of support were measured, living at home and number of best friends. After controlling for initial symptom level, only depression was significantly related to negative life events. The only social support measure directly related to symptom level was living at home. However, number of best friends predicted symptom scores on health, depression and anxiety when there was a high level of negative life events. A greater number of friends was significantly related to fewer symptoms. A surprising finding was that a high number of desirable events and a high number of best friends was also associated with higher levels of anxiety. The 50 authors did not attempt to explain this last finding, (although a u-shaped as compared to a linear model of support and disorder might help), but stated that further study was needed to clarify these complex associations. They also noted that findings from their study were limited in generalizability because of the specific population and situation studied. However, this writer believes it is an important study to build on. Further work focusing on the same population in different situations, or the same situation with different populations, will begin to elucidate cross-situational, cross-population relationships for stress and social support. For example, if the same study by Monroe and his colleagues (1983), was done with graduate students instead of undergraduate students, best friends may be associated with lower levels of anxiety; the impact of close friends may change with an individual's development. Moos and his collegues continued to examine the area of social support in two other studies. Both were longitudinal and controlled for a number of demographic factors and preexisting disorders. The first study by Holahan and Moos (1981), focused on 245 married couples randomly selected in the San Fransisco Bay area. Self- report data were collected on negative major life events, depression, psychosomatic symptoms, work, family, and non- family support. Analyses of these data showed that family 51 and work related support were negatively related to depression and physical symptoms, for both men and women (correlations ranged from -.24 to -.47), while negative life events were positively related to both types of disorder in both sexes (correlations ranged from .23 to .30). All of these results are in the predicted directions but it is also interesting to note that the correlations were significant but low. Much of the variance related to disorder is still unaccounted for, even with prior disorder controlled. A second study (Mitchell and Moos, 1984), offered an even more detailed look at the relationship between social support, stress, and depression. This research separated subjects experiencing major vs. minor depression and looked at the impact of major life events and chronic strain. As expected, their results showed some variation in stress and social support as a result of demographic factors. However, once these factors were controlled, a number of other results were apparent. Decreases in strain and increases in positive life events from time one to time two were associated with increases in family support at time two. Initial levels of family support did not predict positive or negative major life events at time two. However, individuals with more friends at time one were more likely to experience decreases in chronic strain at time two. At this point it appears that different 52 kinds of stresses in this study are differentially affected by different kinds of social support. That is, friends may have an impact on the decrease of chronic strain, while increasing family support may decrease strain and increase major positive events. Neither type of support appears to have an impact on major negative life events. Further analysis of the data showed an interaction between different types of stress. Subjects with a higher number of negative events and higher levels of strain had fewer friends at time two, whereas the combination of higher levels of strain and more positive events did not have this effect. In addition, those with few positive events and an increased number of negative events at time two had less family support at time two. Yet, those with many positive and negative events did not experience this decrease in support. In sum, the impact of negative events seems less apparent when they are linked to positive events. Positive events could have a buffering or direct positive effect on support and disorder. Future research should consider sampling or controlling for the number of positive events experienced as well as controlling for demographic factors. It is also clear that despite recent methodological advances in this area of research, there is much more to discover about the impact of social support. Summary Again, the literature reviewed above has pointed out the negative effects of both stress and alcoholism. Life stress is associated with depression, ulcers, heart disease, and drug use. Alcoholism is associated with marital disruption, physical violence, and poor adjustment in children of alcoholics. Certain coping behaviors and social support structures have also been shown to buffer the effects of stress and have a positive impact on adjustment. Other behaviors, such as avoidance and disengagement appear to have a negative impact on adjustment and are clearly present in alcoholic families. Of the stresses and behaviors described above, interactional stresses and interactional coping behaviors seem inexorably linked to difficulties in families with heavy alcohol consumers. In sum, there does seem to be a connection between stress, drinking, social support, and adjustment. However, due to methodological problems in the literature and few studies comparing stress across populations, the exact nature of this relationship is unclear. Figure 2 describes a possible relationship between husbands, wives, stress, social support, and symptomatology. 53 ////;7180cia1 Support Network / [Husbands' Drinking / \[ Hassles / Hassles Wife's Drinking Symptoms —‘ Symptoms Social Support Network Figgre 2;, Model of the Stress-Illness Relationship in a Drinking Couple 54 55 The present study is an examination of differences in the amount and type of stressors experienced by families who vary in the extent to which they are currently involved in problematic alcohol use, but all of whom are at higher than usual risk for the development of alcohol problems among their offspring. Differences in interactional involvement--via the structure of social support networks--and differences in extent of symptoms manifested as a result of such variations in drinking, are also examined in these families. The following hypotheses will be tested; they are generated on the basis of findings from the research reviewed above: Formal Predictions Hassles and Alcohol Involvement 1A: Husbands and wives i3 families where there lg high lifetime alcohol problem involvement will show greater amounts gf hassles than spouses ig families with low lifetime alcohol involvement. lg; The category g; events chosen gg stressful by an individual and his spouse will differ depending 9g Egg individual's drinking pattern. Ag individual who has 3 high 1232; g; lifetime alcohol problem involvement and his spouse will choose more interpersonal hassles gg stressful than will those individuals and spouses with 3 low alcohol involvement. 56 Alcoholic marriages have been characterized by more frequent disagreements, which include physical and verbal abuse (Ablon, 1976). In addition, men in these families change jobs frequently and there are more separations and divorces in this group (Blane and Hewitt, 1977; Jacob et al., 1978). The majority of these events are interpersonal in nature (Hore, 1971a, b). Symptom Occurrence and Alcohol Involvement 2A: Individuals with high lifetime alcohol problem involvement will exhibit greater amounts pf antisocial behavior than individuals with low lifetime alcohol involement. 2B: Spouses pf high lifetime alcohol involved men will experience greater levels pf depression and somatic symptoms than will spouses pf low lifetime alcohol involved men. Withdrawal and aggressive acting out are recognized patterns of behavior in the wife of an alcoholic. The alcoholic husband feels inadequate and also tends to be antisocial (Ablon, 1976: Zucker, 1976). In addition, depression and somatic complaints appear to be common reactions to stressors of an interpersonal nature (Brown and Harris, 1978; Paykel, 1974). 57 Social Support and Alcohol Involvement pi Individuals with pigp lifetime alcohol involvement and their spouses will have weaker social support networks than will individuals with low lifetime alcohol involvement and their spouses. Again, withdrawal, silence, walking out, antisocial behavior, and high levels of conflict have all be identified as present in alcoholic families (Ablon, 1976; Zucker, 1976; Blane and Hewitt, 1977; Jacob et al., 1978). All of these behaviors can be alienating for other individuals. In addition, drug use and alcohol consumption have also been linked more directly with lower amounts of social support (Rhoads, 1983; Cronkite and Moos, 1984). Hassles, Symptoms, Social Support and Alcohol Involvement ii The level pf symptomatology and stress experienced py one member pf the marital couple will pg related pg the level pf stress and symptomatology ip Egg other member. Research has showed that the husband's drinking is related to wives isolative and aggressive behaviors (Ablon, 1976; Steinglass, 1979). Statistics also show that abusive drinking is related to high divorce rates and incidents of family violence (Morgan, 1982). Lastly, alcohol consumption of both husbands and wives has been 58 interactively linked with psychosomatic symptoms, life events, and social support (Cronkite and Moos, 1984). .1 Differences ip lifetime alcohol involvement, social support, and hassles will influence level 22 experienced symptomatology; Certain coping behaviors and strong social support networks are effective in directly influencing the adjustment of the individual and/or buffering the adverse effects of stress (Cohen and Wills, 1985). Poor social support, as predicted in hypothesis 3 and demonstrated as present in alcoholic families (0rford et al., 1975: Schaffer and Tyler, 1979; Rhoads, 1983), does not have this positive effect (Cohen and Wills, 1985; Parker and Brown, 1982; Cronkite and Moos, 1984). Chapter 2 Method Subjects In order to adequately test the hypotheses in the present study and explore the factors that influence the intergenerational transmission of alcoholism as the Michigan State University Longitudinal Study sets out to do, one needs a sample of families with substantial but varying levels of alcohol consumption. These families should also be fairly young in order to follow them over time and monitor the development of drinking behavior and drinking related problems in the offspring. Finally, given the confounding effects of family stage and parent age on the development of these processes, such families should also be comparitively homogeneous for stage and age and should not be confounded by large differences in socioeconomic status. To recruit such a sample, two sources are utilized. The first source is a 100% population sample including all males convicted in the tri-county area of driving while impaired (DWI) or driving under the influence (DUIL), and who registered a blood alcohol level of 0.15% (150 mg./100m1.) or higher when arrested. To control for 59 60 families' age and stage, in addition, the men had to be currently married, living with their spouse, and have a son between the ages of 3.0 and 6.0 years. The second subset is of families of similar social background, parent age and family stage, whose alcohol problems compliment those of the first subset because their alcohol problems cover the lower range of severity. These families have been located using door-to-door recruitment techniques in the same census tract areas as the more heavy drinking families. Homogeneity of parent age and family stage was maintained by only allowing in families whose children are close in age (+/— 6 months), as the first subset. In addition, the parents in these families also had to be currently married and living with their spouse. Thus, the composite sample of 30 families (i.e. 60 adults), is a homogeneous one for parent and child ages, and family stage which is characterized predominantly by couples in their first marriage with preschool children. In addition, the sample encompasses a substantial diversity of parent drinking, and density of alcohol related problems. Recruitment Procedures Court personnel (probation officers) have been trained to inquire of all men who meet the above criteria about their willingness to discuss "possible participation 61 in a study of child development and family health" being run by Michigan State researchers. Each probation officer is provided with a short speech reviewing in detail what is to be said to each potential subject. If the individual agrees to talk with the MSU Family Study staff (understanding that there is no commitment to participate at this point, and understanding also that the study has no connection with the original court processes), his name is then released to project staff for contact and discussion of possible study participation. Ninety percent of the families contacted in this manner agreed to participate. Again, the remainder of the families in this study were recruited from the same census tracts as the above families. Families from this part of the sample were obtained by a door to door canvass of homes. Ninety percent of the families contacted at this point gave the project staff their name, phone number, and ages, sex and numbers of children. From this point, every family selected as fitting project criteria and asked to participate in the study gave their consent. Data Collection Procedures After the initial contact by the probation officer, a meeting with the family was arranged to explain the project in greater detail and to screen the families more 62 thoroughly for project criteria. Formal consent and the initial demographic information form and health history were also obtained at this time. The health history contained the Short Michigan Alcoholic Screening Test (SMAST; Selzer, 1975), and was used as the initial alcoholism screening inventory. The information on the SMAST insured that the sample did indeed contain individuals that covered a range of drinking behaviors from no drinking to alcoholic drinking. A more complete assessment of drinking behavior was made later in the study using a drinking and drug history. Analysis of these data confirmed the classification of individuals as ranging all the way from alcoholic drinkers to more moderate drinkers and abstainers. Measures Each family that participated in the project completed many questionnaires, direct observation sessions and interviews (see Zucker, 1980: Zucker, et al., 1984: Zucker, Noll and Fitzgerald, 1986). It is beyond the scope of this dissertation to review the methodology for the entire Michigan State University Longitudinal Study. The instruments included in the present study are measures of daily hassles, social support, and measures of behavioral and emotional symptomatology for adults. Table 63 2 presents a summary of the specific measures used here. The remaining text in the chapter describes each measure and its characteristics in greater detail. The complete project contact schedule may be found in Appendix 1. (A) Measurement pf Stressful Life Events: A measurement of life stress was necessary to test the hypotheses. Based on the earlier review of stress research and the controversy over measurement of stress, a measure of daily or chronic stress was chosen for this study. The Hassles and Uplifts Scales are considered to be the most appropriate measures of this sort of stress. These scales include 117 hassles and 135 uplifts. "Hassles are the irritating, frustrating, distressing demands that to some degree characterized everyday transactions with the environment. They include annoying practical problems such as losing things or traffic jams and fortuitous occurences such as inclement weather, as well as arguments, disappointments, and financial and family concerns," (Kanner et al., 1981; p. 3). Daily uplifts are "positive experiences such as the joy derived from manifestations of love, relief at hearing good news, the pleasure of a good night's rest, and so on," (Kanner et al., 1981: p. 6). Each item is rated on a three point scale for how strongly or how often it occurred. Three different scores are derived from these self-ratings: frequency (number of items endorsed), cumulated severity Table 2 Summagy 9; Study Measures A. Life Stress The Hassles and Uplifts Scales: Kanner et al., 1981. B. Moderator Variables The Social Support Questionnaire: Weil and Zucker, 1985. C. 1. Symptomatology The Hamilton Rating Scale for Depression: Hamilton, 1960. The (Short) Beck Depression Inventory: Beck et al., 1974. The Anti-Social Behavior Scale: Zucker and Noll, 1984. The Health History Questionnaire: Carpenter and Lester, 1980. Drinking Behavior The Short Michigan Alcohol Screening Test: Selzer et al., 1975 The Drinking and Drug History: Zucker, 1980 64 65 (summation of the three point ratings), intensity (cumulated severity divided by frequency). In the original research on this scale, 100 Caucasion men and women, aged 45-64, participated in a 12 month study where they completed the Hassles and Uplifts Scale 9 times at one month intervals. The subjects also completed a major life events scale, the Hopkins Symptom Checklist, and the Bradburn Morale Scale which is an index of psychological well-being. An effort was made not to confound Hassles and Uplifts items with psychological or physical symptoms, so items were eliminated from the scale if they were similar to those on the Hopkins Symptom Checklist. Test-retest correlations showed that subjects were experiencing approximately the same number of hassles and uplifts from month to month, but that the amount of distress or pleasure associated with them was more varied. In addition, state measures of hassles (and uplifts) were more highly related to each other than to a trait measure meant to indicate an individual's typical hassles. That is, a simple trait measure did not reveal the amount or type of hassles and uplifts a person was experiencing. Examination of item content showed that themes emerged which differentiated one sample group from another. For example, the middle-aged group reported economic concerns that did not appear in the student or 66 professional sample. On the other hand, students checked more items related to academic and social problems such as wasting time, concerns about meeting high standards, and being lonely. Checks for validity of the Hassles and Uplifts Scale were performed by correlating scale items with the Bradburn Morale scale. Results showed Hassles were significantly correlated with negative and not positive affect, while Uplifts were significantly related to positive and not negative affect. However, when these results were examined by gender, Uplifts were positively correlated to negative emotion for women and not for men. Results also showed a significant relationship between psychological symptoms and Hassles. Averaging over 9 months of Hassles, correlations of Hassles and month 10 scores of the Hopkins Symptom Checklist were .49 for the total, and .41 and .60 for men and women respectively. In sum, these date provide initial construct validation for the Hassles Scale in regard to its relationship to a significant adaptational outcome, i.e. psychological symptoms. (B) Measurement pf Adult Symptomatology: Measures of adult symptomatology are also required, including measures of depression, antisocial behavior, and physical symptoms. 