4 ‘ . tic - .u4—1.|1~.\- LIBRARY MichiganSm _ Wait)? ‘ ll lllllllllflflllll Ll Ill ll l 111 Ill Illlljl l. 3 1 0064SN3 This is to certify that the thesis entitled COLLEGE STUDENT DRINKING AND THE INTERPERSONAL MEASURES OF DIFFERENTIATION AND EXTENSION presented by James Joseph Mullin has been accepted towards fulfillment of the requirements for ' - Ph.D. Psychology degree in @“fil /\ ,’ " ,- :1.‘ Date F453” *3 i \q"f>k) 0-7 639 LFA“, J l 105‘ ": \Q’ A . .. E R?! iM'VlfiGlCE’ $313193 44 ’7 /( I; if. , ‘1" D9 .20., 1393 ‘51:)” V'Su '15 \ - . v in; t“ (315.; ' - . 3:312 55 ‘ ”‘ COLLEGE STUDENT DRINKING AND THE INTERPERSONAL MEASURES OF DIFFERENTIATION AND EXTENSION By James Joseph Mullin A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Psychology 1980 James Joseph Mullin Interpersonal Checklist. Drinking was measured in four ways: (1) a drinking score, in ounces of absolute alcohol, (2) a drinking pattern index, (3) drinking category (light, moderate, heavy), and (4) drinking category including ab- stainers. Analyses of variance for the factors of differentia- tion (high, low), extension (high, low), and sex, disclosed a significant effect for extension (based on the peOple orientation scale). Those classified as high on this ex- tension measure had a significantly lower mean drinking score than those classified as low on this extension mea- sure. Men had a significantly greater mean drinking score than women. However, the mean differences in the drinking score for the differentiation factor levels, while not sig- nificant, were Opposite to that hypothesized for both dif— ferentiation measures. Those classified as low in differ- entiation showed lower average drinking scores than those classified as high in differentiation. Separate discriminant analyses, for males and females, of the drinking pattern and of the drinking category group- ings did not reveal significant distinctions based on the differentiation or the extension measures. Only the dis— criminant analysis for males, which included abstainers, significantly distinguished abstainers from drinkers on the extension measure based on the Lov scale, and significantly distinguished abstainers from heavy drinkers on the James Joseph Mullin differentiation measure based on the differentiated role scale. The male abstainers scored significantly higher on the Lov scale and significantly lower on the differentiated role scale than the drinkers. On the basis of these results, it was prOposed that differentiation-extension might be regarded as a bipolar factor representing a dimension of relatedness to others. This would imply that college students who like people, and who view themselves as like other people, drink less alcohol than those who see themselves as distant from and different from other people. The relevance of this differentiation- extension dimension to Alcoholics Anonymous therapy and implications of the findings for future studies of drinking pOpulations were discussed. Many have helped along the way. Thank you, one and all. ii TABLE OF CONTENTS INTRODUCTION . . . . Differentiation - Extension Drinking Motives Personal Motives and Drinking . Drinking Specific - Nondrinking Specific Theories . College Population Sex Differences Differentiation and Alcohol Use Extension and Alcohol Use The Problem Summary and Hypotheses Subjects . . . . . . . . Sample Selection . . . . . . . Instruments . Measuring Alcohol Consumption The Inventory of Drinking Behavior Drinking Score Quantity- Pattern Index Drinking Categories . Verdicality of Drinking Inventory . Measures of Differentiation and Extension. Attitudes Toward Leadership Preference Inventory Interpersonal Check List Design and Analysis Analysis of Variance Discriminant Analysis Statistical Treatment Statement of Hypotheses RESULTS Drinking Score, Primary Measures of Differ- entiation and Extension, and Sex Drinking Score, Secondary Measures of Differ- entiation and Extension, and Sex Pattern of Drinking . . . . Drinking Category: Light, Moderate, Heavy . . . . . iv Page Amt—a onooouolom 12 12 14 1.4.. 1Q 17 18 19 20 21 22 22 23 24 24 25 t) L. 27 29 Drinking Classification: Abstain, Light, Moderate, Heavy Additional Results Concerning Sex Differences . . . . Summary . . . . . . . DISCUSSION Differentiation and Drinking Extension and Drinking Differentiation- Extension and Drinking Sex Differences and Comparisons Drinking Pattern . Drinking Category . . . Drinking Classification . . Extension - Differentiation and Drinking Implications of the Present Research Reformulation of the Composite Theory Alcoholics-Anonymous Therapy and the Re- formulated Composite Theory . Recommendations and Conclusions APPENDICES REFERENCE NOTES REFERENCES LIST OF TABLES Table ’ Page 1 Number of Subjects by College Year and Sex . . . . . . . . . . . 13 2 Means and Standard Deviations of Raw Drinking Score by Primary Measures of Differentiation and Extension and Sex . . . . . . . 30 3 Analysis of Variance Summary of Fourth Root of Drinking Score for Primary Measures of Differentiation and Extension and Sex . . . 31 4 Means and Standard Deviations of Raw Drinking Score by Secondary Measures of Differentiation and Extension and Sex . .' . . . . . 33 5 Analysis of Variance Summary of Fourth Root of Drinking Score for Secondary Measures of Differentiation and Extension and Sex . . . 34 6 Means and Standard Deviations of Males on Primary and Secondary Measures of Differen- tiation and Extension for Drinking Pattern (Q-P Index) Categories . . . . . . 35 7 Discriminant Analysis of Primary and Secondary Measures of Differentiation and Extension for Drinking Pattern (Q-P Index) Categories (Male Subjects) . . . . . . . . . . 36 8 Means and Standard Deviations of Females on Primary and Secondary Measures of Differen- tiation and Extension for Drinking Pattern (Q-P Index) Categories . . . . . . 38 9 Discriminant Analysis of Primary and Secondary Measures of Differentiation and Extension for Drinking Pattern (Q-P Index) Categories (Female Subjects) . . . . . . . . 39 vi Table Page 10 Means and Standard Deviations of Males on Primary and Secondary Measures of Differ- entiation and Extension for Drinking Cate- gory (Light, Moderate, Heavy) . . . . 4O 11 Discriminant Analysis of Primary and Secondary Measures of Differentiation and Extension for Drinking Category: Light, Moderate, Heavy (Male Subjects) . . . . . . . . 41 12 Means and Standard Deviations of Females on Primary and Secondary Measures of Differ- entiation and Extension for Drinking Cate- gory (Light, Moderate, Heavy) . . . . 42 13 Discriminant Analysis of Primary and Secondary Measures of Differentiation and Extension for Drinking Category: Light, Moderate, Heavy (Female Subjects) . . . . . . . 43 14 Means and Standard Deviations of Males on Primary and Secondary Measures of Differ- entiation and Extension for Drinking Classifi- cation: Abstain, Light, Moderate, Heavy . . . . . . . . . . 45 15 Discriminant Analysis of Primary and Secondary Measures of Differentiation and Extension for Drinking Classification: Abstain, Light, Moderate, Heavy (Male Subjects) . . . . 46 16 Standardized Discriminant Function Co- efficients for Function One and Associated Variables of the Drinking Classification Analysis for Males . . . . . . . 47 17 Means and Standard Deviations of Females on Primary and Secondary Measures of Differen- tiation and Extension for Drinking Classifica- tion: Abstain, Light, Moderate, Heavy . . 50 18 Discriminant Analysis of Primary and Secondary Measures of Differentiation and Extension for Drinking Classification: Abstain, Light, Moderate, Heavy (Female Subjects) . . . 51 19 Summary of Hypotheses Tested . . . . 55 vii Figure LIST OF FIGURES Page Trends toward organizational maturation . . . . . . . . 2 Centroid coordinates for significant dis- criminant function for males on drinking classification . . . . . . . 49 viii Appendix A LIST OF APPENDICES STUDENT SURVEY ON ATTITUDES AND DRINKING DIRECTIONS . . . SAMPLE SIZES BY RESIDENCE HALL ETHNICITY AND SEX . . . . PROCEDURE USED FOR GENERATING THE DRINKING (ABSOLUTE-ALCOHOL) SCORE ALPHA INTERNAL CONSISTENCY RELIABILITIES . . ix Page 67 78 79 82 INTRODUCTION The drinking of alcohol is associated with a variety of social and psychological conditions and individual dif- ferences. Cross cultural observation and research indicate that alcohol consumption is usually regulated by tradition in institutionalized settings (Bacon, Barry 5 Child, 1965; Bacon, 1974; Horton, 1943). Many differences in drinking practices can be understood within a sociocultural rubric. Beyond these considerations, however, an understanding of individual variation in the use of alcohol calls for the in- clusion of personality concepts (Hogan, DeSoto G Solano, 1977; Jessor, Carman a Grossman, 1968; Kessel a Walton, 1965; McClelland, Davis, Kalin G Wanner, 1972). The present study of alcohol use by college students employed Smith's (1979) composite theory of personality. Central to this theory are the developmental trends of differentiation and extension. Figure 1 pictures these trends, and adds the dimension of organization of the self (Smith, 1979, p. 23). Differentiation - Extension In studying differentiation, Smith (1966, 1973) has demonstrated that the tendency to assume similarity or (high) MATURITY / / // EXTENSION / 1 (low) 1/ (high) / / / / / / ORGANIZATION (low) DIFFERENTIATION Figure 1. Trends toward organizational maturation. (Modified from Smith, 1979) dissimilarity to others is a relatively consistent property of a person's social perception. Wegner (1971) in his study of self—other differentiation concluded that assumed simi- larity is an extension of Witkin's (1962) theory of psycho— logical differentiation: "articulated individuals are able to differentiate themselves from others, and assume dis- similarity to most people" (Wegner, 1971, p. 21). What enters into the development to extend ourselves? Smith (1979) states that, "the increasing satisfaction of the need for love results in the increasing extension of the self" (p. 23). The composite theory assumes a developmental sequence to extension and differentiation. However, the natural development of a person may be retarded at any stage: "The composite theory assumes...that it is of critical im- portance in understanding a particular person to know the highest form of self he has attained” (Smith, 1979, p. 24). The present study explores the contribution of the composite theory to an understanding of drinking behavior in college students. Drinking Motives In looking at the variety of reasons for drinking, Jessor, Graves, Hanson and Jessor (1968) point out that, at least in American society, there have been three broad cat- egories emphasized. These categories, based on the works of Mulford and Miller (1959, 1960b) and Fallding (1964), have been termed: social-convivial, social-facilitating, and personal-effects reasons. Social-convivial drinking is done for the enjoyment of the ongoing social interaction associated with it. Social- facilitating drinking is done to overcome social apprehen- sions and anxieties. Drinking for personal-effects reasons involves the use of alcohol for c0ping with many types of personal inadequacies and fears. In a later paper, Fallding and Miles (1974) proposed four types of normal or social drinking: ornamental-community- symbolic, facilitation, assuagement, and retaliation. Ornamental-community-symbolic is drinking that is ritualistic and ceremonial and is seen by Fallding and Miles as appro- priate. Facilitation drinking, done for relieving social tensions, was the typical kind of drinking they found. As- suagement drinking, done to drown one's sorrows, and retal- iation drinking, aimed at harming others, they found to be uncommon . Personal Motives and Drinking A number of studies have shown higher alcohol consump- tion associated with personal meanings and motives for drinking rather than social meanings and motives (Cahalan, Cisin & Crossley, 1969; Mulford 6 Miller, 1960a; Riley, Marden G Lifshitz, 1948). Investigations into the relation of personality dimensions and drinking levels provide a means of insight into psychological bases of drinking be- havior. Heavier drinking involves aspects of personality which have been called "personal" or "immature”. Previous studies have shown that heavy drinkers evidence differences in personality from light or nondrinkers before the devel- opment of their different drinking styles (Jessor, Collins a Jessor, 1972; Jones, 1968; Jones, 1971). These personal- ity differences can be seen as factors leading to variations in drinking behavior. Drinking Specific - Nondrinking Specific Theories Zucker (1979) identifies two types of theory about the development of drinking. Drinking specific theory, which characterizes drinking activity in terms related specifi- cally to