67 (1) Measurement g; Depression: (i) The Hamilton Rating Scale (HRS:amilton, 1960, 1967), is completed by an interviewer after completion of the Diagnostic Interview Schedule, an instrument being used in another part of the MSU Vulnerability Study (Hamilton, 1960). The HRS is based on behavioral and somatic features which account for 50-80% of the total possible score. Affective or psychological symptoms account for the remaining percentage. The maximum possible score is 52 although scores higher than 35 are rare and scores around 30 indicate severe illness. There are a total of 17 items that are rated on 3- or 5- point scales. A factor analysis of the 17 items (N=152 male and 120 female psychiatric outpatients), produced six factors although it is unlikely that the 5th and 6th factors have reached stability so they are not described here. Factor one is considered a general factor of depressive illness, and measures the severity of symptoms. Factor two measures symptoms of anxious or retarded depression; a high score on Factor two indicates anxious depression while a low score indicates retarded depression. Unlike Factors 1 and 2, Factors 3 and 4 vary slightly for men and women. In men, Factor 3 indicates insomnia, loss of appetite, and fatiguability or, guilt, suicide, and loss of insight. In women, Factor 3 indicates loss of libido, fatiguability, and depression or, insomnia, agitation, 68 delayed insomnia and hypochondriasis. Factor four in men, indicates hypochondriasis, loss of weight and of insight or a varied collection of symptoms. Factor four in women, indicates loss of energy, loss of appetite, and hypochondriasis or, guilt, suicide, depression, and agitation. Inter-rater reliabilities have ranged between .80 and .90 for the same interview (Hamilton, 1969). The inter- correlations between the HRS and the Beck Depression Inventory (BDI) was 0.75 indicating some stability across measures. In this case, the BDI was not based on traditional self-ratings (Hamilton, 1969). For this study, the Hamilton Rating Scale reliability was evaluated by way of a series of conjoint interviews conducted by the project interviewer (RAZ) and a psychiatrist who already had established reliability with the Hamilton and had used this instrument with several hundred patients. During these inter-views each rater filled out the Hamilton separately. Neither interviewer's score deviated from the other by more than three points, and the Pearson r of the two raters' scores was 0.94. (ii) The (Short) Beck Depression Inventory: (8801; Beck et al., 1961) is used as a self report measure of depression in this study and is completed by both adults. The original BDI contains 21 items each containing four 69 alternative statements graded in severity from 0 to 3. The items were derived from clinical experience in psychoanalytic psychotherapy with depressed patients. Therefore, each of the items describe specific behavioral manifestations of depression. The BDI shows high internal consistency with statistically significant correlations between each of the 21 categories and the total depression score (N=409). Test-retest reliability was determined by comparing clinicians' ratings of depression with the BDI at two differrent times. The changes in clinicians' ratings were parallel to changes on the BDI scores. According to the researchers, this indicates a consistent relationship between the instrument and the patient's clinical state. Validity was examined by comparing psychiatrists' interview ratings with the BDI across time intervals from two to five weeks. In 85% of the cases, the BDI predicted a change in the clinicians' depth of depression ratings. Though this inventory measures intensity of depression, it does not distinguish among standard diagnostic categories. The Short Beck Depression Inventory (SBDI), the short form of the BDI, consists of 13 of the original BDI items. The 13 items were chosen by ranking all items according to the correlations between each individual item, the total score, and the clinicians's rating. The ranked items were then added one by one, and again correlated with total 70 scores and with clinicians's ratings until cumulative correlations leveled off. This short form correlated .96 with the total score and .61 with the clinician's ratings of depression. Next, a cross-validation study was conducted on a sample of general practice patients with no suicide attempts, and a group of schizophrenic patients, (N-431). Each subject filled out both forms of the BDI and was interviewed by an experienced clinician. Correlations between the two forms ranged from .89 to .96. Correlations between the short form and the clinicians's ratings ranged from .55 to .67. All of these correlations were significant. Indications are that the short form is quite adequate in predictive power (Beck et al., 1974). (2) Measurement g; Antisocial Behavior: The Antisocial Behavior Scale (Zucker & Noll, 1984) represents a modification of an earlier instrument, the Adolescent Antisocial Behavior Questionnaire (Zucker & Barron, 1973) that incorporates 18 of the items from this earlier inventory, as well as a larger number of social and antisocial items that are more suitable for an adult population. These items were drawn from a variety of sources, including antisocial items from the Diagnostic Interview Schedule (Robins, 1981) and from areas of behavior included in antisocial personality disorder, as described in the Research Diagnostic Criteria (Feighner, Robins, Guze, Woodruff, Winokur and Munoz, 1972). The 71 final instrument is composed of 46 items that have been categorized by content into 9 subscales including parental defiance, sexual behavior, delinquent behavior, leaving the field, serious physical aggression, excitement and sensation seeking, job related anti-social behavior, school related anti-social behavior, and trouble with the law. Each item is rated for the lifetime frequency of involvement in that activity. Again, both adults complete this questionnaire. Psychometric properties of the instrument appear adequate. Test-retest reliability over a four week interval with 151 undergraduate college students is 0.91: the coefficient alpha is 0.80 (Zucker & Noll, unpublished manuscript). Due to the nature of these items, social desirability may be a factor in responses. That is, one might assume subjects would under report engaging in the listed activities because of their undesirable nature. To minimize this effect, the directions were phrased in an accepting manner, with the expectation communicated that many people engage in the behaviors being asked about. Initial construct validity studies with this inventory indicate that is successfully distinguishes between college students, community court samples, and prison samples, i.e., it adequately differentiates level of antisocial involvement (Noll and Zucker, 1986). 72 (3) Measurement pf Health History: The final questionnaires used to measure disorder are health history forms completed by both adults. This extensive self-administered history-questionnaire was developed by the Rutgers Longitudinal Study (Carpenter and Lester, 1980) to assess health and illness status in fifteen areas: hospitalization history; current medication use; allergies; prior illnesses; skin and hair problems; eye, ear, nose and throat symptoms; heart and lung; G.I. tract; urinary tract; skeleton and joints; nervous system; alcohol and drug use; general health care patterns; diet and weight control; and physical fitness activities. The wife's form, in addition, contains questions regarding the target child's birth and early developmental history. (C) Measurement g; Social Support: A measure of social support was needed to examine the hypothesis that alcoholic and non-alcoholic families experience different types and amounts of social support and social support is less effective in preventing disorder in alcoholic families than in matched control families. A large number of studies have shown a connection between social support networks, life stress, and illness. There have been an almost equally large number of 73 definitions and instruments to measure social support. Networks have been measured by size, density (the extent to which individual's within the network know each other), multiplexity (the number of kinds of support given), quality or strength of support, reciprocity of support, length of relationships, and frequency and ease of contact (Mitchell and Trickett, 1980). Support has also been catagorized into emotional support, task-oriented assistance, maintanence of identity/worth, and sharing of skills and information (Broadhead, Kaplan, James, Wagner, Schoenback Crimson, Heyden, Tibblin, and Gehlback, 1983). At this point, the literature does not clearly point to one method of measuring support over another in regard to what aspect of support buffers stress most effectively. In addition, there is not a single social support instrument that measures every type of support. As a result, this study modified an already existing instrument, the Social Support Questionnaire, by Norbeck, Jondsey, and Carrieri, (1981), in order to measure the size, density, frequency, duration, quality, and multiplexity of the support networks. This instrument was chosen as a starting point because of the extensive reliability and validity work that has been done on it. (See Appendix 2 for a copy of the Norbeck instrument, and of the revised social support measure used here.) 74 Initial studies on the Norbeck instrument tested a population that consisted of 135 mostly Caucasion senior and graduate level nursing students with a mean age of 28.5 years. The students were also administered the Marlowe-Crowns Social Desirability Scale, the Social Support Questionnaire by Cohen and Lazarus, the Profile of Mood States (POMS), and the Life Experiences Survey (LES), by Sarason et a1. (1978). The same measures were completed twice at a one week interval. Test-retest reliability coefficients on items measuring affect, affirmation and aid ranged between .85 and 0.92 while size of network, duration of relationships, and frequency of contact all had coefficients of .92. Internal consistancy was quite high as well with all of the items correlating with each other from .69 to .98. As a result of these intercorrelations, the authors collapsed their items into three variables: total functional support (affect, affirmation, aid), total network support (size, frequency, duration), and total loss (amount and kinds of support loss). Social desirability was not a factor in this test as social support items only correlated between .01 and .17 with the Marlowe-Crowns scale. Concurrent validity with the Cohen and Lazarus scale was weak for aid support (-.03) and affective support (.51). Unfortunately the 75 scales from the two measure vary somewhat in content, making comparison difficult. The determination of construct validity of this scale was more complex. Total loss, as measured by Norbeck's instrument, was significantly correlated with confusion (.26) and depression (.24) on the POMS subscales but, total functional support and, total network support scales were not. In addition, no significant results were found when social support, the POMS, and the LES were entered in a multiple regression analysis. That is, life events showed no relationship to mood, and social support showed no moderating effects on this relationship for this particular, nonclinical sample. However, the content of this scale is based on common conceptualizations of social support and network properties which have shown significant relationships with stress and adjustment in many other studies (Barnes, 1972; Cobb, 1976; Mitchell and Trickett, 1980; Broadhead et al., 1983). In addition, the internal validity and reliability of the Norbeck scales is quite robust, indicating that it is a sound research instrument. As stated earlier, the Social Support Interview (SSI) used in the current study is an expanded version of the Norbeck measure. It retains the original measures of network size, frequency of contact, duration of relationships, aid, affect, ahd affirmation, but also the 76 581 adds items that allow assessment of network density, interpersonal similarity, and organizational support (e.g. club and church support). These additional variables have also been cited in the theoretical literature as important measures of social support and, various forms of these variables have been used in several studies, though no validity or reliability data were collected at those times (Barnes, 1972; Mitchell and Trickett, 1980). The specific questions used to elicit the data were designed by this researcher and were formed to tap the concepts described in the literature regarding density, similarity, and organizational support. As a result, the validity and reliability of these factors is unknown; however, the addition of these variables should not alter the statistical robustness of the factors derived from the Norbeck Social Support Questionnaire. (D) Measurement pf Drinking Behavior: Measures of drinking behavior, drinking related problems, and drinking history are needed to determine a drinking density score (described earlier), for subjects in this study. The SMAST and the Drinking and Drug History were chosen for this purpose because of their robustness as instruments (reviewed below), and their incorporation of items relevant to formal diagnostic criteria for alcoholism. This latter factor is important 77 as it insures the validity of these items as significant drinking related behaviors. (l) The Short Michiggp Alcoholism Screening Test (SMAST: Selzer, Vinoker & VanRooyan, 1975) is completed by both adults in the families and is a measure of drinking related problems. It is a shortened form of the MAST, and contains a total of 13 items that are written in a yes/no response format. The original MAST contained a total of 25 items and was first administered to hospitalized alcoholics, persons convicted of drunk driving, persons convicted of drunk and disorderly behavior, drivers whose licenses were under review, and a control group. The validity of the MAST was assessed by reviewing legal, medical and social agencies files and reviewing driving and criminal records (Selzer, 1971). Subjects' ages ranged from 19-73 years. After scoring the MAST and comparing results to other sources of data, 204 of 219 subjects were correctly classified as alcoholic or nonalcoholic on the MAST. The remaining 15 subjects were false negatives, i.e., they were alcoholic according to medical, legal and social records but were classified as nonalcoholic on the MAST. Closer examination of these cases showed that these subjects had underreported their legal involvements. As a way of minimizing false negatives, the author recommended using 78 court records as well as MAST data in categorizing individuals as alcoholic or not. The revised form of the MAST was tested on 501 male drivers over the age of 20 (Selzer et al., 1975). Of this sample, 102 drivers were routinely renewing their licenses, 171 drivers had been sent to a driver safety school by the courts a result of one or more moving traffic violations, 129 subjects were undergoing inpatient treatment for alcoholism, and 99 subjects were undergoing outpatient rehabilitation for alcoholism. Mean ages of these subjects were 31.8, 29.5, 47.6 and 37.6 respectively. Mean education level was 3, 2, and 1 year of college and 12th grade, respectively while mean income level was $9,750., 9,500., 12,500, and 9,500, respectively. Reliability of this MAST in terms of internal consistency was .95 for the entire sample and was .83 and .87 for alcoholic and nonalcoholic groups. Validity coefficients ranged from .79 to .90. Although these are fairly high, the investigators were concerned about the effects of denial on these results. Therefore, alcoholic and nonalcoholic groups were given the deny-bad subscale on the Marlowe-Crowns Social Desirability Scale with these scores being correlated with the MAST scores. Correlations equaled -.18 and -.11 (p<.01) for alcoholic and nonalcoholic groups respectively. According to the 79 authors, these results are significant but fairly small, indicating that the effect of denial on the MAST was "negligable". In addition to reliability and validity analyses, a step-wise regression analyses was employed to select the MAST items that significantly improved the prediction of alcoholic or nonalcoholic categories. The result was the 13 item SMAST. Separate computations of SMAST reliability were done and yielded coefficients of .76, .78 and .93 for alcoholic, nonalcoholic and total sample, respectively. Correlations between the MAST and the SMAST yielded r's of .93, .90, and .97, respectively. Validity coefficients with alcoholic and nonalcoholic critierion groups were .94. The authors concluded that both the MAST and the SMAST are effective screening devices for alcoholism and identify drinking related problems. (2) The Drinking and Drug History This self-report questionnaire is also completed by both adults and measures amount and types of alcohol- related problems, quantity, frequency, and variability of alcohol consumption. It also measures the length of time the individual has been drinking and experiencing drinking related problems. The questionnaire is composed of already much tested items from the 1974 NIDA Survey (Johnston, O'Donnell, Voss, Clayton, Slatin, and Ross, 80 1976), from the American Drinking Practices Survey (Cahalan et al., 1969), and from the V.A. Medical Center (U. of California) San Diego, Research Questionnaire for Alcoholics (Shuckit, 1978). All of the items have been extensively used in a variety of survey and clinical settings. Chapter 3 Results Analysis The overall design of this study is cross-sectional, with a sample composed of 60 individuals, (30 couples), who vary in their drinking behavior from abstainers to heavy drinkers, but who as a group are heavily skewed toward the heavy drinking end of the continuum. Frequency distributions of 1) rate of alcohol consumption (QFV), 2) number of drinking problems and 3) Lifetime Alcohol Problem Scores (LAPS), for these subjects confirm that this sample encompasses a wide range of drinking behavior (Figures 3, 4 and 5), although from a population standpoint, these respondents are, again, clearly heavily skewed toward the heavy drinking and of the population (Cahalan, Cissin, and Crosley, 1969).* As would be expected from prior drinking studies, there are clear sex differences in this sample, with women drinking less and having fewer drinking related problems than the men. Socio-demographic characteristics of the sample are presented in Table 3. *Since a skewed distribution may distort the statistical analyses, the distribution of men's QFV scores was normalized via a logrithmic transformation. Results from analyses with the transformed variables were then compared to results from analyses with the original QFV values. There was no significant difference between these results; therefor, analyses with the original values are presented in this dissertation. 81 KxXXNKUIXXKKNKXXXXKKKXKXXXKX x MEN x x (N=30) x x x x x x x 4 3 2 1 x x x x x x x x WOMEN x x x x (N830) x x x x x x x x x x x Note: 5-Heavy Drinkers, 3-Moderate Drinkers, 1-Abstainers Figpre 3; Frequency Distribution of QFV Scores for Men and Women 82 MEN (N830) WOMEN (N=30) XXXXXXXXXXXX X XXXX XXX XX X XX XXX 3215.. HS IDQQOIUIO-UNI-‘O XXX 10 XX KICK g...- N XXXX 17 xx 20 XXX 23 Figpre _4_. Frequency Distribution of Number of Drinking Related Problems for Men and Women 83 MEN (N330) XXXXX XXX XXXX XX XX XX Figpre 5; -4.0 -3.5 -3.0 -2.5 -2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 WOMEN (N330) XX XX XXXXX XX XXXX XX XX X Frequency Distribution of Lifetime Alcohol Problem Scores for Men and Women 84 Table 3 Demographic Characteristics‘gf the Sample MEN (N830) Age in years i 31.0 S.D. 4.2 Years of Education 2 12.6 8.0 2.4 Highest Degree Earned: Elementary 23% High School 57% Technical 7% College (B.A.) 7% Graduate (M.A.) 3% Graduate (Ph.D.) 3% Religion: Protestant 30.0% Catholic 13.3% Jewish 0% Orthodox: Evangelical, Greek, etc. 3.3% Other 3.3% None 40.0% Number of Marriages: one 73% two 23% three 4% Family Income: less than 7,000 8% 7,001 to 13,000 17% 13,001 to 20,000 14% 20,001 to 30,000 30% 30,001 to 50,000 24% greater than 50,000 7% 85 WOMEN (N'30) 29.9 4.3 13.3 2.3 3% 80% 10% 7% 0% 46.6% 16.6% 3.3% 3.3% 10.0% 20.0% 80% 20% 0% TRI-COUNTY AREA a 28.9b 76%c 17%d 20.9% 27.0% 26.2% 20.4% 4.6% Table 3 (cont'd.) MEN WOMEN TRI-COUNTY (N=30) (N-30) AREA Number of Children: one 23% two 39% three 30% four 8% a-Data obtained from the 1980 U.S. Census, includes Clinton, Eaton and Ingham counties in Michigan. b=This figure is based on the median age of the population. c-This figure is based on the years of school completed by individuals who are 25 years or older. dsThis figure is based on the years of school completed by individuals who are 25 years or older, and includes individuals who have completed 16 p; more years of education. e=This is the percent of families who have an income of $10,000 or less. f-This is the percent of families who have an income of $10,001 to $20,000. 86 87 These demographic data show that the sample is composed of husbands and wives who average 30.5 years of age and have completed an average of 13 years of education. This is the first marriage for 73% of the men and 80% of the women, with all of the families having at least one child, and 76% of the families having two or more children. When a religious preference was stated, the largest portion of the sample stated that they were Protestant; however, 40% of the men and 20% of the women stated that they had no religious preference. Finally, 31% of the families have an income between $7,000 and $20,000 and, 54% of the families have an income between $20,001 and $50,000. Turning to the separate measures of stress, symptoms, and social support, the results from each individual questionnaire were examined, with variables within each questionnaire being collapsed into a composite variable via established scoring techniques for that particular questionnaire. This composite variable was then used in further analysis. In addition, when possible, the composite symptom variables were calculated two ways (A), for current symptom level and (B), for lifetime symptom level. 