alcohol variables (for example, the availability of alcohol). Nondrinking specific theory, on the other hand, sees the drinking process in terms that include alcohol con- sumption, but which primarily involve more central human behavior need patterns not intrinsic to drinking. He notes that, "from a deveIOpmental standpoint, nondrinking specific theories have a potentially greater ability to map conti- nuities in process over longer periods of the life span, insofar as they allow for varied phenetypic expression of underlying and presumably more enduring characteristics" (p. 10). The present study takes the viewpoint of a non- drinking specific theory. College Population A college pOpulation is of particular interest in studying drinking behavior since it is during this age of young adulthood (18-22 years) that drinking quantity typi— cally reaches a peak (Harford 8 Mills, 1978; Vogel-Sprott, 1974). This age range also has the highest percentage of drinkers, both of men and women (DHEW, 1974). Zucker (1979) in referring to these statistics of age and drinking states that, "In short, from a drinker vs. nondrinker standpoint, cross-sectional evidence indicates that old age (or a pat- tern similar to it) starts around age 22 and continues unabated from there on" (p. 44). Sex Differences Differences in male and female problem drinkers have been reported by many investigators (Beckman, 1975; Curlee, 1970; Lisansky, 1957; Rimmer, Pitts, Reich 6 Winokur, 1971). Jones (1971) suggests that women who drink excessively are more unlike the average female drinker than men problem drinkers are unlike the average male drinkers. This is con- sistent with the notion that for women, the social prescrip- tions against excessive drinking are greater. At present the differences in drinking behavior be- tween men and women remain substantial (DHEW, 1974; Harford 8 Mills, 1978; Wechsler, Demone G Gottlieb, 1978). Indica- tions of a decreasing of the historic difference in drinking practices between males and females, particularly among teenagers, have been reported recently (Wechsler G McFadden, 1976; Wilsnack G Wilsnack, 1978). Keil (1978), in summarizing studies of men's and wo- men's drinking, reports the general results that women have been found to be less likely than men to drink and to con- sume less on occasions that they do drink. He suggests, however, that as sex roles continue to change in American society, women can be expected to drink more and to drink more frequently. Differentiation and Alcohol Use Psychological Differentiation, which was originally considered in a more limited form as field.dependence/inde- pendence, has received wide attention in studies of alco- holism. People who are less differentiated tend to have dif- ficulty in delimiting an object from its surrounding visual field, as on the Rod and Frame Test (Witkin, Lewis, Hertzman, Machover, Meissner 6 Wapner, 1954). The less differentiated display a limited capacity for analyzing their internal pro- cesses and depend a good deal on external cues. Psycholog- ically differentiated people, by contrast, perceive an ob- ject as distinct from its field and are more independent of the influence of context (Witkin, Dyk, Paterson, Goodenough & Karp, 1962; Witkin, 1965). The results of numerous studies (Harley, Cohen & Silverman, 1974; Hoar, 1977; Karp, Witkin & Goodenough, 1965a, 1965b; Witkin, Oltman, Cox, Ehrlichman, Hamm a Ringler, 1973) have shown alcoholics to be a field dependent and less differentiated population than their non- alcoholic counterparts. Moreover, a number of studies show that alcoholics manifest a significant decrease in field dependence following a treatment program (Chess, Neuringer a Goldstein, 1971; Goldstein 8 Chotlos, 1966; McWilliams, Brown & Minard, 1975). Extension and Alcohol Use Lolli (1956) characterized alcoholism as a disorder of the love disposition. He described alcoholics as having difficulty in reaching out to love another. Bacon (1974), in considering the results of a large-scale cross-cultural study of drinking, pointed to the need for an affiliative dimension in the study of alcohol use. Other studies re- lating problem drinking to facets of alienation (Blane, Hill a Brown, 1968; Kane 8 Patterson, 1972), social deviance (Jessor, Graves, Hanson G Jessor, 1968), and weak family af- fectional ties (Zucker, 1979) implicate conceptualizations that we would term low extension of the self. Jessor, Carman and Grossman (1968), in their study of drinking patterns of college students, point out that, along with academic achieve- ment, the goal of social affection or inter-personal liking is one of pervasive importance for students. In addition, they note that, drinking alcohol is one of the learned activities available to college students for dealing with disappointment in this area. The Problem The primary purpose of the present study is to explore the relation of Smith's (1979) composite theory of personal- ity to drinking in a college student pOpulation. The two major dimensions (Figure l) of the composite theory are differentiation and extension. Smith (1979) de- fines differentiation as the degree to which a person senses a clear and articulated distinction from the personal and physical environment. A differentiated person recognizes clear differences between self and others. Smith (1979) de- fines extension as the degree to which a person senses a feeling of being united with others. An extended person feels close to other peOple. Smith's (1979) picture of develop- ment (Figure 1) shows a growing organization of the self in the direction of increasing differentiation and extension. This is seen as a progression of maturation. The notion that alcohol consumption and maturity are connected was investigated by Jung (1977). He surveyed male and female college students to determine the relationship between the Jessor, et al. (1968) quantity - frequency index of drinking and the maturity level of motives for drinking. Jung found that both male and female students with immature motives for drinking (e.g., to get "smashed") drank about 10 three times as much as those with mature motives (e.g., just to be friendly). Men drank about twice as much as women regardless of motive. Summary and Hypotheses Indications that greater drinking is associated with a diminution of differentiation and extension have been re- viewed. Looking at alcohol use, and its relation in the deve10pmental process within the framework of the composite theory, gave rise to the following hypotheses: 1. Persons classified as low on differentiation will have greater alcohol use scores than those clas- sified high on differentiation. 2. Persons classified as low on extension will have greater alcohol use scores than those classified as high on extension. 3. Persons classified as low on both differentiation and extension will have greater alcohol use scores than all others. Congruent with alcohol use history, it is hypothesized that: 4. Males will have greater alcohol use scores than females. For discrete categorizations of drinking, it is hypoth- esized that: 5. Differentiation and extension measures will dis- criminate among the drinking pattern groups. 11 Differentiation and extension measures will dis- criminate among the light, moderate, and heavy drinking groups. Differentiation and extension measures will dis- criminate among the abstainer, light, moderate, and heavy drinking groups. METHOD The principle aim of the present study was to test hypotheses relating the composite theory of personality (Smith, 1979) to college student drinking. A survey of campus residence halls was undertaken to gather data for measuring the differentiation and extension dimension of the composite theory and for determining the amount of alcohol use by students. Subjects Of the 600 questionnaire booklets distributed to stu- dents in the campus residence halls of Michigan State Uni- versity, 495 were returned (83%). This sample population, by college year and sex, is displayed in Table 1. Sample Selection A modification of the procedure described by Stratou- dakis (1976) for sample selection from residence halls at Michigan State University was used as follows: a printout of the campus residence hall pOpulation by college year and sex was used to select a residence hall from each of the four major residence complexes (north, south, east, and west) that, in the judgement of the present investigator, had an 12 13 TABLE 1 Number of Subjects by College Year and Sex College Year Males Females Total Freshman 70 121 191 SOphomore 78 78 156 Junior 42 48 90 Senior 21 15 36 T73 Note: 22 respondents did not report college year or sex. optimal balance of males and females and an optimal repre- sentation by college year and which, according to the di- rector of residence hall programs, had not recently been overly used for survey research. Students present in their rooms of the selected residence halls were asked to partici- pate in the study, on a voluntary basis, by completing a questionnaire booklet (see Appendix A for booklet and in- structions used). Approximately one hour after distribution the questionnaires were personally collected at the students' rooms. The sample population by residence hall.and ethnicity is shown in Appendix B. 14 Instruments The questionnaire booklet (reproduced in Appendix A) contains the following: (1) Drinking Inventory, (2) Atti- tudes Toward Leadership, (3) Preference Inventory, (4) Inter- personal Check List, and (5) Demographic Items. The study utilized two measures of differentiation: (1) The differentiated role score from Attitudes To- ward Leadership. (2) The dissimilarity score from the Preference In- ventory (Note 1). And two measures of extension: (1) The people orientation score from Attitudes To- ward Leadership. (2) The Lov score from the Interpersonal Check List. The Drinking Inventory yielded three measures: (1) A score giving the amount of absolute alcohol consumed per day. (2) An index for pattern of alcohol consumption. (3) An index for category of drinking. Measuring Alcohol Consumption In a comparison of three common methods of describing alcohol consumption Little, Schultz, and Mandell (1977) pro- posed "a new scoring system which not only combines the strength of the three methods ... but also preserves much information on volume and variability of alcohol consumption" (p. 554). 15 The three methods they examined were: (1) the Quantity-Frequency—Variability Index, and (2) the Volume- Variability Index, both devised by Cahalan, Cisin, and Crossley (1969); and (3) the Jessor, Graves, Hanson and Jessor (1968) method of estimating average daily ounces of absolute alcohol ingested. Little, et a1. (1977) called this latter average the absolute-alcohol (A-A) score. In addition, Little and her colleagues devised a pattern of consumption measure they called the Quantity-Pattern (Q-P) index. This Q-P measure is a four point scale designed to detect a pattern of heavy consumption. Little, et al. (1977) described this A-A -- Q-P index as "an estimator which per- mits maximum information on alcohol consumption to be re- tained without recourse to the original instrument. Its flexibility and precision allow a simple yet thorough des- cription of reported drinking behavior" (p. 562). The method for measuring alcohol consumption used in this study follows that suggested by Little, et a1. (1977). The A-A score, based on the Jessor, et al. (1968) scoring procedure, is used to measure aggregate volume, and the pat- tern of consumption is identified by the Q-P index. The aggregate volume of total ounces of alcohol ingested per day is a composite score for the three beverages. Since the data included a range of consumption, and since beer, wine and liquor differ in the amount of absolute alcohol, a rather complicated scoring procedure was used to calculate 16 the absolute-alcohol (A-A) score. (See Appendix C for de- tailed computational procedure of A-A score). The average ounces of absolute alcohol consumed per day is referred to by Jessor, et a1. (1968) as an individ- ual's Quantity-Frequency or "Q-F” score; and referred to by Little, et al. (1977) as an Absolute-Alcohol or "A-A” score. Because quantity-frequency has at times been used differ- ently by different writers, the term drinking score will be used in this study to refer to this calculated composite absolute alcohol (A-A) score. The Inventory of Drinking Behavior The inventory of drinking behavior (see Appendix A) used in this questionnaire survey is from: Scale 10: Drinking, of the "Community Survey Interview Schedule” in Jessor, et a1. (1968). Frequency data were obtained for each of the three major beverage types by the following categories: Three or more times a day Two times a day About once a day Three or four times a week Once or twice a week Two or