88 The primary drinking variable used in the analyses is a composite measure designed to express (and assess) differences in extent of drinking difficulty over the life course. This measure was developed by Zucker (1986) as a way of scaling and summarizing the multiplicity of lifetime troubles that contribute to alcoholic careers with different levels of "density". His description of the measure and rationale follows: ”From a longitudinal perspective, the term alcoholism is conceived as a dimensional concept, that reflects in differing degrees-- beyond a particular threshold value--that drinking extent and problems have been present in varying severity over the life course of the individual. This particular measure, the Lifetime Alcohol Problems Score (LAPS), scales the dimension by way of a multiple index composed of three equally weighted sets of information on drinking. The first is an index of earliest drinking difficulty (age of first drunkenness), the next is an index of current drinking difficulty (Cahalan, Cissin & Crossley's (1969) Quantity-Frequency-Variability (QFV) index of extent of drinking in the last six months), and the last is the sum of number of years between earliest and most recent occurrence of a varied list of drinking problems (i.e. an integrated summation of "total drinking problem-years"). 89 The specific formula for calculating LAPS follows: LAPS I a + b + c Where a - QFV (standardized score); x - 10, SD I=1. b - drinking problem-years/current age (standardized score), where drinking problem-years - drinking problem x years elapsed between first occurrence and most recent occurrence of the problem* (Summed over all drinking related problems); x - 10, SD - 1. c = [1/age first drunk]2 (standardized score); x - 10, SD - l." (Zucker, 1986) The data necessary to calculate the LAPS measure were obtained from information on the Drinking and Drug History questionnaire. * Thus if a respondent only had an occurrence of a problem at age 20, the score would be 1 ”problem year"; if the first occurrence of the problem were at 20 and the most recent was age 21, the score would be 2 ”problem years" for that problem, etc. 90 The QFV score is the second major drinking variable used in these analyses and is considered a measure of current drinking behavior. This score is derived by assigning a one to five rating to each individual based on the quantity, frequency, and variablility of their drinking in the last six months (Cahalan, Cisin, and Crossley, 1969). After examining the distribution of these ratings in Figure 3, it was apparent that current drinking for men in this sample did not follow a normal distribution. This violated a basic statistical assumption for further analyses; therefore, a logarithmic transformation of men's QFV scores was used, along with the original QFV scores, throughout the analyses. The transformed QFV did not alter the significance of the results, so the following tables contain the statistics calculated with the original QFV score for men. One question needs to be raised at the outset, namely whether the observed effects, between drinking and symptomatology, are to some extent spurious since a number of the symptom measures incorporate items that also deal with drinking events. (For example, on the Hassles scale, one item is "Had a fight with spouse".) Given this possibility, a second set of analyses were performed involving the Hassles, Social Support, and Symptom scales, (i.e. all the scales where there was potentially confounding drinking related content). In these 91 additional analyses the potentially confounding variables were removed. In no case did the significance of the results change. Consequently, the correlations reported and discussed here involve the original scale measures. After these preliminary steps, the separate relationships between drinking behavior and social support, stress, and symptomatology were examined by Pearson product moment correlations (SPSS-9). The interrelationships of these variables were then explored via a series of hierarchical multiple regression analyses (SPSS-9). Hassles and Alcohol Involvement: In the first hypothesis, it was predicted that spouses in families with a high lifetime density of alcohol consumption would experience more daily hassles, particularly interpersonal hassles, than would spouses in families with a lower lifetime density of alcohol consumption. The data gathered via the Hassles Scale and the alcohol measures described in the previous section, were used to test the hypothesis. The index of hassles used was frequency of hassles in the last month. Because measures of intensity of hassles produced similar results, only frequency measures are reported. When lifetime drinking was examined (Table 4), the correlational analyses showed no significant relationships Table 4 The Relationship Between Hassles and Measures g; Self Drinking and Spouse Drinking (Pearson R's) Self Drinkingi a b Husbands (N=30): QFV LAPS Overall Hassles -.33* .23 Interpersonal Hassles -.41* -.13 Self Drinking; Wives (N=30): QFV LAPS Overall Hassles -.12 .12 Interpersonal Hassles .01 .25+ Spouse Drinking: QFV LAPS -.22+ -.01 -.25+ -.12 Spouse Drinkingi QFV LAPS -e26+ e22 -.42** -.02 Note: a-The Quantity-Frequency-Variability Score, the measure of current drinking. b-The Lifetime Alcohol Problem Score, the measure of lifetime alcohol involvement. Note: +2<.1o, *2<.os, **p<.01, ***p<.001 92 93 between drinking and hassles for individuals or between husbands and wives; the relationship between wives' interpersonal hassles and LAPS tended toward significance (p<.10). However, when the quantity-frequency-variability measure of drinking (QFV) was correlated with current hassles, the results were quite different (Table 4). In this case, the husbands' current drinking practices were negatively related to their own level of interpersonal hassles (r--.41, p<.05), as well as general hassles (r-- .33, p<.05). That is, the heavier a man was drinking, the fewer hassles he perceived experiencing. Similarly, wives were experiencing fewer interpersonal hassles when their husbands were drinking more heavily (r--.42, p<.01). Wives' current drinking practices were not significantly related to their own or their husbands' amount of hassles experienced (Table 4). In sum, men's level of lifetime alcohol involvement was not significantly related to the level of daily hassles they reported experiencing. However, heavier current drinking was unexpectedly related to lgggp levels of perceived hassles. In contrast, women's level of lifetime alcohol involvement tended toward a significant relationship with women's level of perceived interpersonal hassles, while current drinking and hassles were unrelated. When looking at the interaction in the couples, the results showed that heavier current drinking 94 by either spouse tended toward a signficant relationship with an individual's lower level of perceived hassles. Generally, the above results did not support the hypothesis that higher levels of lifetime alcohol involvement would lead to higher levels of hassles. And, when current drinking was examined, heavier alcohol consumption was actually related to lower levels of hassles. Since these results were unexpected, further analyses were performed in an attempt to understand the relationships. Specifically, it was felt that the effects of daily hassles might be minimized by the occurrence of daily uplifts. (Uplifts are the theoretical opposite of daily hassles in that they are everyday positive things that may occur.) If uplifts do buffer the effects of hassles, the occurrence of uplifts would be expected to weaken an otherwise positive relationship between hassles and drinking. To test this idea, lifetime and current drinking measures were correlated with uplifts, the difference in frequency between hassles and uplifts, and the ratio of hassles to uplifts. None of these relationships reached significance for men or women, indicating that uplifts were not a significant factor in the relationship between an individual's level of drinking and his amount of hassles perceived. Uplifts did not appear to buffer the effects of hassles on drinking 95 behavior. (A table describing these results can be found in Appendix 3.) Symptom Occurrence and Alcohol Involvement: Part One of the second hypothesis states that individuals with a high lifetime pattern of alcohol involvement would have higher levels of antisocial behavior when compared to individuals with low lifetime alcohol involvement. The results in Table 5 strongly support this hypothesis for both men and women; as lifetime alcohol involvement increases, the individual's level of antisocial behavior also increased. However, Table 5 indicates that alcohol involvement is related to other symptoms as well. To recapitulate, for men, both lifetime alcohol involvement and current alcohol consumption are positively related to antisocial behavior in childhood and adulthood. This positive relationship between childhood antisocial behavior and LAPS implies that childhood behavior may be predictive of long term adult drinking practices for men. In addition, LAPS is positively related to men's self- reported and interview-rated depression. In other words, higher levels of lifetime alcohol involvement are related to higher levels of both internalzing and externalizing symptoms for men. Table 5 The Relationship Between Symptomatology and Measures g; Self Drinking and Spouse Drinking (Pearson R's) Externalizing Symptoms Self Drinkingi Spouse Drinkingi Husbands (N-30): QFV LAPS QFV LAPS Antisocial Behavior- .43** .75*** .34* -.12 Total Antisocial Behavior in .38* .66*** .26+ -.04 Childhood Antisocial Behavior in .33* .61*** .36* -.15 Adulthood Self Drinkingi Spouse Drinkingi Wives (N-30): QFV LAPS QFV LAPS Antisocial Behavior- .09 .45** -.21 .30* Total Antisocial Behavior in .08 .27+ -.21 .06 Childhood Antisocial Behavior in .11 .46** -.18 .33* Adulthood 96 Table 5 (cont'd.) Internalizing Symptoms Self Drinking: Spouse Drinking: b a Husbands (N330): QFV LAPS QFV LAPS Physical Illness .09 .11 .03 .01 Depression (Self-Reported) .14 .56*** -.15 .21 Depression (Hamilton Scale) .23+ .33* .14 .02 Self Drinking; Spouse Drinkingi Wives (N830): QFV LAPS QFV LAPS Physical Illness .00 .27+ -.29+ .25+ Depression Depression (Hamilton Scale) .15 .15 -.23+ .11 Note: a-The Quantity-Frequency-Variability Index, the measure of current drinking. b-The Lifetime Alcohol Problems Score, the measure of lifetime alcohol involvement. Note: +p<.10, *p<.05, **p<.01, ***p<.001 97 98 For women, only lifetime alcohol involvement is positively related to adulthood antisocial behavior and tends toward significance when related to childhood antisocial behavior. There is also a positive relationship between LAPS and women's self-reported depression and another positive trend between LAPS and physical illness. Again, like the men, increased lifetime alcohol involvement is related to higher levels of both internalizing and externalizing symptoms. Part Two of the second hypothesis also predicted a relationship between one spouse's drinking and the other spouse's adjustment; specifically, when one spouse had a high level of lifetime alcohol involvement, it was anticipated that the other spouse would report high levels of depression and physical illness. This hypothesis did not stand up for men in that their internalizing symptoms of depression and illness were not related to their wives' drinking. However, higher levels of men's externalizing antisocial behavior is related to higher levles of wives' current drinking. That is, men are acting out more when their wives are drinking more. Again, the results are slightly different for women. Increases in husbands' LAPS tend to be related to increases in wives' physical illness and current depression, as well as being significantly related to increases in wives' antisocial behavior. These findings 99 are all in the expected direction. However, contrary to expectations, increases in husbands' current levels of drinking tended to be related to decreases in wives' level of physical illness and depression. Wives appeared better adjusted when their husbands were currently drinking more. To summarize this last set of results, it is apparent that antisocial women are married to men with heavy lifetime drinking histories and, similarly, antisocial men are married to women who are currently drinking heavily. In addition, the women married to the heavier lifetime drinkiers are experiencing more physical illness and depression. Whereas, women married to husbands who are currently heavier drinkers are experiencing less physical illness and less depression. Social Support and Alcohol Involvement: The third hypothesis predicts that individuals with high lifetime drinking densities, and their spouses, will have less social support than those individuals with low drinking densities and their spouses. The data do not support this hypothesis as none of the social support measures were related to lifetime or current drinking patterns. Furthermore, there appears to be a clear trend in the data for a low order positive relationship between an individual's current drinking pattern and his or her level of perceived social support (Table 6); the more an Table 6 The Relationship Between Social Support and Measures 93 Self Drinking and Spouse Drinking (Pearson R's) Self Drinkipgi Spouse Drinking: a b Husbands' (N=30): 93! AAES ggy AAES Number in Network .23+ .17 -.13 -.02 Function of Network .19 .12 -.09 -.02 Density of Network .24 .19 -.09 -.08 Total Network Score .20 .12 -.09 -.02 Self Drinking; Spouse Drinking: Wives' (N=30): 9:2 AAES 9:! AAES Number in Network .21 .21 .13 .07 Function of Network .21 .16 .13 .10 Density of Network .21 -.23+ .16 -.13 Total Network Score .21 .16 .14 .10 Note: asThe Quantity-Frequency-Variability Index, the measure of current drinking. b-The Lifetime Alcohol Problems Score, the measure of lifetime alcohol involvement. Note: +p<.10, *p<.05 100 101 individual currently drinks, the more social support he is reporting. This trend is not as clearly apparent when examining the relationship between social support and lifetime alcohol involvement and there are no spouse effects present. 2A3 Interaction 2; Stress and Symptomatology Ag Couples: Previous hypotheses have looked at the direct effects of alcohol consumption on individuals and their spouses. However, it is possible that one individual's drinking is having an indirect effect on his or her spouse. That is, alcohol consumption may directly effect an individual's level of symptomatology; symptomatology, in turn, may impact on the spouse and his or her adjustment. The fourth hypothesis is aimed at this possibility, predicting that the stress and symptoms of one member of a couple will be related to the stress and symptoms experienced by the other member. The data relating to this hypothesis are presented in Tables 7 thru 9. A review of these data shows that there is a strong relationship between husbands' and wives' stress, their symptoms, and their stress and symptoms. First, there is a significant positive relationship between husbands' and wives' level of overall hassles (p-.32, p<.05) and Table 7 The Relationship between Husbands' and Wives' Daily Hassles (Pearson R's) Husbands (N-30): Wives (N-30): Total Hassles Interpersonal Hassles Total Hassles r-.32* r-.37* Interpersonal Hassles r=.23 r-.35* Note: +p<.10, *p<.05 102 103 interpersonal hassles (p-.35, p<.05) (Table 7). However, the relationships between hassles and symptoms and, symptoms and symptoms, varies depending on whether the symptoms are internalizing or externalizing. Therefore, these relationships will be discussed based on this distinction. To aid this discussion, the data have also been divided into tables based on these internalizing and externalizing categories (Figure 6). Table 8A illustrates the relationship between spouses' internalizing symptoms. It is apparent that current self-reported depression in husbands and wives is positively related as is physical illness and depression. The relationship between each spouses' physical illness also shows a tendency to be positively associated. Table 88 illustrates the relationship between spouses' externalizing symptoms which generally show a low order positive trend toward significance. The relationship between spouses' externalizing symptoms is of a lower order than the relationship between spouses' internalizing symptoms. 3The high correlations between daily hassles for husbands and wives may exist as a function of the design; hassles may not be independent events for two individuals living in the same house. However, the Hassles measure that was used examines an individual's perception of an event as stressful, and thus focuses on the potentially more independent aspects of the Hassles event. Husbands' Symptomatology: Internalizing Externalizing Wives' Symptomatologyi I I I I I I Internalizing I Table 8A I Table BC I I I I I I I I I I I I I Externalizing I Table 8D I Table 8B I I I I I I I Figpre 9; Outline of Tables on the Relationship Between Husbands' and Wives' Symptomatology 104 Table 8A The Relationship Between Husbands' and Wives' Internalizing Symptomatology (Pearson R's) Husbands' Symptoms: (N830) Physical Depression Depression Illness (Beck) (Hamilton) Wives' Symptoms: (N-30) Physical .29+ .50** .33* Illness Depression .49** .65*** .25+ (Beck) Depression -.18 .08 .19 (Hamilton) Note: +p<.10, *p<.05, **p<.01, ***p<.001 105 Table 8B The Relationship Between Husbands' and Wives' Externalizing Symptomatology (Pearson R's) Husbands' Sypptoms: Childhood Antisocial Behavior Wives' Symptoms: (N=30) Childhood Antisocial .19 Behavior Adulthood Antisocial .27+ Behavior Total Antisocial .25+ Behavior Note: +p<.10, *p<.05, **p<.01, 106 (N830) Adulthood Antisocial Behavior .12 .22 .19 ***p<.001 Total Antisocial Behavior .22 .30* .29+ 107 Table 8C presents the correlations between husbands' externalizing symptoms and wives' internalizing symptoms. None of these relationships are signficant. However, the relationships between husbands' internalizing symptoms and wives' externalizing symptoms are quite strong and positive (Table 80). Specifically, husbands' physical illness and current depression are both positively related to their wives' adulthood antisocial behavior. In addition, husbands' current depression is positively related to wives' childhood antisocial behavior. The next step in these analyses was to examine the association between spouses' stress and both internalizing and externalizing symptoms. These results are presented in Table 9. Again, there is a difference in the significance of these relationships by type of symptom. A positive relationship exists between spouses' hassles and individuals' current level of self-reported depression, while there is no significant relationships between spouses' hassles and individuals' antisocial behavior. This stronger relationship between hassles and internalizing symptoms is also apparent when an individual's hassles and symptoms are examined (Table 9). Both husbands' and wives' hassles are positively related to their own level of physical illness and current depression but, hassles are not related to antisocial behavior. Table 8C The Relationship Between Husbands' Externalizing and Wives' Internalizing Symptomatology (Pearson R's) Husbands' Symptoms: (N-30) Childhood Adulthood Total Antisocial Antisocial Antisocial Behavior Behavior Behavior Wives ' Symptoms: (N-30) Physical .12 .07 .20 Illness Depression .00 .04 .07 (Beck) Depression .19 .14 .23 (Hamilton) Note: +p<.10, *p<.05, **p<.01, ***p<.001 108 Table 8D The Relationship Between Husbands' Internalizing and Wives' Externalizipg Symptomatology (Pearson R's) Wives' Symptoms: (N-30) Childhood Antisocial Behavior Adulthood Antisocial Behavior Total Antisocial Behavior Physical Illness .26+ .46** .41** Husbands' Symptoms: (N-30) Depression Depression (Beck) (Hamilton) -.01 .39* .37* .47** .20 .48** Note: +p<.10, *p<.05, **p<.01, ***p<.001 109 110 This second set of analyses, which examines the relationship between self-hassles and self-symptomatology, verifies the existence of a positive relationship between these factors. Past literature has consistently shown a connection between stress and illness, but stress has usually been defined as the occurence of major life events (Rabkin and Struening, 1978). The pilot study for the current project (Weil, 1984), supported the earlier literature, finding the strongest relationships to exist between major events and both antisocial behavior and worst-ever interview-rated depression. No relationship was observed between this type of stress, i.e. major life events, and physical health. Conversely, results of the present study show strong positive relationships both between daily hassles and and physical illness as well as between daily hassles and depression. However, there was no relationship between daily hassles and antisocial behavior. In other words, different types of stress appear to have some similar and some different effects. Major life events are related to antisocial behavior. Daily hassles are related to physical illness. And, both major events and daily hassles are related to depression. These results are also presented in Table 9. Finally, the relationship between internalizing and externalizing symptoms within individuals is examined (Appendices 4A and 4B). Summarizing the results for Table 9 The