three times a month About once a month Less than once a month, but at least once a year Less than once a year Never 17 "This order of presentation...is intended to make it eas- ier for fairly regular daily drinkers to admit to their high frequency by suggesting that ever higher frequencies among some subjects were anticipated..." (p. 167). Following the frequency question were three quantity questions concerning "5 or more drinks", ”3 or 4 drinks", and "l or 2 drinks", for each beverage type, with the fol- lowing categories: Nearly every time More than half the time Less than half the time Once in a while Never In describing this procedure Jessor, et a1. (1968) re- marked that: The advantages of such an obviously time-consuming ap— proach are several. First, and most general, by break- ing down a complex behavioral syndrome such as drinking into several component parts, the accuracy with which the respondent was able to recall his own behavior may have been increased. Second, it enables a beverage-by- beverage analysis of drinking habits. Third, it differ- entiates the pattern or regularity of consumption of varying amounts of a beverage enabling a distinction to be made, for example, between a drinker who alwa 5 drinks moderately and one who is usuall a l1ght drinker but sometimes drinks heavily. (p. 168) Drinking Score The measures of alcohol use were obtained by asking, for each type of beverage (beer, wine, spirits): (1) how of- ten (frequency) the beverage is drunk, (2) how much at one time (quantity) the beverage is drunk. The quantity responses were converted to units of absolute alcohol and multiplied by 18 a weighted frequency to give an absolute alcohol score for each type of beverage. The three beverage scores summed yield a total absolute alcohol score. This drinking score represents an average amount of absolute alcohol consumed per day. . Jessor, et a1. (1968) point out that this composite measure of volume of alcohol has a number of distinct vir- tues: 1. It summarizes, in a single figure, quantity, fre- quency, and range of consumption of beverages dif- fering in alcoholic content; 2. this figure, average ounces of absolute alcohol consumed per day, is readily comprehensible and can easily be translated into other meaningful units, such as "beers" or "drinks", 3. the score permits the calculation of group statis- tics such as means and medians which are not pos- sible when discrete descriptive categories are em- ployed; and, 4. the score is continuously distributed along a single dimension permitting correlational analyses, tests of differences between group means, and other stat- istical comparisons. (pp. 168—169) Quantity-Pattern Index While they alluded to this scale's ability to reflect pattern of drinking, Jessor, et a1. (1968) only used 3 vol- ume measure of alcohol consumption. It was Little, et a1. (1977) who devised a method to score this scale for pattern of alcohol consumption. The Q-P categories indicate the average frequency of heavy drinking over the time measured. 7 The Q-P index values provide an ordered scale of four points reflecting how frequently a drinker has five or more drinks 19 on a single occasion. Category A = zero times, Category B = more than none but less than one time, Category C = one or more but less than 4.2 times, Category D = 4.2 or more times. The value 4.2 was chosen so that subjects reporting five or more drinks "once or twice a week" "more than half the time" are included in Category D (Little, et al. 1977, p. 559). Drinkinngategories In order to classify amount of drinking into discrete categories an index devised by Yano, Rhoads, and Kagan (1977) was employed. Their index of drinking is based on convert- ing beverage intake to the amount of absolute alcohol con- sumed per day. The drinking score computed for this present study can be directly converted to this index. The categor- ies, called light, moderate, and heavy by Yano, et al. (Note 3), were defined as follows: Light drinkers consume some, but not exceeding 1/3 ounces of absolute alcohol per day. Moderate drinkers consume more than 1/3 ounces but not exceeding two ounces of absolute alcohol per day. Heavy drinkers consume more than two ounces of absolute alcohol per day. Putting these categories roughly in everyday terms, the light drinker would have less than one liquor based drink, bottle of beer or glass of wine a day. The moderate drinker would 20 have approximately one to three liquor based drinks, bottles of beer or glasses of wine per day. The heavy drinker would have approximately four or more liquor based drinks, bottles of beer or glasses of wine per day. The term "heavy drinking" has been criticized as per- haps suggestive of problem drinking (Marden, Zylman, Fillmore & Bacon, 1976). In this present study we are using the terms "light", "moderate", and "heavy" to refer only to rela- tive position, each progressive category indicates more drinking than the previous one. Thus, the category "heavy drinking" is not synonymous with problem drinking or alco- holism. Verdicality of Drinking Inventory In order to maximize the truthfulness of respondents' self-reporting, they were assured complete confidentiality and anonynimity. No names or student numbers were recorded on the questionnaire or answer sheets and only summary stat- istics were used. Following the practice of Jessor, et a1. (1968) the drinking inventory was placed near the end of the question— naire in the hope that the respondent would have become used to answering questions of a personal nature. Respondents were asked questions of the type: "how often" a particular drinking behavior was engaged in, rather than "have you ever”. Thus, simple acquiescence was not sufficient to 21 complete the task; the respondent had to give consideration to the frequency of engaging in various drinking behaviors. Although there is reasonable concern with accepting at face value the alcohol use score as a wholly accurate repre- sentation of the amount of drinking taking place in the col- lege residence hall community, Jessor, et a1. (1968) note that in this measure: If we make the assumption that the distorting effects of memory and self-protection have a roughly equal influence upon all respondents, then we should at least be able to order the respondents fairly accurately with respect to their drinking behavior. Even without such an assump- tion, groups of respondents can probably be ordered ac— curately Since individual differences in amount of dis- tortion can then be readily assumed to cancel each other out. The actual scores reported by the respondents are simply considered as s mbolic of the extent of the par- ticular behavior being studied. (P. 179) Measures of Differentiation andTExtension The definitive characteristic of differentiation is that of distinguishing oneself from others. The scale used in this study as the primary measure of differentiation was the differentiated role scale from Attitudes Toward Leadership (Smith, 1973). The scale used as the secondary measure of differentiation was the dissimilarity score from the Pref- erence Inventory (Smith, 1973). The definitive character- istic of extension is feeling towards others in a concerned and united way. The scale used in this study as the primary measure of extension was the people orientation scale from 22 Attitudes Toward Leadership. The scale used as the secon- dary measure of extension was the Lov scale from the Inter- personal Check List (LaForge, 1973). Attitudes Toward Leadership This is a thirty-two item instrument requiring the respondent to pick which of two statement pairs is more important for a leader to do. The first sixteen items pro- vide a measure of peOple orientation, and are scored for those responses chosen where the leader shows a positive consideration for the work group. The second sixteen items provide a measure of role differentiation, and are scored for those responses chosen where the leader is responsible for a differentiated role in the work group. Both scales have reported internal consistency reli- abilities above .80 (Smith, 1973). Preference Inventory This test was designed by Smith (1973) to provide a measure of assumed similarity to others. It contains twenty four self-description items of the true-false type. To answer the question, what is a person who generally assumes a great deal of similarity like?, Smith (1974) used the following procedure: From a group of several hundred college students, 25 who consistently assumed a great deal of similarity were matched with 25 of the same sex and intelligence who consistently assumed little similarity. Both groups 23 answered the 200 statements in the five-trait scale [of a large personality inventory]. The answers of the empathic [assumed similarity] and the nonempathic groups were compared and the 48 statements that they answered most differently were isolated. These 48 statements were answered by a second group of empathic and non- empathic students. The 24 statements [chosen] ... are those that most sharply separated the high-empathy from the low-empathy groups. (p. 160) In scoring the scale for use in this study as a sec- ondary measure of differentiation, the items were reflected in the direction of dissimilarity. Thus, a respondent's high score is in the differentiated direction. The alpha reliabilities computed from the present data for the Preference Inventory and the two scales from Atti- tudes Toward Leadership are presented in Appendix D. Interpersonal Check List The Interpersonal Check List "is a l34-item list of words or phrases which may be used to obtain self-description or description of others with respect to an interpersonal domain of personality” (LaForge, 1973, p. l). The summary factor Lov (love-hate) used in this study as a secondary measure of extension was obtained by calculating a weighted sum of a person's set of responses to the list items con- stituting the Lov scale. Internal consistency reliabilities (based on commonalities) for self description of the Lov scale are .90 (LaForge, 1973). The Interpersonal Check List was scored using a computer program developed by Krus and Krus (1977). 24 Design and Analysis In this study we are investigating college student drinking and two exploratory measures of extension along with two exploratory measures of differentiation. In order to test hypotheses relating drinking score to high and low levels of differentiation and extension, as well as their interaction and the effect of sex, an analysis of variance design was used. The assumption was that greater mean alco- hol use was associated with both low extension and low dif- ferentiation factor levels, and that men would show greater mean alcohol use than women. In order to test hypotheses relating differentiation and extension to categories of drinking pattern or to discrete categories of drinking, discriminant analysis was used. The assumption was that students belonging to different drinking patterns or to dif- ferent drinking category groups could be distinguished by their differentiation and extension scores. Analysis of Variance The analysis of variance (ANOVA) is concerned with the analysis of variation about means. ANOVA consists of par- titioning the total variation present in the data set into appropriate components of interest and error. In order to test for effects of high and low extension and of high and low differentiation, the medians of these respective scores were used to divide subjects into above the median (high) and below the median (low) components. 25 Discriminant Analysis "Discriminant analysis begins with the desire to sta- tistically distinguish between two or more groups of cases. These 'groups' are defined by the particular research sit- uation" (Nie, Hull, Jenkins, Steinbrenner G Bent, 1975, p. 435). The differentiation and extension measures, both pri- mary and secondary, are the variables on which the quantity- pattern groups and the drinking category groups are hypoth- esized to differ. In discriminant analysis, these variables, the discriminating variables as they are called, are lin- early combined into "discriminant functions". Mathemati- cally, the objective of discriminant analysis "is to weigh and linearly combine the discriminating variables in some fashion so that the groups are forced to be as statistically distinct as possible" (Nie, et al., 1975, p. 435). Once the discriminant functions are derived, this technique provides several means for determining their significance, number, and importance. Nunnally (1978) in discussing the multiple discrim- inant function (MDF) states that, "both conceptually and mathematically the MDF constitutes a powerful tool which has not been employed nearly as much as it should have been in the behavioral sciences" (p. 464). His view is that in applied problems