Relationship Between Symptomatology and Measures pf Self Hassles and Spouse Hassles (Pearson R's) Husbands (N-30): Physical Illness Depression (Self-Reported) Depression (Hamilton Scale) Wives (N-30): Physical Illness Depression (Self-Reported) Depression (Hamilton Scale) Internalizing Symptoms Self Hassles .44** .59*** ,57*** Self Hassles .40** .56*** .11 111 Spouse Hassles .23+ .36* .11 Spouse Hassles .07 .46*** .14 Table 9 (cont'd) Externalizing Sympggms Husbands (N-30): Self Hassles Spouse Hassles Antisocial Behavior in .15 -.06 Childhood Antisocial Behavior in .11 .09 Adulthood Antisocial Behavior- .14 .02 Total Wives (N-30): Self Hassles Spouse Hassles Antisocial Behavior in .12 .15 Childhood Antisocial Behavior in .24+ .03 Adulthood Antisocial Behavior- .08 .04 Total Note: +p<.10, *p<.05, **p<.01, ***p<.001 112 113 women: physical illness shows a strong positive relationship to self-reported depression (p-.68 p<.001), to interview-rated depression (£-.31, p<.05), and to antisocial behavior (3?.46, p<.01). There is also a positive relationship between self-reported depression and antisocial behavior (p-.38, p<.05). And, there is a very strong positive relationship between women's childhood and adulthood antisocial behavior (p-.89, p<.001). In men, physical illness is also related to self- reported depression (£-.51, p<.01), and to interview-rated depression (p-.46, p<.01). However, physical illness and self-reported depression are not related to antisocial behavior in men while interview-rated depression is (3!.44, p<501). Last, there is a very strong relationship between men's childhood and adulthood antisocial behavior (_z_-.92, p<.001) . To summarize, results for both sexes show A gAggg link between physical illness and depression, depression and antisocial behavior, and antisocial behavior Ap childhood and adulthood. However, 3 clear difference between men and women Ag the relationship between physical illness and antisocial behavior; this Ag gAgnificant for women but not for men. A less clear difference between the sexes is the link between depression and antisocial behavior. That is, for women, self-reported depression is related to antisocial behavior but, for men, interview- 114 rated depression is related to antisocial behavior. The explanation for this is not readily apparent, except that the measures of self-reported and interview-rated depression appear to measure slightly different things for women while they are tapping a more similar theme in men. Their intercorrelation is r-.24 (p<115), among women and, r=.56 (p<.001), among men. Interaction g; Alcohol Consumption, StressI Support and Symptomatolggyg The first four hypotheses called for an examination of the separate relationships between individual's and spouses' alcohol consumption, hassles, social support and symptomatology. However, the model outlined in the literature review describes these factors as interactive, and as impacting the adjustment of an individual. The final hypothesis is based on this interactive model and predicts that differences in lifetime alcohol involvement, social support, and hassles, in individuals and their spouses, will influence symptomatology in individuals. This hypothesis was tested using a series of hierarchical multiple regression analyses with measures of physical, depressive, and anti-social symptomatology being used as the dependent variables. As illustrated in the model (Figure 2, page ), it was predicted that drinking involvement would have a 115 direct effect on hassles and social support; social support, spouse's hassles, and spouse's symptoms would have a direct effect on hassles; and hassles would than directly impact an individual's symptoms. Therefore, the subject's lifetime alcohol problems score, his level of hassles, his social support, his spouse's lifetime alcohol problems score, his spouse's level of depression, and his spouse's hassles were used as independent variables, in this order, in the regression equations. (Depression was used as the spouse symptom measure because it appeared to correlate most highly with the other spouse symptom measures, making it the strongest variable to represent spouse's symtomatology.) Separate regressions were run for each of the following dependent variables: physical health, self-reported depression, interview-rated depression, adult anti-social behavior. The results for both men and women are presented in Table 10. In general, the findings indicate clear sex differences in the relationships of the above factors and in their predictive value of adjustment. Regarding predictive value, the total variance accounted for in the equations predicting men's symptomatology ranged from .35 to .73. The total variance accounted for in the regression equations predicting women's symptomatology ranged from .06 to .56. Overall, men's symptomatology tended to be better predicted by the independent variables Table 10 Hierarchical Regressions Predicting Men's and Women's Symptomatology (N-30) Men's Physical Illness 2 2 Significance Independent Step A Change Ag 3 (g; Change Variable Beta a l .01 .01 .57 LAPS .11 2 .31 .30 .002 Hassles .56 Support 4 .31 .00 .89 Spouse's .02 LAPS 5 .43 .12 .04 Spouse's .42 Self-reported Depression 6 .51 .09 .05 Spouse's -.33 Hassles 116 Table 10 (cont'd.) Women's Physical Illness Step 1 B. .07 .24 .25 .28 .35 .35 2 Change AA A .07 .16 .02 2 Significance Independent 117 2; Change .15 .02 .47 .34 .13 .94 Variable Beta LAPS .27 Hassles .41 Social -.13 Support Spouse's .17 LAPS Spouse's .36 Self-reported Depression Spouse's .02 Hassles Table 10 (cont'd.) Step A Change A_’A .32 .27 .02 .04 .08 .00 2 Significance Independent 2; Change .001 .000 .25 .11 .01 .87 Variable Beta LAPS .56 Hassles .54 Social -.15 Support Spouse's .19 LAPS Spouse's .35 Self-reported Depression Spouse's .02 Hassles 2 Significance Independent .03 .14 Men's Self-Reported Depression 2 l .32 2 .59 3 .61 4 .65 5 .73 6 .73 Women's Self-Reported Depression 2 Step R Change Ap R 1 .ll 2 .29 3 .39 4 .42 5 .56 6 .56 .00 118 9A Change .07 .02 .05 .25 .01 .98 Variable Beta LAPS .34 Hassles .42 Social -.33 Support Spouse's .18 LAPS Spouse's .52 Self-reported Depression Spouse's .00 Hassles Table 10 (cont'd.) Men's Interview Rated Depression 2 2 Significance Independent Step A Change Ap’A 9; Change Variable Beta 1 .11 .ll .08 LAPS .33 2 .34 .23 .005 Hassles .49 3 .34 .006 .62 Social .08 Support 4 .34 .000 .91 Spouse's .02 LAPS 5 .34 .00 .85 Spouse's -.04 Self-reported Depression 6 .35 .01 .55 Spouse's -.ll Hassles Women's Interview Rated Depression 2 2 Significance Independent Step A Change Ap A {pg Change Variable Beta 1 .02 .02 .43 LAPS .15 2 .03 .01 .62 Hassles .10 Support 4 .05 .01 .65 Spouse's .10 LAPS 5 .05 .00 .71 Spouse's -.10 Self-reported Depression 6 .06 .01 .63 Spouse's -.14 Hassles 119 Table 10 (cont'd.) Men's Adulthood Antisocial Behavior 2 2 Significance Independent Step A Change Ap A 9; Change Variable Beta 1 .37 .37 .00 LAPS .61 2 .37 .00 .76 Hassles .05 3 .46 .09 .05 Social .30 Support 4 .49 .03 .26 Spouse's -.l7 LAPS 5 .49 .00 .78 Spouse's .05 Self-reported Depression 6 .49 .00 .80 Spouse's -.04 Hassles Women's Adulthood Antisocial Behavior 2 2 Significance Independent Step A Change in A g; Change Variable Beta 1 .22 .22 .01 LAPS .46 2 .28 .06 .14 Hassles .25 3 .29 .01 .60 Social -.09 Support 4 .36 .08 .10 Spouse's .28 LAPS 5 .36 .00 .82 Spouse's .05 Self-reported Depression 6 .36 .00 .88 Spouse's .03 Hassles 120 Table 10 (cont'd.) Note: a=The Lifetime Alcohol Problems Score, the measure of lifetime alcohol involvement. 121 122 than was women's symptomatology, though not at a statistically significant level. A summary of the amount of variance accounted for in predicting each of the dependent variables is presented in Table 11. (A summary of the significant independent variables for predicting each symptom may be found in Table 12.) More specifically, for men, each of the predicted symptoms have multiple determinants. Physical illness was best predicted by men's own level of hassles, spouse's level of depression, and spouse's level of hassles. That is, physical illness was high when self-hassles and spouses' depression were high and spouses' hassles were low. However, men's level of self-reported depression is high when their own lifetime alcohol involvement and level of hassles were high, along with high levels of self- reported depression in their spouses. The prediction of interview-rated depression is quite similar, with high levels of depression again predicted by high levels of LAPS and hassles. Lastly, high levels of antisocial behavior in men are predicted by high LAPS and high levels of social support. In sum, hassles, LAPS, and spouse's depression turn out to be major variables in predicting adjustment in men. For women, only self-reported depression is predicted by multiple factors. These included women's own lifetime alcohol involvement, hassles, social support, and spouse's Table 11 Percent g; the Variance Accounted Men's and Women's Symptomatology Predicted Symptom Physical Illness Depression (Self-reported) Depression (Interview-rated) Antisocial Behavior in Adulthood 2 for When Predicting 2 3 Men Women (N=30) (N=30) .51 .35 .73 .56 .35 .06 .49 .36 Note: R is derived from hierarchical regression analyses with the predicted symptoms functioning as the dependent variables. 123 124 depression. That is, self-reported depression is high when LAPS, hassles, and spouse's depression are high and social support is low. In contrast, none of these, or any other independent variables, help predict interview-rated depression. Again, this supports the belief that self- reported and interview-rated depression measures are actually tapping different areas of behavior for women. The only significant predictor of physical illness is women's own hassles, and the only significant predictor of antisocial behavior is women's own lifetime alcohol involvement. In these latter two relationships, high levels of one variable are associated with high levels of the other variable. Again, hassles and LAPS turn out to be the more frequent and stronger variables in predicting adjustment. Summarizing the results for men and women, LAPS and hassles and up as the most frequent variables in predicting adjustment for both sexes. However, LAPS is a significant factor in estimating both internalizing and externalizing symptoms while hassles is only influential in estimating internalizing symptoms. This is true for both sexes. Spouses' depression is another significant factor in predicting men's and women's internalizing symptoms. Each of the above variables have similar effects on adjustment with high levels of drinking, hassles, and spouses' depression associated with higher 125 symptom levels. In contrast, social support has opposite effects for men and women, with increases in support predicting antisocial behavior in men and decreases in support predicting self-reported depression in women. A more detailed comparison of the factors contributing to the prediction of various symptoms for each sex can be made by reviewing Table 12. Most noteably, a man's level of physical illness is predicted by his level of hassles and by his wife's level of depression and hassles; whereas, a woman's level of physical illness is not predicted by her husband's state. It is also noted that none of the variables included in these analyses aid in the prediction of women's interview-rated depression, while LAPS and hassles account for 34% of the variance in men's current interview-rated depression. To follow up these hierarchical analyses, a set of stepwise regressions were performed to determine if the model described earlier did indeed represent the best fit of the independent variables in predicting physical health, depression, and anti-social behavior. These results are presented in Tables 13 and 14. Briefly, LAPS, hassles, and spouses' depression are still the significant factors in the prediction of adjustment but their relative importance is shifted slightly for certain symptoms in the stepwise analyses. Table 12 Summagy 25 Independent Variables Contributing gAgnificantly pg the Prediction 2; Men's and Women's Symptomatology \ Men WOmen (N330) (N=30) significant Significant Independent Independent Predicted Symptom Variable's Variable's Physical Illness Hassles Hassles Spouse's Depression Spouse's Hassles Self-reported LAPSa LAPS Depression Hassles Hassles Spouse's Depression Social Support (-) Spouse's Depression Interview-rated LAPS Depression Hassles Adulthood LAPS LAPS Antisocial Social Support (+) Behavior Note: The independent variables listed here are all significant at the p .05 level or better, and are listed in order of decreasing importance. a=Lifetime Alcohol Problems Score, the measure of lifetime alcohol involvement. 126 Table 13 Stepwise Regressions for Predicting Men's Symptomatology (N830) Physical Illness 2 2 Significance Independent Step A Change 12.3 g; Change Variable Beta 1 .31 .31 .001 Hassles .56 Self-Reported Depression 1 .42 .42 .000 Spouse's .65 Self-reported Depression 2 .60 .18 .002 Hassles .45 a 3 .72 .12 .002 LAPS .37 Interview Rated Depression 1 .29 .29 .002 Hassles .54 Adulthood Anti-social Behavior 1 .37 .37 .000 LAPS .61 a=The Lifetime Alcohol Problems Score, the measure of lifetime alcohol involvement. 127 Table 14 Stepwise Regressions for Predicting Women's Symptomatology (N830) Physical Illness 2 2 Significance Independent Step A Change Ap_A 2; Change Variable Beta 1 .25 .25 .005 Spouse's .50 Self-reported Depression Self-Reported Depression 1 .42 .42 .000 Spouse's .65 Self-reported Depression Interview Rated Depression 1 e -- 0-- a--- ------- e -- Adulthood Antisocial Behavior a l .22 .22 .01 LAPS .46 a=The Lifetime Alcohol Problems Score, the measure of lifetime alcohol involvement. 128 129 More specifically, for men, hassles maintains its position as the single strongest predictor for physical illness. LAPS is still the strongest single predictor of antisocial behavior and, hassles still accounts for the largest amount of the variance in predicting interview- rated depression. However, each of the independent variables cited are the only significant variables in predicting their respective symptoms; there are not multiple determinants for physical illness, interview- rated depression, or antisocial behavior. Self-reported depression still had multiple predictors but with slightly different weights than in the hierarchical regression model. That is, in the stepwise regression, wives' depression has the largest beta weight (.65) with hassles (Beta-.45), and LAPS (Beta-.37), following in the second and third positions. For women, LAPS maintains its place as the single significant predictor of antisocial behavior and, again, no variables are found to predict interview-rated depression. However, the prediction of physical illness and self-reported depression changes in the stepwise regression analysis. Specifically, husband's depression is the only significant predictor of women's physical illness gpg women's self-reported depression. The effects of hassles, LAPS, and social support drop out, attesting to the high power of this factor that appears to be under- 130 represented in the originally constructed model. In sum, the model presented in this study, and tested in the hierarchical regressions, appears to more descriptive of men than of women. These data also indicate that the importance of the specific independent variables varies somewhat when predicting different types of symptoms. To briefly summarize the above results is impossible; yet, what has become clear is that current drinking and lifetime alcohol involvement have quite different relationshipts to hassles and symptoms and, the interaction of drinking, hassles, support, and symptomatology is quite different for men and women. A discussion of these generalizations and of the more specific findings is presented in the following chapter. 131 Chapter 4 Discussion The results presented in the previous chapter are numerous and fairly complicated. Some of the data ended up supporting the hypotheses that had been put forth while other data suggested new ways of conceptualizing heavily drinking couples. This discussion will attempt to systematically review the results and attempt to set them in the context of recent work on drinking behavior and alcoholism. In this vein, it is also important to note that the results from the present study differ from other work on two significant points. First, because the current work was designed as a study of a high-risk population, the present sample was selected so that it adequately covers drinking levels at the high end of the drinking continuum. Therefore, this sample contains individuals with high levels of alcohol involvement and high levels of symptomatology. That is, this is not a sample of convenience but, a population sample that is representative of heavy to very heavy drinkers. As such, results from this study are not generalizable to normal, moderate drinkers with low levels of trouble. The findings are more relevant to families at clinical risk who are experiencing high levels of stress early on in 132 their development. However, because of the special way the sample was selected, results are potentially more generalizable than could be the case for a treatment sample with its attendant self selection biases. This type of sample does not coincide with samples from many of the studies reviewed in this dissertation, limiting the similarity (but not validity) of the results between the present study and other work. Second, the present sample is homogeneous with regard to family stage, thereby eliminating any variation of drinking, stress, social support, etc., that would be confounded by this factor. The majority of studies reviewed earlier and presented below made little attempt to control for the families' stage of development and, as a result, would be expected to show different, and possibly confused patterns of relationships between variables and stages of development. For example, the study by Pachman and Foy (1978) looked at the symptoms of anxiety and depression in alcoholics whose marital status broke down as follows: married-49%, separated-18%, divorced-21%, never married-11%, widowed-1%. The majority of these subjects also ranged in age from 35-57. Both age and marital status can be expected to affect the stresses an individual experiences, the tasks he is involved in, and the other individuals that he interacts with (Steinglass, 1980). Similarly, work in this area suggests 133 that age, as it reflects years of abusive drinking, will affect the physical health and type of alcohol related symptoms an individual experiences (Jellinek, 1952; Bacon, 1973; Mulford, 1977). Over time, an alcoholic is likely to progress through different stages of alcoholism. Jellinek (1952) originally described these stages as prodromal, crucial, and chronic and identified them by the appearance of various types and intensities of symptoms. In sum, developmental stage is likely to influence the relationship between drinking, stress, symptoms, and support; as stages change, these relationships may change. The present work addresses this issue more definitively, albeit within one particular stage of one family life cycle. Hassles and Alcohol Involvement: Returning to the current work, the first hypothesis predicted that overall hassles and interpersonal hassles would be higher in individuals with high versus low lifetime alcohol involvement. The data do not support this hypothesis. At best they suggest that a positive relationship might be present between women's interpersonal hassles and their own long term history of heavier alcohol involvement (the relationship is at the trend level). This association must be confirmed with a larger sample, but more generally, the hypothesis is not supported. In fact, a negative relationship between 134 drinking and hassles is observed when current drinking is examined. Specifically, men report experiencing significantly fewer hassles when their current level of drinking is higher, and tend to experience fewer hassles when their spouse is currently drinking more. And, women report experiencing significantly fewer hassles when their husbands are currently drinking more. This lack of a significant relationship between LAPS and current hassles and the unexpected negative relationship between current drinking and current hassles may be due to three different alcohol-related phenomena. First, the heavy drinkers and the individuals with high lifetime alcohol problem scores may be denying the occurrence of negative or unpleasant events, particularly if they are related to that individual's drinking (Bacon, 1973; Royce, 1981). Second, the negative relationship between current drinking and hassles may be due to the pleasant arousal effects of alcohol. Alcohol consumption and its related affects may serve to block out unpleasant stimuli so that, while stress exists, it is not perceived as stress. This buffering phenomenon was examined by Allan Williams (1966) when he studied the effects of alcohol consumption on anxiety and depression in a sample of male college students that included both problem and non-problem drinkers. Anxiety and depression