discriminant analysis is more fruitfully used in the areas of noncognitive attributes such as 26 interests or with personality characteristics and attitudes. Such is the case in the present study. Statistical Treatment As shown in Appendix B, the number of nonwhite, non- American born respondents is small. In order to allow for a more clear-cut presentation of results these cases were excluded from the statistical analyses. Since the raw drinking scores were highly skewed, the transformation, fourth root of the drinking score, was used to bring about a good approximation to normality. A 2 X 2 X 2 factorial unequal cell analysis of variance de- sign for the three independent variables of differentiation (high, low), extension (high, low), and sex (male, female) was used with the dependent variable fourth root of drinking score. A separate ANOVA was done for the primary measures of extension and differentiation and for the secondary mea- sures of extension and differentiation. All main effects and interactions were tested with the residual error term. The four variables used for the set of discriminating variables in the present analysis are the two measures of differentiation and the two measures of extension. This present study has been primarily concerned with college student drinkers. Discriminant analysis was used to see if measures of differentiation and extension could dis- tinguish: (l) drinking pattern categories, and (2) drinking 27 categories (light, moderate, heavy). Additionally, dis- criminant analysis was used to look at possible distinctions among abstainers, light drinkers, moderate drinkers, and heavy drinkers with regard to measures of differentiation and extension. The analyses of variance, discriminant analyses, and reliability computations were performed by using the Statis- tical Package for the Social Sciences (Nie, et al., 1975) on the CDC 6500 computer system at Michigan State Univer- sity. Statement of Hypotheses The statistical treatment dealt with seven hypotheses. Restated, these hypotheses were: Hypothesis 1. Among subjects who drink alcoholic beverages, the mean drinking score will be lower for those classified high in differentiation compared with those classified low in differentiation. Hypothesis 2. Among subjects who drink alcoholic beverages, the mean drinking score will be lower for those classified high in extension compared with those classified low in extension. Hypothesis 3. Among subjects who drink alcoholic beverages, the mean drinking score will be lowest for those classified high in both differentiation and ex- tension compared with all other classifications. 28 Hypothesis 4. Among subjects who drink alcoholic beverages, females will have lower mean drinking scores compared with males. Hypothesis 5. Among subjects who drink alcoholic beverages, differentiation and extension measures will discriminate among groups identified by drink- ing pattern categories. Hypothesis 6. Among subjects who drink alcoholic beverages, differentiation and extension measures will discriminate among groups identified by drinking category (light, moderate, heavy). Hypothesis 7. Differentiation and extension measures will discriminate among groups identified as ab- stainers, light drinkers, moderate drinkers, and heavy drinkers. RESULTS The primary purpose of this study was to test assump- tions relating college student drinking and measures of differentiation and extension. The hypotheses that a high level of differentiation and a high level of extension would be related to lower drinking scores were tested using anal- ysis of variance (ANOVA). The ANOVA design also provided for testing interactions as well as the influence of sex differences in alcohol use. An examination of the ability of differentiation and extension variables to "discriminate" between groups iden- tified by drinking pattern and drinking category was carried out using multiple discriminant analysis. Along with the derivation of the discriminant functions, multiple discrim— inant analysis provides tests for statistically determining the success with which the variables used actually discrim- inate. Additionally, a look at aspects of the data intended to articulate the sex difference findings is provided. This information will, hopefully, round out the picture of college student drinking drawn from the present investigation. 29 30 Drinking Score, Primary Measures of’Differentiation and Extension, and—Sex A 2 X 2 X 2 factorial unequal cell analysis of vari- ance with differentiation (high, low), extension (high, low) and sex as the independent variables was used to test hy— potheses one through four on the primary differentiation and extension measure. The transformation fourth root of drink- ing score was used as the dependent measure. Table 2 shows the means and standard deviations for subject's raw drinking score by level (high, low) of the primary measures of dif- ferentiation and extension and sex. The analysis of vari- ance summary is displayed in Table 3. TABLE 2 Means and Standard Deviations of Raw Drinking Score by Primary Measures3 of Differentiation and Extension and Sex Mean SD N Sex Female .69 .74 203 Male 1.05 1.31 173 Differentiationp Low .83 1.03 199 High .88 1.09 177 Extensionp Low .97 1.21 216 High .69 .78 160 Total .85 1.06 376 aSubscript p in table indicates primary measure. 31 TABLE 3 Analysis of Variance Summary of Fourth Root of Drinking Score for Primary Measures3 of Differentiation and Extension and Sex Source df MS F Differentiationp l .002 .039 Extensionp l .290 4.927* Sex 1 .419 7.120** Differentiationp é Extensionp l .005 .079 Differentiatio 5 Sex 1 .196 3.339 Extensionp i Se 1 .000 .005 Differentiationp x Extensionp l .035 .600 y Sex Error 368 .059 aSubscript p in table indicates primary measure. *Probability less than .05 **Probability less than .01 The analysis of variance disclosed significant main effects for extension and sex. None of the interactions (including hypothesis three) were significant. Nor was the effect of differentiation (hypothesis one) significant. Hypothesis two was confirmed. Those subjects classified high on the primary measure of extension (based on the peOple orientation scale from Attitudes Toward Leadership) had sig- nificantly lower mean drinking scores than those subjects classified as low on extension. Hypothesis four was also confirmed; men having a highly significant greater mean 32 