were measured prior to a 135 stag cocktail party, after subjects had consumed 4 oz. of alcohol during the party, and at the end of the party. Results showed that both anxiety and depression were initially higher for the problems drinkers but decreased for both problem and non-problem drinkers after they ingested 4 oz. of liquor. This decrease occurred at the same rate for both groups. However, anxiety and depression began to increase for both groups after they consumed 8 oz. of liquor and more. Williams concluded from these results that alcohol has a dosage effect; "with increasing dosage levels, the anesthetic effect builds, resulting in a progressive impairment of function,” (Williams, 1966; p. 692). This impairment is likely to cause anxiety and depression. Williams also concluded that the relatively high levels of anxiety and depression in problem drinkers would create a stronger motivation for them to drink, and temporarily reduce their negative feelings. He goes on to state that this condition may have contributed to this group becoming problems drinkers and may lead then to becoming alcoholics. Recent research in the area of alcohol use and stress also supports the idea that alcohol can reduce anxiety. Sher and Walitzer (1986), found that both subjects' heart rate and anxiety responses to a stressful social situation were reduced more when male drinkers ingested more alcohol. They noted that this change was not mediated by 136 self-consciousness or expectancies for tension-reducing alcohol effects; this implied that alcohol mediated the stress, with drinking subjects no longer reacting to the situation as stressful, at least via heart rate and anxiety responses. Another recent study also supports the stress-reduction effects of drinking, particularly when it is combined with other distracting stimuli (Steele et al., 1986). This study had a 2 by 2 design which included subjects who had, or had not, received alcohol. Respondents rated pleasant art slides or did nothing, after receiving negative feedback from an IQ test (the stressful event). Results showed that subjects' mood improvement was greatest in the alcohol/slides condition with alcohol/no slides also yielding improved mood, again implying that alcohol consumption improves mood. Lastly, a study by Collins and Marlatt (1985), suggests that individuals will increase alcohol consumption as a strategy for coping with aversive social interactions. This again is consistent with the notion that a reduction in stress occurs with drinking. Thus, one would expect stress to be reduced as current drinking increases. Regarding the significant association between lower levels of wives' interpersonal hassles and higher levels of husband's current drinking level, it may be that a husband's current drinking distracts and mobilizes his wife so that she perceives fewer hassles in her own life. 137 By this arguement, it would also follow that a wife's current depression would decrease as her spouse's current drinking increases. There is a trend in the data supporting this; wifes' depression is negatively associated with husbands' QFV (Table 5). This same process also may be operating for men. Their level of interpersonal hassles shows a negative trend when related to wives' current drinking; their hassles are lower when their wives' current drinking is higher. More generally, these findings provide some support for the hypothesis that spouses' current drinking serves to distract and mobilize an individual so that they perceive fewer personal stresses. In sum, lifetime alcohol involvement showed a low order positive relationship with overall hassles while current drinking showed a negative relationship with hassles. Current drinking may act as a temporary reliever of stress but heavy drinking over an individual's lifetime tends to increase the level of overall stress experienced. These notions will need further testing in other research. Finally, there is another alternative; findings may be developmentally specific to the fairly young families studied here (mean age around 30, with young children). Steady heavy alcohol involvement may have a differential relationship to hassles at later stages of the life cycle. Symptom Occurrence and Alcohol Involvement: The first part of hypothesis two, predicting a relationship between high lifetime patterns of alcohol involvement and high levels of antisocial behavior in individuals, received strong support in the data for both men and women. It also appears, unpredicted by the hypothesis, that husbands' lifetime alcohol involvement is positively related to wives' adulthood antisocial behavior and, wives' current drinking is positively related to their husbands' antisocial behavior. These findings clearly link drinking and antisocial behavior both within individuals and within couples. This co-occurrence of heavy drinking and antisocial behavior in individuals is well documented in past research (Zucker and Gomberg, 1986), with the current data supporting this co-occurrence in couples. That is, the data show that heavy drinking individuals are antisocial and are married to heavy drinking, antisocial spouses. Further support for this link between antisocial behavior and drinking in couples is found in Reider's research on drinking and family violence on this same sample (1987). Her results show that increased alcohol involvement is significantly related to higher levels of individual's and spouse's aggression toward their children and their mate. Again, drinking in one spouse is related to antisocial behavior of the other. However, if there is any direction of 138 139 effect in this linkage within the marriage, it will have to be explored in future research. An individual's drinking may lead a spouse to act out, or an individual's antisocial behavior may lead a spouse to drink more heavily. The second part of the hypothesis, that predicted a positive relationship between spouse's level of lifetime alcohol involvement and an individual's level of depression and physical illness, received weak support among the women. Higher levels of husbands' lifetime alcohol problems tend to be related to higher levels of wives' health problems and wives' self-reported depression. However, interestingly, husband's level of current drinking shows a tendancy to be negatively related to their wife's level of physical health and current depression. Again, it appears that wives' higher level of adjustment is related to their husbands higher level of current drinking. As stated earlier, this may be because the wives are mobilized by their husband's current drinking and perceive their own problems as diminishing. However, in the long run, husbands' heavier lifetime drinking appears to lead women to develop more internalizing problems. This is another causal hypothesis that also needs to be tested with longitudinal data, though, the relationship between current drinking and a spouse's diminished stress and symptoms has also been 140 found in other cross-sectional work (Jacob et al., 1986). Jacob and his colleagues studied drinking and adjustment in intact families containing an alcoholic father and at least one child between 10 and 18 years of age (Jacob et al., 1986). He found that alcoholic husbands who were steadily consuming large amounts of alcohol in the home had wives who (1) had relatively low scores on certain MMPI scales (2) had relatively low scores on the Beck Depression Inventory, and (3) reported relatively higher levels of marital satisfaction, when compared to non-alcoholic couples. Weaker relationships were found between these variable for out-of-home and binge drinkers. Based on these findings, Jacob also hypothesized that family stress is minimized during periods of high alcohol consumption and marital/family satisfaction is heightened, thus reinforcing drinking behavior. This breakdown of drinking into steady or binge drinking also helps explain what had previously been viewed as conflicting results between Steinglass' 1981 and 1982 drinking studies. In the first study, Steinglass found a decrease in effective communication and positive affect in wet versus dry alcoholic families. In his 1982 study, Steinglass found increases in communication and positive affect in wet versus dry alcoholic families. Jacob (1986) also studied these factors in dry and wet 141 conditions but divided his sample of drinking families into the steady and binge drinking groups mentioned above. Analyses on this group x drinking interaction indicated that binge drinkers showed higher rates of negativity than steady drinkers, binge drinkers were less task-focused in problem solving than steady drinkers and, steady drinkers became more positive in drinking versus no-drinking conditions. Binge drinkers showed little change in positivity (Jacob, 1986). In sum, current drinking may also reduce symptomatology and stress by contributing to a more positive family atmosphere. It is also noted that, while there is a trend for a positive relationship between husbands' drinking and wives' internalizing symptoms in the current work, there is no such relationship between wives' drinking and husbands' internalizing symptoms. The present data suggest that wives are more interpersonally tied to their husbands than vice versa. However, this might change over time as the family adapts to the husband's drinking. Work in this area suggests that families of alcoholics go through stages of adaption which progress as follows: denial and minimizing; tension and social isolation; frustration and disorganization; attempts to reorganize- shifts in roles; separation-escape; reorganization without the alcoholic; and recovery and reorganization with the alcoholic (Royce, 1981). It is clear that as these stages 142 progress and drinking persists, the wife becomes increasingly detached from her husband. While the above explanations are interesting, they are only based on trends. The lack of a strong relationship between health and spouses' drinking may be attributable to other factors. Stage again becomes a factor, in that participants in this study may be young enough such that heavier alcohol consumption has not yet had a sufficient cumulative impact on their physical status. As already noted, individuals in this study average 30 years of age and, if alcoholic, are likely to fall into an early phase of the addiction process. They experience a variety of drinking related physical symptoms, such as blackouts, but not at the intensity or frequency of older, later stage alcoholics (Mulford, 1977). In this vein, it is important to underscore that measures of current drinking (QFV) and lifetime alcohol problems (LAPS) consistently produced different patterns of findings. The positive relationships between gp individual's lifetime alcohol problem score and both internal and external symptomatology are always stropggg gAgp Ag the relationship between gp individual's current drinking and symptoms. These data suggest that, with regard to symptomatic status, drinking has a more powerful effect over an extended period of time; a person has to 143 drink heavily over a long period before it adversely effects adjustment and symptoms. This finding is also compatible with the developmental theories of alcoholism which describe alcohol addiction as occurring in phases (Jellinek, 1952; Bacon, 1973; Park, 1973; Mulford, 1977). Again, these phases are identified by the number, type, and seriousness of alcohol related problems that an individual is experiencing. As drinking persists, these problems increase and worsen, causing deterioration of individual's adjustment in physical, emotional, and social spheres. As described earlier, another aspect of this process may be that current drinking has a buffering effect, either by altering the perception of problems so that they are not reported as such or, by increasing the denial of problems (Williams, 1966; Sher and Walitzer, 1986; Steele et al., 1986). Either of these phenomena would strongly influence self-report data. In sum, within individuals, men's high levels of lifetime alcohol involvement and current drinking are positively related to high levels of antisocial behavior. Lifetime alcohol involvement is also positively related to men's level of current depression. For women, a high LAPS is also positively related to high antisocial behavior and, LAPS is related to current depression and physical illness. Generally speaking, heavy drinking Ag related pg 144 poor adjustment Ap both men and women, althougA Egg intensity 23 this relationship varies across ggAA Across individuals, husbands' lifetime alcohol involvement and wives' current drinking are both positively related to their spouses' antisocial behavior; the co-occurrence of antisocial behavior and drinking within individuals also exists as a phenomenon in couples. In addition, husbands' lifetime drinking tends to be positively related to wives' physical illness and depression, while husbands' current drinking tends toward a negative relationship with wives' physical illness and depression. These contrasting results may indicate that men's current drinking temporarily mobilizes their wives so that their symptoms decrease but, as this heavy drinking persists over time, women are adversely affected. Social Support and Alcohol Involvement: The third hypothesis predicted that individuals with high lifetime levels of alcohol involvement, and their spouses, would have less social support than individuals with low lifetime alcohol involvement and their spouse. This hypothesis does not receive support from this study. In fact, there appears to be a low order positive relationship between high levels of lifetime alcohol involvement and high levels of social support. Similarly, there is a low order positive correlation between high 145 levels of current drinking and high levels of social support. If these findings are accurate and not a product of type II error, this lack of support for the predicted hypothesis and tentative support for the opposite of the hypothesis may be a function of the effects of alcohol consumption in certain phases of the family's functioning. Early on in these young families, drinking may be an extraversive activity; increased drinking has been shown to occur in social situations where others are drinking heavily (Collins et al., 1985). Since drinking may be part of a social interaction, perceived social support may increase. Also because of being in an early drinking stage, individuals may reach out more when problems occur, before major deterioration of the marriage (Royce, 1981). And, the initial reaction of others, including spouses, may be one of support. However, as abusive drinking persists over time, family members, co-workers, and friends become more alienated and the drinker becomes more isolated, (Paolino and McCrady, 1977; Cronkite and Moos, 1984). Again, this fits with the developmental theories of alcohol addiction where early phase alcoholics are seen as still drinking in social contexts and as experiencing less social isolation and hostility than are later phase drinkers (Jellinek, 1952; Bacon, 1973; Mulford, 1977). In light of these other studies, the current results showing 146 increases in support with drinking may be indicative of an early phase in a developmental process in heavy drinking families. Differences between these results and those of previous research may be explained by the fact that the majority of previous research that finds a signficant relationship between heavy drinking and decreased support (Jackson, 1954; Rhoads, 1983) has focused on already identified alcoholics and drug abusers who were in treatment, and on families that are considerably further along in their marriage and in parenting. According to Williams (1966), this type of identification usually occurs in individuals who are older and have admitted they cannot control their drinking and need help. "This is a crucial, psychologically meaningful admission in the life of an alcoholic, and those who take this step may have different personality characteristics from those who do not,” (Williams, 1966; p 245). Also, studies with older samples of alcoholics cannot "...dismiss the possibility that their findings reflect social and psychological consequences of 15 to 20 years of excessive drinking and loss of control over alcohol,” (Williams, 1966; p 246). An example of research done with this type of sample is Steinglass' study on interaction in alcoholic families (1981). His subjects were described as middle-aged (x-4s years old), well into their second decade of marriage, and 147 having a minimum of 5 years of alcoholic drinking history. In contrast, the respondents in the current study are, on the average, in their early 30's, and have not been self or other selected by the social system as alcoholics. A second possibility is that the lack of significant results is better explained by way of the poor state of methodology in social support research. This issue is discussed in more detail below, in the section on mea- surement issues. However, it is interesting to note here that researchers are finding the whole range of negative and positive correlations between various factors and social support, raising the question: is social support correctly conceived? And, is it functioning the way we think it does (Heller, 1986)? We think of all support as positive, but it may be that certain types of support have negative effects on individuals. An illustration of the above theoretical problem and a relevant example for the current research is discussed by Coyne and DeLongis (1986). Using their own data, they comment on the dual supportive and stressful relationships of marriage: "Although married persons generally have numerous psychological and social advantages over unmarried persons, much of this may be limited to happily married persons. There is some evidence that unhappily married people are worse off than unmarried people in terms of physical health and psychological well-being. People who are unhappily married report more physical illness and depression, heavier drinking, and more isolation outside of their marriage than do happily married people," (Coyne and DeLongis, 1986; p 455). 148 Clearly, having a spouse in a support network can be stressful, which could also be true of other network members. When determining what or who is actually "supportive", it may therefore be necessary to go into different and greater detail about the relationship between two people than has been done in the present study. Thus, an examination of what is unsupportive about a relationship would be particularly important to examine in heavy drinking or alcoholic families, where many marriages are known to be conflicted (McCrady, 1982; Reider, 1987), and where other friends and family listed in a social network may be in conflict with the drinking individual. Evidence for a dual supportive and stressful relationship in EAAg study comes from a tally of questionnaire responses, which showed spouses listed as a support person 100% of the time on the Social Support Questionnaire, yet frequently described as a source of stress on the Hassles scale. The Interaction 2; Stress and Symptomatology Ap Couples: The purpose of this phase of the analyses was to extend the well established link between stress and adjustment within individuals to the interaction of stress and adjustment in couples. With this in mind, the fourth hypothesis predicted that the stress and symptomatology of one member of a couple would be related to the stress and symptomatology experienced by the other member. In fact, 149 the data here continue to (1) support the presence 9; Egg stress-illness connections for individuals and (2) gAyg clear support pg their interaction within couples. Again, there are clear differences in these relationships by sex and by type of symptom (i.e. internalzing and externalizing symptoms). Within both husbands and wives, daily hassles are significantly and strongly related to internalizing symptoms; in neither sex are hassles systematically related to externalizing symptoms. The present results thus show that the relationship between stress and symptoms is not an all encompassing one, but must be examined by the type of the symptomatic display being examined. This pattern of relationships is also seen in couples. Again, among both spouses, an individual's hassles show a positive relationship to their spouse's internalizing symptoms, but not to the spouse's externalizing symptoms. similarly, we see a varied pattern of results when the relationship of husbands' and wives' symptoms are examined (Tables 8A, 8B, 80, 8D). When the relationship between spouses' internalizing symptoms is examined (Table 8A), strong positive relationships are evident between husbands' and wives' self-reported depression and, between spouses' depression and physical illness. In contrast, the relationship between husbands' and wives' 150 externalizing symptoms is only significant when husbands' total antisocial behavior and wives' adulthood antisocial behavior are compared. There is a strong trend towards significance for the remaining externalizing variables. When relationships across types of symptoms are examined, pgpg 2A the correlations between husbands' externalizing Symptoms gpg wives' internalizing symptoms 35g significant (Table 88). Whereas, wives' externalizing symptoms of childhood and adulthood antisocial behavior are positively related to husbands' internalizing symptoms of current depression and physical illness. In other words, husbands' and wives' adjustment are strongly linked pg each other butL the role that spouse's adjustment plays varies py Egg ggA g; the spouse. The role relationships are different in these marriages. To summarize up to this point, the results indicate that the interactions between drinking, stress, and symptomatology need to be examined on a variety of levels. Current drinking and lifetime alcohol involvement show different patterns of relationship to the individuals' level of stress, and to their levels of internalizing and externalizing symptoms. In addition these relationships vary with the sex of the individual. More specifically, for men, higher lifetime alcohol involvement is strongly related to AAgher levels 9A current depression, and childhood and adulthood antisocial 151 behavior. Heavier current drinking is associated with Agygg levels of hassles perceived, and to greater amounts of childhood and adulthood antisocial behavior. Hassles are related to more depression and physical illness. In other words, lifetime alcohol involvement Ag positively related pg both internalizing and externalizing sypptoms yAAAg current drinking Ag positively related pp externalizing symptoms. Daily hassles are positively related to internalizing symptoms. Thus, heavier lifetime and current drinking may cause more stress and poorer adjustment. However, heavier current drinking may also act as a buffer, reducing levels of stress and internalizing symptoms. Higher levels of current drinking are related to lower levels of hassles which, in turn, are related to internalizing symptoms. For women, lifetime alcohol involvement is strongly related to depression and adulthood antisocial behavior and, tends to be positively associated with daily hassles, physical illness and childhood antisocial behavior; current drinking is not significantly associated with any of these variables. Hassles are positively related to physical illness and current depression and, at the trend level, with adulthood antisocial behavior. Therefore, like men, women's lifetime alcohol involvement is positively related to both internalizing and externalizing symptoms. Yet current drinking is not related to women's 152 symptomatology. Daily hassles are positively related to both internalizing ggg externalizing symptoms and again, lifetime alcohol involvement is related to increases in interpersonal hassles for women. It seems possible that heavier lifetime alcohol involvement may cause more hassles and poorer adjustment in women, and these outcomes are unaffected by current drinking level. The relationships between drinking, symptoms, and hassles continues to vary across the sexes when the influence of spouses' variables are considered. A wife's current drinking Ag related, gg Egg ggggg level, pg Agggg levels 9A hassles for her husband and, pg gAgher levels 9A antisocial behavior Ag her husband. On the other hand, gAgher levels 2; wives' hassles are related pg higher levels 2; men's internalizing symptoms 2; physical illness and current depression. Husband's current drinking also shows a negative relationship with women's hassles and, is negatively related to women's internalizing symptoms of physical illness and current depression. However, unlike wives' LAPS, husbands' lifetime alcohol involvement Ag positively related pg gggAg wives' physical illness ggg current depression. Finally, like men and their spouses, higher levels of husbands' hassles are related to higher levels of women's current depression. When examining additional non-drinking variables, husbands' and wives' hassles are positively related to each other, as are 153 spouses' levels of physical illness, current depression, and antisocial behavior. Women's antisocial behavior is also positively related to their husbands' levels of physical illness and current depression. In brief, husbands' and wives' levels of drinking, symptoms, and hassles take on both enhancing and buffering effects when related to their spouses' drinking, symptoms, and hassles. At this point we have a model where heavier current drinking Ag related pg gAgher levels 2; antisocial hassles and internalizing symptoms Ag men and women. Hassles are positively related to internalizing symptoms, but heavier current drinking is associated with lower levels of hassles, in both individuals and across spouses. Heavier current drinking in men is also associated with lower levels of internalizing symptoms in women. However, as heavy alcohol use persists and alcohol related problems accumulate into high levels of lifetime alcohol involvement, drinking becomes positively related to both internalizing and externalizing symptoms, and tends to be related to higher levels of overall hassles. This is true for both men and women and, across spouses. Last, men's and women's level of adjustment is also strongly related to their spouses' level of adjustment. Interaction g; Drinking Behavior, Stress, Support, and Symptomatologyg This research also examined the inter-relationships of the above factors by way of regression analysis, anticipating that differences in individuals' and spouses' lifetime alcohol involvement, social support, and hassles would be related to differences in the person's symptomatic outcome. Differences in the data were clearly apparent and may be summarized as follows: Overall (1) lifetime alcohol involvement Ag the most significant factor Ag predictingi current self-reported depression and adulthood antisocial behavior Ag both men and women; (2) Ag Ag also the major predictor 2; current interview-rated depression Ag men. (3) Hassles Ag the stroggest predictor 2; physical illness and the second strongest predictor, next pg LAPS, 2A current self-reported depression. This Ag ggain true for both men and women. (4) Increases Ag AgygApgA both LAPS and hassles predict increases Ag AgygA pg Symptomatology. However, after the symptom variance due to LAPS and hassles is accounted for, the major predictors for men's and women's symptoms began to vary. These differences vary symptom by symptom and are 4It is important to underscore that the term predictor is used here in an associational sense and not a time-lagged sense. Ultimately the longitudinal data will have to be used to establish order of precedence of effect, and of genuine "predictors". 154 155 summarized in Table 12. Briefly, social support shows itself to be operating in opposite ways for men and women. Higher levels of support are predictive of higher levels of antisocial behavior in men, while lower levels of support are predictive of higher levels of current self- reported depression in women. Another difference between the sexes is the occurrence of spouse's depression and spouse's hassles as predictors of men's physical illness but not of women's physical illness. In sum, the regression analyses lent support to the earlier correlational findings: LAPS is related to internalizing and externalizing symptomatology; hassles is related to internalizing symptomatology; spouses' symptoms helped predict a person's adjustment; and the pattern of relationships between drinking, hassles, and adjustment varies for men and women. Modifications and Additions pg the Model 9A Egg Stress- Illness Relationship Ag Drinking Couples: At the most general level, a review of the relationships among patterns of alcohol consumption, stress, support and symptoms indicates a need to revise the model. Such a model should clearly be constructed differently for men and women and for type of symptom being predicted. For example, the hierarchical regression analyses and earlier Pearson correlational analyses 156 indicate that drinking by the individual, and his or her spouse, has a different relationship to stress and symptom levels across the sexes. The stepwise regression analyses also indicated that the relative importance of these variables differs from the original model and, again differs between men and women. These sex differences also appeared earlier in the analyses when drinking levels and number of drinking related problems were explored (Figures 3, 4 and 5). Women in this sample drink less and have fewer alcohol related problems than men. This and other sex differences found between drinking and adjustment variables also appear reflective of more general population differences. For example, in a study of 795 men and women (Parker, Parker, Wolz & Harford, 1980), women drank less alcohol less often than men. The demographic factors influencing consumption also varied across sex with marital status and age showing a relationship to drinking patterns in men and educational background and employment status showing a relationship to drinking patterns in women (Parker et al., 1980). Antecedents for women's drinking also vary from men's and more frequently include escapism, feelings of powerlessness and inadequacy, and sex-role conflicts (Beckman, 1980; Scida & Vannicelli, 1979). To add to this, sex differences in emotional adjustment are also evident in the general population. For example, there are 157 clear sex differences in incidence of depression with women experiencing more major depressive episodes than men (18-23% for women versus 8-11% for men), and men experiencing a higher rate of antisocial personality disorder than women (3% versus <1%), (DSM-III, 1980). In other words, the sex differences that appear in the current work are important to understanding the environments in alcoholic and heavy drinking families, but they are also reflective of broader population differences between men and women. In addition, despite the strong predictive value of the variables outlined in the above regression analyses, the total amount of adjustment variance accounted for ranges from 35% to 73% for men, and from 6% to 56% for women. While this is a significant portion of the variance (for men), there is still variance that is unaccounted for. Based on the initial correlational analyses it seems likely that current drinking level may be one such additional element. The data show that current drinking often operates in an opposite manner than lifetime alcohol involvement when associated with hassles and symptoms. Higher current drinking level is significantly related to lower amount of hassles reported, and is less strongly related to severity of symptoms experienced when compared to lifetime alcohol involvement. Higher levels of spouse's current drinking are actually 158 significantly related to lower amounts of hassles experienced by men and women, and tend to be related to lower amounts of current interview-rated depression and health problems for women. In trying to understand the above results, the variable relationships of current drinking appear to be similar to the variable effects of wet and dry phases of drinking as described by Williams (1966), Sher and Walitzer (1986), Jacob (1986) and Steinglass (1981). For example, in his research, Steinglass described communication and relationship relevant messages (i.e., "healthy" interaction), as increasing when couples contained an actively drinking member versus when couples were in a dry, non-drinking phase. In other words, the current work and Steinglass' work lead to the conclusion that current drinking may be considered as a buffer in the stress-illness relationship in a drinking couple. In the current work, this buffer may function by increasing an individual's denial of problems, causing these problems to be perceived as "less problematic" or, by having a mobilizing effect on an individual's spouse, which in turn reduces her own symptomatology. This process may also be an earlier manifestation of the buffering effect of current drinking, since the sample in the present work is markedly younger in age and family stage than is the sample in Steinglass' work. 159 Another factor that may be important in accounting for level of adjustment in these couples is their place on the drinking continuum from early to middle, to later stage in the developmental process of alcoholism (Jellinek, 1952; Bacon, 1973; Mulford, 1977; Zucker, 1987). For example, in discussing the results of this study, it has been suggested that drinking in the early stages of alcoholism is associated with increases in perceived support. In this early phase, marital conflict would not be as high, so that spouses would still be supportive. In addition, the drinker's level of denial would not be as high so that he would be more likely to seek out support. However, as drinking becomes more identified as a problem and persists over a greater period of time, the nature of the relationship between social support and drinking would be expected to change. For example, among alcoholic populations in treatment settings, where individuals are typically in later drinking stages, marriages are quite conflicted and individuals are more isolated from friends and families (McCrady, 1982; Rhoads, 1983). From this perspective, the predictive model should vary along a continuum of drinking; stage of drinking is a group variable that should indicate where on the continuum a couple belongs. For the present sample, stage of drinking was not precisely identifed. That is, a range of heavy drinking 160 patterns was sampled and subjects were not chosen because of societal identification as alcoholics, implying that these families are in an early drinking stage, but no attempt was made to classify them as being in prodromal, crucial, or chronic stages of addiction (Jellinek, 1952). This distinction is important in that, as the alcoholic progresses through these stages, (s)he is experiencing an increasing number of drinking related problems which are likely to increasingly disrupt everyday functioning. If indeed the stress-illness relationship differs for these groups, failure to divide a sample in this way may wash out certain patterns in the data. It is also noted that a much larger sample that taps a wider range of drinking behavior is required to test this hypothesis. One more factor that may influence the impact of drinking and stress on adjustment in couples is the constellation of the couple involved. Couples can be composed of two heavy drinkers, two light drinkers, a heavy male drinker and a light female drinker or, a heavy female drinker and a light male drinker. The present results indicate that differences in stress and adjustment are related to differences in individual's and spouse's drinking patterns. This leads one to hypothesize that similar differences in individual adjustment appear as a couple's drinking pattern varies. Again, past research supports this idea with findings that show men are more 161 likely to use approach coping techniques in dealing with stress when their wives are the heavy drinkers. But, when they are the heavy drinkers, men are more likely to use avoidant coping techniques (Cronkite and Moos, 1984). A test of this additional hypothesis would require a different population than in the current work, to insure that couples of each drinking combination are represented. Last, the type of stress experienced needs to be accounted for when attempting to predict adjustment. As stated earlier, the pilot study for this project (Weil, 1984), used Holmes and Rahe's (1967) measure of major life events as an indicator of stress. With this measure, significant positive relationships were found between stress and antisocial behavior and, between stress and worst-ever depression. In the current study, stress was measured via the Daily Hassles Scale (Kanner et al., 1981). Increases in hassles were significantly associated with increases in current depression and with increases in physical health problems. A logical conclusion is that there are multiple types of stressors, and that each type may have its own unique or shared impact on adjustment. The continuing argument about what kind of event is stressful becomes a moot arguement; different types of events are stressful in different ways. Along this line, it would also be interesting to determine if uplifts (i.e. the opposite of hassles), or 162 desirable life events, have a different relationship to adjustment than do hassles. However, past research has shown little support for a direct relationship between pathology and positive change. For example, Sarason et a1. (1978), found that extraverted expression was related to positive life changes but maladjustment was related to negative changes. Work by Fontana et al. (1979), found that outside social participation was related to perceived desirability of events but again, perceived undesirability of events was related to psychological disturbance but not to social participation. Lastly, Kanner et a1. (1981), also found that daily hasssles were better predictors of symptoms than were daily uplifts. From these findings and from the present study, it may be concluded that uplifts or desirable events do not have a direct effect on maladjustment but they may influence factors that do, such as outside social participation or extraverted expression. Therefore, uplifts and related phenomena may eventually be important factors in completing a comprehensive model of the stress-illness relationship. I Incorporating these ideas, a revised model would need to be developed for internalizing and externalizing symptoms , and separately for men and women. For men, internalizing symptoms would be predicted by men's daily hassles, lifetime alcohol involvement, current drinking, spouse's current drinking, and spouse's level of 163 internalizing and externalizing symptoms. The hypothesized direction of effects is outlined in Figure 7. In this model there are both direct and indirect effects, with the latter taking on both enhancing or buffering functions. Externalizing symptoms for men would be predicted by lifetime alcohol involvement, current drinking, social support, major life eventss, and spouse's externalizing symptoms. These relationships are diagrammed in Figure 8. For women, internalizing symptoms would be predicted by lifetime alcohol involvement, hassles, social support, spouse's internalizing symptoms, spouse's lifetime alcohol involvement, and spouse's current drinking (Figure 9).' Externalizing symptoms, for women, would be predicted by lifetime alcohol involvement, hassles, major life events, spouse's internalizing and externalizing symptoms, and spouse's lifetime alcohol involvement (Figure 10). In sum, internalizing symptoms would be predicted more by daily hassles, with current drinking minimizing the effects of hassles, and spouse's level of symptoms. 5 The correlations that are presented in Figures 7, 8, 9, and 10 were drawn from the earlier results. Correlations for Major Life Events and Social Support were not calculated in the earlier analyses; therefore, only the hypothesized direction of affects are presented. (+) Spouse's Lifetime Current . Alcohol Drinking \\\ ’//Involvement -e18 t0 -e21 e27 to 046 Externalizing Symptoms .30 to .33 .12 to .48 Spouse's I Spouse's Lifetime .09 to .75 Internalizing Alcohol and Involvement Externalizing ) Symptoms Note: Correlations listed in the model are the Pearson product moment correlations obtained from the earlier analyses. Correlations were included in the model if they were significant at the p<.0 level, or better. , Internalizing symptoms include physical illness, self-reported depression and, interview- rated depression. Externalizing symptoms include childhood and adulthood antisocial behavior and, total antisocial behavior. Figgre Ag. Model Predicting Externalizing Symptoms in Women 167 168 Externalizing symptoms would be predicted more by lifetime alcohol involvement and by major life events and spouse's symptoms. There also are more spouse influences when building this model. Last, stage of drinking is a necessary factor in predicing each type of symptom, for both men and women. As stage of drinking progresses, the intensity of symptoms should increase. Methodological Issues and Problems: As mentioned at the beginning of this chapter, the sample studied here is representative of a population of heavy drinkers with high levels of symptomatology. Thus the findings are not generalizable to moderate drinkers with low levels of trouble. similarly, the drinking patterns of couples in this sample were not systematically varied to cover the whole range of drinking combinations. The data presented here are representative of a particular kind of family where the husband/father is a heavy drinker and in the majority of cases has already achieved a lifetime diagnosis of alcoholism. (A substantial subset of these men are currently in remission.) Wives in these families have a greater variation in level of drinking involvement, although they are representataive of the range of women who marry alcoholic men. Last, the sample is drawn from a midwestern area and consists of Caucasion, 169 primarily blue collar, individuals. In these respects, the results may not be generalizable to other social statuses and races, or conceivably, to other geographic areas. A small sample always presents the problem of sample bias due to selective recruitment, selective participation, a selective population, or some other idiosyncratic characteristics. Much work in the present study was done to minimize such flaws. The potential bias in recruitment was minimized by gathering subjects from two different sources: the entire population of district courts in the greater Lansing area, and a variety neighborhoods in the greater Lansing area, with all of the possible subjects from both of these groups being approached regarding participation in this project. Selective participation can be ruled out in that 90% of the court referred families agreed to participate and 100% of the selected neighborhood families agreed to participate. (When collecting demographic information in the neighborhoods via door-to-door canvassing in order to select neighborhood families that fit the study criteria, 90% of the families approached offered the requested information.) Another concern in this work is the potential bias originating from the fact that the data are heavily based on (adult) self-reports. In order to minimize this 170 problem, a strong effort was made to insure high quality data. All of the study families were seen over several months and personal relationships were developed with each family member. As a result, the rapport with these families was exceptionally good. In addition, data were collected in family homes with sessions scheduled at the family's convenience; when data collection was completed, families were paid substantially for their participation. Therefore, if a bias exists due to self-report, it is likely to come more from unconscious denial or from a lack of subjective experience of the phenomenon being measured. Measurement Issues: In trying to understand the pattern of results in this data set, the actual measures used must also be considered. In measuring depression, we used two types of instruments, a self-report measure of current depression (The Short Beck Depression Inventory) and an interviewer- rated measure of current depression obtained during a long diagnostic interview (The Hamilton Rating Scales of Depression). The review of the literature showed the inter-correlation between the HRS and the regular BDI, when it was interview-rated for a clinically depressed outpatient population, was .75, indicating substantial stability across measures. Inter-correlations between the BDI and the SBDI on this same population were .96 indicating that the SBDI is a similar measure of 171 depression. However, when the interview-rated HRS and the self-report SBDI are compared here, the results are quite different . For example, the correlation between the two measures of current depression for women is .25 (p<.10) while the correlation for men is .56 (p<.001). In addition, the data also show that current depression on the SBDI and on the HRS have different relationships to hassles, social support, and drinking. All of the above findings suggest that the Beck and Hamiliton scales are measuring conceptually different components, and/or the parallels obtained in earlier work to not apply to the current population. Other conceptual issues may be raised concerning the measurement of social support. The current state of the construct validity of instruments in this field is quite poor. Concurrent validities between the Norbeck Social Support Questionnaire (1981) and a social support questionnaire by Cohen and Lazarus (19 ) ranged from - .03 to .51. This type of data is unavailable for many other social support instruments, but one would be tempted to conclude that it would be equally poor since the content of the instruments varies widely. In sum, it is still not clear what constitutes a valid measure of social support...whether it is the number of people an individual knows, the quality or type of their relationships, the frequency of contact with others, etc. Therefore it is 172 possible that the lack of significant social support results in the current study is at least in part due to the current state of conceptual and empirical weakness of the existing network of reasearch. Future Directions: The results of this dissertation clearly indicate that future research in the area of drinking, stress, and adjustment needs to differentiate between current drinking and lifetime alcohol involvement, between the stress of major life events and the stress of daily hassles, and between internalizing and externalzing symptoms. In addition, the relationships of the above factors needs to continue to be examined separately for the two sexes. In addition, a marked increase in sample size is also called for in order to reduce type II error. By increasing the range of drinking behavior sampled, and by increasing the number of subjects in each section of the distribution, additional manipulations of the data will also be possible. This would allow division of the sample by age of subject so that the effects of this developemental variable can be measured and, exploration of the ecology of variations in couples' drinking pattern. As the Michigan State University Longitudinal Study progresses, longitudinal data are to be collected on an increasingly larger sample, enabling the project to 173 reliably examine drinking, stress and symptom changes over time. In addition, a larger N will allow for statistical control of initial symptom level, as this may relate to changes over time. Analyses with the longitudinal data set should also help determine if stage of drinking (early-, middle-, and later stage alcoholism), has an impact on adjustment, and if the "type" of marriage has an impact on adjustment. This latter factor may also account for a remaining portion of the variance in adjustment (Hammen et al., 1986). In addition, the longitudinal research will allow a more definitive causal analysis of the data. Finally, the larger study is to be expanded to include measures of children's adjustment. These measures will be entered as dependent variables in the model outlined in the present study, with the goal of discovering what factors in a drinking couple's interaction place their child at high or lower risk for maladjustment. APPENDICES Appendix 1 Contact Schedule NAME: ADDRESS: TARGET CHILD: PHONE: FIRST CONTACT Project Informed Consent Medical Release Form Video/Audio Release Health History Demographic First Visit Rating Reading Level SECOND CONTACT DOTS - Child Beck Activity Inventory (ASS) DOTS - Self Drinking a Drug Symbol Digit Modality Moos Child Behavior Checklist Conners Parent Questionnaire DOTS - Spouse FACES II OSLC Family Crisis List Yale Stanford-Binet Miniature Situation Delay of Gratification California Child Q-Sort THIRD CONTACT Medical history - Child Physical Exam Neurodevelopmental Assmt. Sturmer Vision Eval. FOURTH CONTACT Coddington Life Stressors Life Events Questionnaire Smell Task Beverages Task Alcohol Concepts Task 1'74 W FIFTH CONTACT Eybergs Data Recording Hassles 6 Uplift Child Behavior Rating Videotape Work Satisfaction (Only to those who work) Social Support Parenting Q-Sort SEVENTH CONTACT Preschool Home Inventory HOME (Scales I - IV) EIGHTH CONTACT 015 Hamilton Psychiatric Conflict Tactic Axes V of DSM 111 Progress Evaluation (PES) Benjamin (Answer Sheets) Coping Questionnaire Height 6 Weight Buss-Durkee (Opinion Inv.) Pearlin Coping & Mastery WAIS 5 Digit Symbol Moos (Sibs 10 6 up) Fears California Adult Q-Sort NINTH CONTACT Parent Daily Report (6) rsnrn CONTACT Recontact Form Genogram 6 Payment Appendix 2 9 Social Support interview (Next) I'd like to ask you some questions about the different people that you know well. First. l’d like to ask you who the people are who you feel provide help and personal support for you or who are important to you. These people might include the following: -your spouse or partner -family members or relatives -friends -work or school associates -neighbors —health care providers -counselor or therapist -minister/priest/rabbi -others Can you give me the first names or initials. so that i can list them: then. l'm going to ask you a number of other questions about them. OK. 90 ahead. {GET LIST) Is there anyone else who is important to you or who provides you with personal support? {ADD ANY ADDITIONAL NAMES} Now. as I read the list back to you. could you tell me each person's relationship to you? (RECORD RELATIONSHIP} OK. now l'm going to ask you some questions about the people you have mentioned. l'll read the first name on your list and then ask you some questions about him or her. i want you to think of each person. and answer each question according to the following scale: (GIVE RESPONDENT A CARD OF THE SCALE) Question #3 to #12 l s not at all 2 . a little 3 - moderately 4 a quite a bit 5 s a great deal ' Hell.C.H. and Zucker.R.A.. H.S.U. Family Study. Michigan State University. l984. 175 Appendix 2 (cont'd.) You can answer the first two questions by Just giving me a number. {INTERVIEHER SHOULD THEN TRANSFORH RESPONDENT'S ANSWER INTO CORRESPONDING SCALE SCORE) Question #1 Question #2 l - once a year or less I . less than 6 months 2 - a few times a year 2 n 6 to IZ months 3 - monthly 3 . l to 2 years 4 . weekly 4 c 2 to 5 years 5 - daily 5 - more than 5 years {INTERVIEHER SHOULD SUBSTITUTE A NAME FOR "THIS PERSON”) 1. How often do you usually have contact with this person (phone calls. visits. letters)? 2. How long have you known this person? 3. How alike are you and this person? 4. How much does this person believe in you and accept your thoughts and ideas? 5. How much can you trust this person and confide in her/him? 6. How much would this person be likely to help if you needed to borrow a few dollars or get a ride somewhere? - 7. How much would this person help if you were upset. depressed. blue? 8. If you were sick in bed for more than a\week. how much would this person be likely to help? 9. How much can you depend on this person in a crisis? 10. How much does this person make you feel liked or loved? How much is this person someone you'd like to be like? {AFTER ALL OF THE ABOVE QUESTIONS HAVE BEEN ASKED FOR EACH INDIVIDUAL LISTED. GO ON TO THE FOLLOWING QUESTIONS) 176 Appendix 2 (cont'd.) IS. How close does this person live to you? (Or how close is the office if referring to a professional person). I6. Hhen was the last time you had contact with this person? Question #15 Question #I6 I u ( IS miles I s < 2 days 2 s < 50 miles 2 s < 1 week 3 s in state 3 s < I month 4 - midwest 4 u < 6 months 5 s farther 5 I > 6 months 12. Now I want to find out how well the people you have mentioned know each other. Remember to use the scale card for your answers. OK. how well does (first name on list) know (each remaining names on the list)? (THE INTERVIEHER SHOULD PROCEED IN THIS FASHION UNTIL THE ATTACHED HATRIX IS COMPLETE. NOTE THAT ONLY I/Z OF THE HATRIX NEED BE COMPLETE T0 COVER ALL POSSIBLE COHBINATIONS.) NOW I HAVE A COUPLE HORE GENERAL QUESTIONS FOR YOU. 13. How many clubs and organizations do you belong to? 14. How often do you attend religious services? 177 Appendix 2 (cont'd.) Family Number: Given By: Date: Support interview Form Social PRE PTI PTZ PT3 Post Ans. Chk: Test: 5 6 7 B 9 IO II I2 I3 Id I5 l6 4 QUESTION I 2 3 RELATION NAHE J I J J L L J_ J J J J J J J J L L J_ J J J J J J J L l J J L J 2 3 l J J J J J J l J L L J J 4 l J L L J J J J J J J J J l J J J J J J J J J l J J J J J J J J J l J J J J J L J J J J J J l J J J J J J I J J J l J J J J J J L J J J l J l J l J J J J J J J_-J--l--1--L--J--J--l l l J ID I J J J J J J L J J l I2 J J J J J J J I L J J J I J J J J J J J J J L L J J I J 13 J J J J J J J J J I I L L J J I I4 lJJJJJJjJJJiLJJl IS 16 l J J J J J L L J I I I J J J I7 J J J J J J I I L L J I I I8 I J J J J J L L L J L L I9 I J J J I L L L J I _L 178 20 Appendix 2 (cont'd.) Ow o~ m~ hm o~ m— "xzu .nc< anon man ~hL uhn mun "amok "025 ">m :U>_U “popes: >~asom v~ Mu - ~— O— o m p w w v n ~ a shoe seasusucn uuoqaom He‘uom 179 Appendix 3 The Relationship Between Daily Uplifts and Measures 9; Self Drinking and Spouse Drinking (Pearson R's) Husbands (N-SO): Overall Uplifts Hassles-Uplifts Hassles/Uplifts Wives (N=30): Overall Uplifts Hassles-Uplifts Hassles/Uplifts Self Drinkingi a m -e13 .02 -.O4 LAPS .18 -.08 -.02 Self Drinkingi 9E! -.09 .08 .09 LAPS -s09 .17 .22 b Spouse Drinkingi QFV LAPS -0244. -s03 .14 .02 .01 .07 Spouse Drinkingi QFV LAPS .05 .10 -s17 001 -029+ .07 Note: a=The Quantity-Frequency-Variabi1ity Index which is a measure of current drinking level. b-The Lifetime Alcohol Problem score which is a measure of lifetime alcohol involvement. 184 The Relationship Between Different Symptoms Among Women Appendix 4A (N=30), (Pearson R's) 1) 2) 3) 4) 5) 5) Note: Physical Illness Depression (Beck) Depression (Hamilton) Antisocial Behavior- Total Antisocial Behavior- Childhood Antisocial Behavior- Adulthood 2 3 i .68*** .31* .46** .24 .38* .13 **p .01, ***p .001 185 IUI .29 .28 .15 .89*** Im .64*** .41* .O7 .89*** .60*** Appendix 43 The Relationship Between Different Symptoms Among Men (N=30) (Pearson R's) 2 .3. 5. E IO\ 1) Physical .51** .46** .01 -.06 .06 Illness 2) Depression .56*** .25 .21 .29 (Beck) 3) Depression .44** .30* .49** (Hamilton) 4) Antisocial .92*** .93*** Behavior- Total 5) Antisocial .73*** Behavior- Childhood 6) Antisocial Behavior- Adulthood Note: *2 .05, **p .01, ***p .001 186 BIBLIOGRAPHY References Ablon, J. (1976). Family structure and behavior in alcoholism: A review of the literature. In B. Kissin & H. Begleiter (Eds.), The biology of alcoholism: Vol. 4, social aspects of alcoholism (205-242). New York: Plenum Press. Achenbach, T. & Edelbrock, C. (1981). Behavioral problems and competencies reported by parents of normal and disturbed children aged four through sixteen. Monographs of the Society for Research in Child Development— 46, (Serial No. 188). Aldwin, C., Folkman, S., Schaefer, c., Coyne, J., & Lazarus, R. (1980, August). Ways of co in A process measure. Paper presented at the Meeting of the American Psychological Association, Montreal, Canada. Alterman, A. J. a Tartar, R. E. (1983). The transmission of psychological vulnerability: implications for alcoholism etiology. The Journal of Nervous and Mental Disease, 171(3), 147-15T. Aneshensel, C. & Stone, J. (1982). Stress and depression: A test of the buffering model of social support. Archives 2; General Psychiatry, 32, 1392-1396. Bacon, 5. D. (1973). The process of addiction to alcohol: social aspects. Quarterly Journal 9; Studies pg Alcohol, 31, 1-27. Barnes, J. (1972). Social networks. Reading, Mass.: Addison-Wesley, 1-29. Baxter-Hagaman, J. A. (1986). Young alcoholic families and the transmission of risk: environmental and family interaction differences from the MSU longitudinal stud . Unpublished dissertation. Michigan State University, East Lansing. Beck, A., Rial, W. & Rickels, K. (1974). Short form of depression inventory: cross validation. Psychological Reports, 35, 1184-1186. Beck, A. T., Ward, C. H., Mendelson, H., Mock. J. & Erbaugh, J. (1961). An inventory for measuring depression. Archives 2; General Psychiatry, 1, 5-63. 187 188 Beckman, L. (1980). Perceived antecedents and effects of alcohol consumption in women. Journal 23 Studies pg Alcohol, 33, 518-530. Billings, A. 8 Moos, R. (1980). The role of coping responses and social resources in attenuating the stress of life events. Journal 23 Behavioral Medicine, 3, 139-157. Blane, H. T. 8 Hewitt, L. E. (1977). Alcohol and yputh: Ag anal sis 23 the literature, 1960-1975. National Institute on Alcohol Abuse and Alcoholism; Alcohol, Drug Abuse and Mental Health Administration. Rockville, MD: Department of Health, Education and Welfare. Block, J. (1961). The Qrsort method in personality assessment and psychiatric research. Springfield, IL: Charles C. Thomas Publisher. Block, J. H. 8 Block, J. (1980). The role of ego-control and ego-resiliancy in the organization of behavior. In Collins, W. A. (Ed.), Development of Cognition, Affect, and Social Relations. The Minnesota Symposia 93 Child Psychology, New Jersey: Erlbaum Associates, 33, 39-102. Block, J. 8 Block, J. H. (1981). The California child 9: set: orientation and introduction. Unpublished Manuscript. University of California, Berkely. Bloom, B. L. (1985). Stressful life event theory and gesearch: implications for primary prevention. Rockville, Md.: U.S. Department of Health and Human Services Publication No. (ADM)85-1385. Bowlby, J. (1973). Attachment and loss. London: Hogarth. Brandt, P. 8 Weinert, C. (1981). The PRQ - A social support measure. Nursing Research, 33, 277-280. Britt, D. 8 Campbell, E. (1977). A longitudinal analysis of alcohol use, environmental conduciveness and 1640-1647. Broadhead, W. E., Kaplan, B., James, S., Wagner, H., Schoenbach, V., Grimson, R., Heyden, S., Tibblin, 6., 8 Gehlbach, S. (1983). The epidemiologic evidence for a relationship between social support and health. American Journal 23 Epidemiology, 331, 521-537. 189 Brown, G. W. 8 Harris, T. (1978). Social origins g3 depression: A study 23 psychiatric disorder 33 women. London: Tavistock Publications, 63-140. Brown, S. (1985). Treating the alcoholic; A developmental model 23 recovery. New York: John Wiley and Sons. Burchfield, 8., Hamilton, E. 8 Banks, W. (1982). Affiliative needs, interpersonal stress and symptomatology. Journal 93 Human Stress, 3, 5-9. Burks, N. 8 Martin, B. (1985). Everyday problems and life change events: ongoing versus acute sources of stress. Buss, A., Fisher, H. 8 Simmons, A. (1962). Aggression and hostility in psychiatric patients. Journal 23 Consulting Psychology, 33, 84-89. Buss, A. 8 Durkee, A. (1957). An inventory for assessing different kinds of hostiity. Journal 23 Consulting Psychology, 33, 343-349. Cahalan D. 8 Cisin, I. (1976). Drinking behavior and drinking problems in the United States. In Kissin B. and Begleiter, H. (Eds.) The Biology 23 Alcoholism: Social Aspects g3 Alcoholism: Vol. 3 (77-115). New York: Plenum Press, Cahalan, D., Cisin, I., 8 Crossley, H. (1969). American drinking practices: A national study 93 drinking behavior and attitudes. New Brunswick, N.J.: Publications Division, Rutgers Center of Alcohol Studies. Carpenter, J., 8 Lester, D. (1980). The etiology g3 alcoholism: 2 longitudinal study. Unpublished Manuscript. Rutgers Center for Alcohol Studies, New Brunswick, N. J. Carey, W. 8 McDevitt,s. (1978). Stability and change in individual temperament diagnosis from infancy to early childhood. Journal 93 the American Academy 23 Child Psychiatry, 33, 331-337. Carrol, B., Fielding, J. 8 Blashki, T. (1973). Depression rating scales: A critical review. Archives 23 General Psychatry, 33, 361-366. Casey, R., Masuda, M. 8 Holmes, T. (1967). Quantitative study of recall of life events. Journal 93 Psychosomatic Research, 33, 239-247. 190 Chiles, J., Stauss, F. 8 Benjamin, L.S. (1980). Marital conflict and sexual dysfunction in alcoholic and non- alcoholic couples. British Journal 23 Psychiatry, 137, 266-273. Chiriboga, D. 8 Dean, H. (1978). Dimensions of stress: perspectives from a longtudinal study. Journal 93 Psychosomatic Research, 33, 47-55. Cobb, S. (1976). Social support as a moderator of life stress. Psychosomatic Medicine, 33, 300-314. Cobb, S. (1974). A Model for life events and their consequences. In Dohrenwend, B. s. and Dohrenwend, B. P. (Eds.) Stressful Life Events: Their Nature and Effects (151-156). New York: John Wiley and Sons. Cochran M. 8 Brassard, J. (1979). Child development and personal social networks. Child Development, 33, 601- 616. Coddington, R. D. (1972a). The significance of life events as etiologic factors in the diseases of children I - A survey of professional workers. Journal 93 Psychosomatic Research, 33, 7-18. Coddington, R. D. (1972b). The significance of life events as etiologic factors in the diseases of children II - A study of a normal population. Journal 93 Psychosomatic Research, 33, 206-213. Cohen, S, 8 Wills, T. (1985). Stress, social support, and the buffering hypothesis. Psychological Bulletin, 33, 310-357. Collins, R. L., Parks, G. A., 8 Marlatt, G. A. (1985). Social determinants of alcohol consumption: the effects of social interaction and model status on the self-administration of alcohol. Journal 23 Consulting and Clinical Psychology, 33, 189-200. Cooley, E. 8 Keesey, J. (1981). Moderator variables in life stress and illness relationship. Journal 23 Human Stress, 1, 35-40. Coyne, J.C., 8 DeLongis, A. (1986). Going beyond social support: the role of social relationships in adaption. Journal 23 Consulting and Clinical Psychology, 33, 454- 460. 191 Cronkite, R. 8 Moos, R. (1984). The role of predisposing and moderating factors in the stress-illness relationship. Journal 23 Health and Social Behavior, 33, 372-393. Cotton, N. S. (1979). The familial incidence of alcoholism: A review. Journal 23 Studies 93 Alcohol, 33, 89-116. Depue, R. A. 8 Monroe, S. M. (1986). Conceptualization and measurement of human disorder in life stress research: the problem of chronic disturbance. Psychological Bulletin, 99(1), 36-51. Diagnostic and statistical manual of mental disorders: DSM-III (1980). The American Psychological Association. Dohrenwend, B. S. 8Dohrenwend, B. S. (1981). Stressful life events and their contexts. New Brunswick, N. J.: Rutgers University Press. DeLongis, A., Coyne, J., Dakof, G., Folkman, S., 8 Lazarus, R. (1982). Relationship of daily hassles, uplifts, and major life events to health status. Health Psychology, 3, 119-136. Dohrenwend, B. S., Dohrenwend, B. P., Dodson, M., 8 Shrout, P. (1984). Symptoms, hassles, social supports, and life events: problem of confounded measures. Journal 93 Abnormal Psychology, 33, 222-230. Dunn, J., Kendrick, C. 8 MacNames, R. (1981). The reaction of first-born children to the birth of a sibling: mother's reports. Journal 23 Child Psychology and Psychiatry, 33, 1-18. El-Guebaly, M. 8 Offord, D. (1977). The offspring of alcoholics: A critical review. The American Journal 93 Psychiatry, 134, 357-365. Feighner, J., Robins, E., Guze, 8., Woodruff, R., Winokur, G., 8 Munoz, R. (1972). Diagnostic criterion for use in psychiatric research. Archives 93 General Psychiatry, 33, 57-63. Folkman, S. 8 Lazarus, R. (1986). Stress processes and depressive symptomatology. Journal 93 Abnormal Psychology, 33, 107-113. Folkman, S. 8 Lazarus, R. (1980). An analysis of coping in a middle-aged community sample. Journal 93 Health and Social Behavior, 33, 219-239. 192 Fontana, A., Hughes, L., Marcus, J. 8 Dowds, B. (1979). Subjective evaluation of life events. Journal 93 Consulting and Clinical Psychology, 33, 906-911. Forbes, R. (1979). Life Stress. Garden City, New York: Doubleday, 104-113. Garmezy, N. (1981). Children under stress: perspectives on antecedants and correlates of vulnerability and resistance to psychopathology. In Rabin, A., Aranoff, J., Barkley, A., and Zucker, R. (Eds.) Further Explorations 3A Personality (196-269). New York Wiley. Gersten, J., Langner, T., Eisenberg, J. 8 Orzek, L. (1974). Child behavior and life events: undesirable change or change per so? In Dohrenwend, B. S. and Dohrenwend,, B. B. (Eds.) Stressful Life Events: Their Nature and Effects. New York: John Wiley and Sons. Goldman, H. H. 8 Goldston, S. E. (Eds.) (1985). Ageventing stress-related psychiatric disorders. Rockville, Md.: U.S. Department of Health and Human Services Publication No. (ADM)85-1366. Guilford, J. (1956). Fundamental statistics 33 psychology and education. New York: McGraw-Hill Book Company Inc. Hamilton, M. (1960). A rating scale for depression. Journal 93 Neurology, Neurosurgggy and Psychiatry. 33. 56-62. Hamilton, M. (1967). Development of a rating scale for primary depressive illness. British Journal 23 Social and Clinical Psycholggy, 3, 278-296. Hamilton, M. (1969). Standardized assessment and recording of depressive symptoms. Psychiatry, Neurologia, and Neurchirurgia, 33, 201-205. Hammen, C., Mayol, A., deMayo, R., 8 Marks, T. (1986). Initial symptom levels and the life-event-depression relationship. Journal 93 Abnormal Psychology, 33, 114- 122. Hatzenbuehler, L., Parpal. M. 8 Matthews, L. (1983). Classifying students as depressed or nondepressed using the beck depression inventory: an empirical analysis. Journal 23 Consulting and Clinical Psycholo , 33, 360- 366. 193 Heller, K. (1986). Introduction to the special series: diaggregating the process of social support. Journal 93 Consulting and Clinical Psychology, 33, 415. Heller, H., Swindle, R., 8 Dusenbury, L. (1986). Component social support processes: comments and integration. Journal 23 Consulting and Clinical Psychology, 33, 466-470. Herjanic, B., Herjanic, M., Penick, E., Tomelleri, C. 8 Armbruster, R. (1977). Children of alcoholics. In Seixas, P. (Ed.) Currents 3A Alcoholism, Vol. 33. New York: Grune and Stratton. Hesselbrock, M., Hesselbrock, V., Tennan, H., Meyer, R. 8 Workman, K. (1983). Methodological considerations in the assessment of depression in alcoholics. Journal 93 Consulting 229 Clinical Psychology, 33, 399-405. Hetherington, E. M. (1972). The effects of father absence on personality-development in adolescent daughters. Developmental Psychology, 1, 313-326. Holahan, C. H. 8 Moos, R. H. (1987). Risk, resistance, and psychological distress: A longitudinal analysis with adults and children. Journal 23 Abnormal Psychology, 33, 3-13. Holahan, C. 8 Moos, R. (1981). Social support and psychological distress: A longitudinal analysis. Journal 23 Abnormal Psychology, 33, 365-370. Holmes, T. 8 Rahe, R. (1967). The Social readjustment rating scale. Journal Psychosomatic Research, 33, 213- 218. Hore, B. D. (1971a). Factors in alcoholic relapse. British Journal 33 Addiction, 33, 89-96. Hore, B. D. (1971b). Life events in alcoholic relapse. British Journal 23 Addiction, 33, 83-88. Hubert, N., Wachs, T., Peters-Martin, I. 8 Gandour M. J. (1982). The study of early temperament: measurement and conceptual issues. Child Development, 33, 571-600. Hunter, J. (1980). Partial correlations. Unpublished manuscript. Michigan State University, East Lansing. Hurme, H. (1981). Life changes during childhood. University of Jyvaskyla, Jyvaskyla. 194 Jackson, J. (1962). Alcoholism and the family. In D. J. Pittman and C. R. Snyder (Eds.), Society, Culture, and Drinking Patterns. New York: John Wiley and Sons. Jackson, J. (1954). The problem of alcoholic tuberculous patients. In P. F. Sparer, Personality, Stress and Tuberculosis. New York: International Universities Press. Jacob, T. (1987). Alcoholism: a family interaction perspective. In C. Rivers (Ed.) Nebraska S osium 33 Motivation, 33, Lincoln Nebraska: University of Nebraska Press. Jacob, T., Favorini, A., Meisel, S. 8 Anderson, C. (1978). The alcoholic's spouse, children, and family interactions. Journal 93 Studies 93 Alcohol, 3, 1231- 1251. Jacob, T. 8 Seilhamer, R. A. (in press). Alcoholism and Family Interaction. In T. Jacob (Ed.) Famil interaction and psychgpathology: theor es, methods and findings, New York: Plenum Publishing Corporation. Jellinek, E. M. (1952). Phases of alcohol addiciton. Johnston, L. D., Bachman, J.G., 8 0'Malley, P.M. (1979). Drugs and the class 23 333. National Institute on Drug Abuse: Division of Research. U. S. Department of Health, Education, and Welfare. Kanner, A., Coyne, J., Schaefer, C., 8 Lazarus, R. (1981). Comparison of two modes of stress measurement: daily hassles and uplifts versus major life events. Journal 93 Behavioral Medicine, A, 1-39. Kaufman, E. 8 Pattison, E..M. (1982). The family and alcoholism. In, Kaufman and Pattison (Eds.), Encyclopedic Handbood g3 Alcoholism. New York: Gardner Press. Kessler, L.G. 8 Eaton, W.W. (1985). Epidemiologic field methods 33 psychiatry: The NIMH catchment area program. Orlando, Fla: Academic Press, Inc. Kessler, R. C., Price, R. H. 8 Wortman, C. B. (1985). Social factors in psychopathology: stress, social support, and coping processes. Annual Reviews 93 Psychology, 33, 531-572. 195 Lazarus, R. (1984). Puzzles in the study of daily hassles. Journal 23 Behavioral Medicine, 3, 375-389. Lazarus, R. 8 DeLonges, A. (1983). Psychological stress and coping in aging. American Psychologist, 33, 245- 254. Masuda, M. 8 Holmes, T. (1967). The social readjustment rating scale: A cross-cultural study of japanese and americans. Journal 23 Psychosomatic Research, 33, 219- 225. McCrady, B.S. (1982). Marital dysfunction: alcohol and marriage. In E.M. Pattison and E. Kaufman (Eds.), Encyclopedic Handbook 23 Alcoholism. New York: Gardner Press. McCubbin, H. (1979). Integrating coping behavior in family stress theory. Journal 33 Marriage and the Family, 33, 237-244. McDevitt, S. 8 Carey, W. (1978). Measurement of temperment in 3- to 7-year-old children. Journal of Child Psychology and Psychiatry and Allied Disciplifies, 32, 245-253. Mehrabian, A. 8 Ross, M. (1979). Illnesses, accidents, and alcohol use as functions of the arousing quality and pleasantness of life changes. Psychological Reports, 33, 31-43. Menaghan, E. (1982). Measuring coping effectiveness: A panel analysis of marital problems and coping efforts. Journal 33 Health 333 Social Behavior, 33, 220-234. Mitchell, R. 8 Trickett, E. (1980). Social-network research and psychosocial adaptation: implications for community mental-health practice. In, Insel, P. M. (Ed.), Environmental Variables and the prevention 23 Mental Illness. Lexington, Mass.: Lexington Books - D. C. Heath. Mitchell, R. 8 Moos, R. (1984). Deficiencies in social support among depressed patients: antecedents or consequences of stress. Journal 23 Health and Social Behavior, 33, 438-452. Monroe, S. M., Bromet, E. J. Connell, M. M. 8 Steiner, S. C. (1986). Social supports, life events, and depressive symptoms: A 1-year prospective Study. Journal 23 Consulting and Clinical Psychology, 33, 424- 431. 196 Monroe, 8., Imhoff, D., Wise, B., 8 Harris, J. (1983). Prediction of psychological symptoms under high-risk psychosocial circumstances: life events, social support, and symptom specificity. Journal 23 Abnormal Psychology, 23, 338-350. Monroe, S. 8 Steiner, S. (1986). Social support and psychopathology: interrelations with preexisting disorder, stress, and personality. Journal 23 Abnormal Psychology, 33, 29-39. Moos, R.H., Bromet, E., Tsu, V. 8 Moos, B. (1979). Family characteristics and the outcome of treatment for alcoholism. Journal 33 Studies pp Alcohol, 33, 78-88. 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. Morgan, P. (1981). Alcohol and family violence: A Review of the Literature. In J. DeLuca (Ed.), Alcohol and Health Monographs, Vol. 3, Alcohol Consumption and Relited Problems. NIAAA, U.S. Government Printing Off ce. Morrisey, E. R. 8 Schuckitt, M. A. (1978). Stressful life events and alcohol problems among women seen at a detoxification center. Journal 33 Studies pp Alcohol, 39(9), 1559-1576. Mulford, H. (1977). Stages in the alcoholic process: toward a cumulative, nonseguential index. Quarterly Murphy, I. B. 8 Moriarty, A. E. (1976). vulnerability, coping and growth from infancy 32 adolescence. New Haven: Yale University Press. Neff, J. A. (1984). The stress-buffering role of alcohol consumption: the importance of symptom dimension. Journal 93 Human Stress, (Spring), 35-42. Nie, N. H., Hull, C. H., Jenkins, J. G., Steinbrenner, F., Bent, 8 D. H. (1975). Statistical packagg for the social sciences, Second Edition. New York: McGraw Hill Book Co. 197 Noll, R. B. 8 Zucker, R. A. (1986). The antisocial behavior inventory: psychometric characteristics and discrimant validity. Unpublished paper, E. Lansing: Department of Psychology, Michigan State University. Noll, R. 8 Zucker, R. (1983). Developmental findingg from 33 alcoholic vulnerability study. Paper presented at the Annual Meeting of the American Psychological Association, Anaheim, California. Norbeck, J. (1982). Social su ort estionnaire. Unpublished Manuscript. Un vers ty of Cal fornia, San Francisco. Norbeck, J. (1981). Revised scorin instructions 333 333 Norbeck social support guest onnaire (N8393. Unpubiished Manuscript. Department of Mental Health and Community Nursing, University of California, San Fransisco. Norbeck, J., Lindsey, A., 8 Carrieri, V. (1981). The development of an instrument to measure social support. Nursipg Research, 33, 264-269. Nylander, I. 8 Rydelius, P. A. (1982). A comparison between children of alcoholic fathers from excellent versus poor social conditions. Acta Paediatria Scandinavia, 13, 809-813. O'Donnell, J., Voss, H., Clayton, R. Slatin, G. 8 Room, R. (1976). Young men and drugg 2.3 nationwide survey. Rockville, Maryland: National Institute on Drug Abuse. 0rford, J., Guthrie, S., Nicholls, P., Oppenheimer, E., Egert, S., 8 Hensman, C. (1975). Self-reported coping behavior of wives of alcoholics and its association with drinking outcome. Journal 33 Studies 33 Alcohol, 33, 1254-1257. __ __ —— Pachman, J. S. 8 Foy, D. W. (1978). A correlational investigation of anxiety, self-esteem and depression: new findings with behavioral measures of assertiveness. Journal 33 Behavioral Therapy and Egperimental Psychiatry, 3, 97-101. Paolino, T. 8 McCrady, B. (1977). The alcoholic marriage: alternative perspectives. New York: Grune and Stratton. Park, P. (1973). Developmental ordering of experiences in alcoholism. Quarterly Journal 33 Studies 33 Alcohol, :3, 473-488. _ " 198 Parker, G. B. 8 Brown, L. B. (1982). Coping behaviors that mediate between life events and depression. Archives 33 General Psychiatry, 33, 1386-1391. Parker, P. A., Parker, E. s., Wolz, M. W. 8 Harford, T. C. (1980). Sex roles and alcohol consumption: a research note. Journal 33 Health and Social Behavior, 33, 43-48. Pascoe, J. 8 Earp, J. (1984). The effect of mothers' social support and life changes on the stimulation of their children in the home. American Journal 33 Public Health, 13, 358-360. Paykel, E. s. (1974). Life stress and psychiatric disorder: applications of the clinical approach. In Dohrenwend, B. S. and Dohrenwend, B. P. (Eds.) Stressful Life Events: Their Nature and Effects. New York: John Wiley and Sons. Rabkin, J. 8 Streuning, E. (1976). Life events, stress, and illness. Science. 194, 1013-1020. Rahe, R. (1979). Life change events and mental illness: an overview. Journal 33 Human Stress, 3, 2-10. Ray, C., Lindop, J., 8 Gibson, S. (1982). The concept of coping. Psychological Medicine, 33, 385-395. Reider, E. (1987). Unpublished Dissertation, Department of Psychology, Michigan State University, East Lansing, Michigan. Rhoads, D. (1983). A longitudinal study of life stress and social support among drug abusers. The International Journal 33 the Addictions, 33, 195-222. Robins, L.N., Helzer, J.H., Croughan, J. 8 Ratcliff, R.S. (1980). The NIMH dia ostic interview schedule: it's history, characterist cs and validity. St. Louis, MO: Washington University School of Medicine. Royce, J.E. (1981). The spouse and the family of the alcoholic. In, Alcohol problems and alcoholism: A comprehensive survey. New York: The Free Press. b Rutter, M. (1981). Stress, coping and development: some issues and some questions. Journal 33 Child Psychology and Psychiatry, 33, 323-356. 199 Rutter, M. (1987). Psychosocial resilience and protective mechanisms. American Journal 33 Orthopsychiatry, 33, 316-331. ' Sands, J. 8 Parker, J. (1979-80). A cross-sectional study of the perceived stressfullness of several life events. International Journal 33 Aging and Human Development, 33, 335-341. Sarason, 1., Johnson, J. 8 Siegel, J. (1978). Assessing the impact of life changes: development of the life experiences survey. Journal 33 Consulting and Clinical Psychology, 33, 932-946. Schaefer, C., Banner, P., Cohen, J., Folkman, 8., Kanner, A., Lazarus, R., 8 Wrubel, J. (19 ). Rationale and instructions for the ways 33 coping checklist coping guestionnaife, and ways 33 cop ng. Unpublished Manuscript. Stress and Coping Project, University of California, Berkeley. ------------- (1985). Ways p3 coping: revised. Unpublished Manuscript. Stress and Coping Proiect. University of California, Berkeley. Schaffer, J. 8 Tyler, J. Degree of sobriety in male alcoholics and coping styles used by their wives. British Journal 33 Psychiatry, 135, 431-437. Schuckitt. M.A. (1978). Research guestionnaire. Unpublished Manuscript, Alcoholism Treatment Program, V.A. Medical Center, University of California, San Diego. Scida, J. 8 Vannicelli, M. (1979). Sex-role conflict and 28-44. Seigel, S. (1956). Nonparametric statistics for the behavioral sciences (68-74). New York: McGraw-Hill Book Company, Inc. Selzer, M.L. (1971). The Michigan alcoholism screening test: the quest for a new diagnositc instrument. American Journal 33 Psychiatry, 127, 1653-1658. Selzer, M., Vinokur, A. 8 VanRooyan, L. (1975). A self- administered alcoholism screening test (SMAST). Journal of Studies on Alcohol, 33, 117-126. 200 Sher, K. J. 8 Walitzer, K. S. (1986). Individual differences in the stress-dampening effect of alcohol: A dose-response study. Journal 33 Abnormal Psycholggy, 23, 159-167. Shulamith. L., Straussner, A., Weinstein, D. 8 Hernandy, R. (1979). Effects of alcoholism on the family system. Health and Social Work, 3, 111-127. Sidle, A., Moos, R., Adams, J., 8 Cady, P. (1969). Development of a coping scale: a preliminary study. Archives 33 General Psychiatry, 33, 226-232. Steele, C. M., Southwick, L., 8 Pagano, R. (1986). Drinking your troubles away: the role of activity in mediating alcohol's reduction of psychological stress. Journal 33 Abnormal Psychology, 33, 173-180. Steinglass, P. (1979). The alcoholic family in the interaction laboratory. The Journal 33 Nervous and Mental Disease, 167, 428-436. Steinglass, P. (1980). A life history model of the alcoholic family. Family Process, 33, 211-226. Steinglass, P. (1981). The alcoholic family at home: patterns of interaction in dry, wet, and transitional stages of alcoholism. Archives 33 General Psychiatry, 33, 578-584. Suls, J. 8 Mullen, B. (1981). Life events, perceived control and illness: the role of uncertainty. Journal 33 Human Stress, 3, 30-34. Tahka, V. (1966). The alcoholic ersonalit : A Clinical Study. Publication No. 13, Helsinki: The Finn sh Foundation for Alcohol Studies. Thoits, P. A. (1986). Social support as coping assistance. Journal 33 Consulting and Clinical Psychology, 33, 416-423. Vaillant, G. 8 Milofsky, E. (1982). The etiology of alcoholism: a prospective viewpoint. American Psychologist, 33, 494-503. Weil, C. M. (1984). The relationship between life stress and gy3ptomatology 33 alcoholic and nonalcoholic famil e . Unpublished Masters Thesis, Department of Psychology, Michigan State University, East Lansing, Michigan. 201 Weil, C. M. 8 Zucker, R. A. (1985). The Social Support Interview. Unpublished Manuscript, Department of Psychology, Michigan State University, East Lansing. Whitehead, L. (1979). Sex differences in children's responses to family stress: a re-evaluation. Journal ‘33 Child Psychology and Psychiatry, 33, 247-254. Wilcox, B. (1981). Social support, life stress, and psychological adjustment: a test of the buffering hypothesis. American Journal 33 Community Psychology, 3, 371-386. Williams, A. F. (1976). The alcoholic personality. In B. Kissin and H. Begleiter (Eds.) The Biol3gy 33 Alcoholism, Vol. 3. New York: Plenum Press. Williams, A. F. (1966). Social drinking, anxiety, and depression. Journal of Personality and Social Psychology, 3, 689-695. Wilson, C. 8 0rford, J. (1978). Children of alcoholics: report of a preliminary study and comments on the literature. Journal 33 Studies 33 Alcohol, 33, 121- 142. '——-———‘ Woititz, J.G. (1977/78). Alcoholism and the family: a survey of the literature. Journal 33 Alcohol and Drpg Education, 33, 18-23. Wolff, S. (1973). Children under stress. Great Britain: Penguin Books. Wolin, S.J., Bennett, L., Noonan, D. 8 Tertelbaum, M. (1980). Disrupted family rituals: a factor in the intergenerational transmission of alcoholism. Journal Zarski, J. (1984). Hassles and health: a replication. Health Psychology, 3, 243-251. Zucker, R. A. (1987). The four alcoholisms: a developmental account of the etiologic process. In P. C. Rivers (Ed.) Nebraska Symposium 33 Motivation: Alcohol and Addictive Behaviors, 33, Lincoln, NE.: University of Nebraska Press. Zucker, R. (1966). Community adolescent drinking study-- research questionnaires. Unpublished Manuscript. New Brunswick, N.J.: Rutgers Center for Alcohol Studies. 202 Zucker, R. (1976). Parental influences upon drinking patterns of their children. In Greenblatt, M. and Schuckit, M. (Eds.) Alcoholism Problems 33 Women and Children. New York: Grune and Stratton. Zucker, R. (1980). The Michigan State University vulnerability study: overv ew, design, and method. Unpublished Manuscript, Department of Psychology, Michigan State University, East Lansing. Zucker, R. 8 Barron, P. (1973). Parental behaviors associated with problem drinking and antisocial behavior among adolescent males. In Chafetz, M. (Ed.) Research 33 Alcoholism 33 Clinical Problems and §pecia1 Populations. Washington D.C.: U.S. Government Printing Office (DHEW Publications 74-675). Zucker, R., Baxter, J., Noll, R., Theado, D., 8 Weil, C., (1982). A3 alcoholic risk study: design and early health related finding_. Paper presented at the annual meeting of the American Psychological Association, Washington, D.C.. Zucker, R. A. 8 Gomberg, S. A. (1986). Etiology of alcoholism reconsidered: the case for a biopsychosocial process. American Psychologist, 33, 783-793. Zucker, R. A. 8 Noll, R. B. (1987). The interaction of child and environment in the early development of drug involvement: a far ranging review and a planned very early intervention. Drugs and Society, 3. Zucker, R. 8 Noll, R. (1980). The antigocial behavior checklist. Unpublished Manuscript, East Lansing, Michigan: Department of Psychology, Michigan State University. Zucker, R. 8 Noll, R. (1982). Precursors and developmental influences on drinking and alcoholism: etiology from a longitudinal perspective. In Deluca, J. (Ed.) Alcohol and Health Monographs Z 33 Alcohol Consumption and Related Problems. National Institute on Alcohol Abuse and Alcoholism, Washington D.C.: U.S. Government Printing Office. Zucker, R. 8 Noll, R. (1984). Psychometric properties 33 333 antisocial behavior scale. Unpublished manuscript, East Lansing, Michigan: Department of Psychology, Michigan State University. 203 Zucker, R., Noll, R., Draznin, T., Baxter, J., Weil, C., Theado, D., Greenberg, G., Charlot, C., 8 Reider, E. (1984). The ecolo of alcoholic families: conceptual framework for the M chig33 State University Longitudinal Study. Paper presented at the National Council on Alcoholism, National Alcoholism Forum, Detroit, Michigan. "‘iiiiiiiiiiiii“