drinking score than women. These results support the notion that greater drinking is associated with a low level of ex- tension. The results also support the commonly held view that on the average, men drink more than women. Drinking Score, Secondary Measures of Differentiation and Extension, and Sex This examination of the secondary measures of differ- entiation (high, low), extension (high, low), and sex em- ployed a 2 X 2 X 2 factorial unequal cell analysis of vari- ance to test hypotheses one through four. The transformation fourth root of drinking score was used as the dependent var- iable. Table 4 shows the raw drinking score means by level (high, low) of the secondary measures of differentiation and extension and sex. The analysis of variance summary dis- played in Table 5 reveals a significant main effect for sex only. None of the other main effects or interactions reach significance (hypotheses one through three). Hypothesis four, that men have higher mean drinking scores than women, is again confirmed. Pattern of Drinking The Quantity-Pattern index provides a four point order- ed scale that reflects a subject's pattern of heavy drink- ing. These four pattern groups (A, B, C, D) can be charac- terized as follows: A = null, B = slight, C = moderate, and D = a pronounced pattern of heavy drinking. Discriminant 33 TABLE 4 Means and Standard Deviations of Raw Drinking Score by Secondary Measuresa of Differentiation and Extension and Sex Mean SD N Sex Female .69 .74 203 Male 1.05 1.31 173 Differentiation Low .77 .99 209 5 High .96 1.13 167 Extension Low .95 1.11 188 5 High .76 .99 188 Total .85 1.06 376 aSubscript s in table indicates secondary measure. analysis was used to determine whether, for drinking sub- jects, the four pattern groups differ with regard to dif- ferentiation and extension (hypothesis five). As signifi- cant sex differences in drinking were shown above (hypothe- sis four), all discriminant analyses were carried out for males and females separately. The means and standard deviations, for male subjects, of the primary and secondary differentiation and extension scores for the drinking pattern (Q-P Index) categories are presented in Table 6. The discriminant analysis of male subjects for the differentiation and extension variables and the drinking pattern groups is shown in Table 7. Since there are four groups, it is possible to derive three 34 TABLE 5 Analysis of Variance Summary of Fourth Root of Drin ing Score for Secondary Measures of Differentiation and Extension and Sex Source df MS F DifferentiationS l .167 2.849 Extension 1 .047 .799 Sex 1 .370 6.292* Differentiation x Extensions 1 .088 1.505 DifferentiationS x Sex 1 .061 1.044 Extensions y Sex — 1 .015 .249 DifferentiationS x Extensions 1 .079 1.348 5 Sex Error 368 .059 aSubscript s in table indicates secondary measure. *Probability less than .01 discriminant functions. As each function is derived, start- ing with no (zero) functions, Wilks' lambda is calculated. Wilks' lambda can be used to test for the significance of discriminating information still remaining in the original variables and not yet removed by the discriminating func- tions. The larger the Wilks' lambda, the less information residual. Lambda can be transformed into chi-square for a statistical test of significance. The eigenvalues, and their associated canonical correlations computed in the derivation of the discriminant functions, signify the rela- tive capability of each function to separate the groups. 3S .oASmmoE 5pmwcooom mopmofiwcfi m paflgomQSm ”693mmoe 5hmeflum moumowwcfi a unapomnzmw mm HN.5 aw. oo.m H5.HH mH.m mo.w mH.m Q5.w a me 5m.o mN.H wo.m mN.HH mH.m om.w mm.m om.w u m 55.5 No. mv.m mm.m mm.e mm.m oo.m mo.5 m mm Ho.m mm. oo.m ow.o~ w5.N om.m m5.m o¢.w < am cam: mm :moz mm :mmz mm cmmz m m Q a z :oflmcouxm cofiumflucohmmmfia :ofimamuxm :oHumflucowmwmfl: :pouumm wcfixewuo mowuomoumu mxoecH muov chopped mcflxcflpa pow :ofimcopxm was :oflumflucmpmmwflm wo WGHNE WC WCOHHMM>®Q @HGUCMHm Uflm mfimmz mmmHSmmmz 5hmwcoomm mam xumefium :o o m4mfipom :oflumaoupou 63Hm>cmwflm :ofluocsm .mxafiz macapocsm Hmoflcoamu ucmcfisflgomflo mmuoonnsm mamzu mowpomoumu Axoch muov chopped mcfixcdha wow scamcouxm wan :oflumfiucopomWAo mo moHSmmoz 5pmwcooom vcm 5Hmepm mo mflm5amc< pcwcflefipomfla 5 m4mma epmwcmum was mane: w mqmwwoo :oflumaopwou 63Hm>comflm :oflpocsm .mxafiz mcofipocsm Hmowcocmu acmcfiefipomflm amuoonnsm mHmEomv mmfihommumu mxmecH m10v :Houumm wcflxcwha How :oflmcoaxm cam :oflumflucopommfi: mo moHSmmoz zhmmcooom wan 5gmeflhm mo mfim5amc< ucmcfleflpomfla m mqmmmz 45.6 H6.H mm.~ ae.HH mN.m “a.” q¢.m mN.m mampmeoz om.5 5s. maum mo.OH Hm.m wo.o Hm.m m5.w bequ om cam: am 2am: am cam: am cam: m m a a cofimcouxm :ofiumfipcoeomwfla :oflmcouxm :ofiumflucopommfla zeomoumu mawxawwo 55>mm: .oumuowoz .unwfiqv 5powmumo wcflxcfipn how :oflmcouxm paw :ofiumflucopommam mo wmopsmmoz 5Hmwcooom ecu 5pmefipm :o mmfiwz mo mcofiumfi>mm mpmecmum was memo: OH mqmfipom :oflumampuou 65Hm>cowfim coauocsm .mxafiz mcofluocsm Hauwcocmu acmcfleflpomfia Amuoonnam mamzv 5>moz .oumumpoz .unwflq uzhomoumu wcflxcfluo now :ofimcouxm use :ofiumflpcouomwfio mo moNSmmoz 5pmvcooom paw zumefihm mo mflmxamc< ucmcfiefluomfla HA m4ma62 NNN ON.N Nm.m Na.N NH.NN NN.N am.N 0N.N NN.N opawoeoz Hm No.5 No.4 mo.m 95.ON Nm.m mN.a NN.N mN.w NENNN mm cam: am cam: am new: am cam: m m a a z :ofimcmpxm cofluwwucoumMMHo coflmcouxm cofiumwucoNoMMMa 5Nomoumu NcwxeNpm 55>wo: .oumuoeoz .uzmfiqv 5powopmu mcfixcwhm pom :oflmcmuxm paw cofiumflucmpmmmwn mo mmmusmmoz xhwwcoomm use 5pmefiwm :o monEom mo maoflumw>oa ppmpcwwm wan mcmoz NH mamflnmo coflumaoppou enam>cowflm cofluocsm .mxafiz macauocsm Hmoflcocmu unm:NENNUmwQ mmuoohnsm onEmmv 5>mmm .oumpmwoz .u:maq ”zpommumu mcfixcfiua pom cowmcopxm can :oflumflucouommNa mo monummoz zymwaooom mam zumeflhm mo mfimxamc< acmcfieflpomfio ma mqma62 mm 55.6 No.N mm.N 55.NN mN.m NN.” 44.N NN.N oaaNmeoz mm om.N Ne. me.m mo.ON Nm.m No.5 Nm.m NN.N NENNN N mm.m m5.oN om.m NN.ON NN.m NN.NN No.4 5m.m :Naumn< am :am: am cam: om cam: om cam: m m a m cofimcouxm :oNumwuconwwwc cofimcouxm :ofipmfiucopowwNa .mmmHu chxcflpa >>mmz .mumpmwoz .pcmNN .cflmumn< "coflumoflmflmmmao mcflxcflum Now cofimcouxm cam cofiumfiucoquMHo mo mmopsmmoz 5umecooom can 5NmEHNm :o moan: mo mcoNumN>oa ppwwcmpm paw menu: «a m4m<9 46 Noo. :mEN mmoN NNNNNQanoeaw Ho.H N «am. N u we. mmoo. m n . mw.NN o mmm. N H mm. mmoo. m ” ~ «N@.Nm NH mmw. o H mm. mNNN. N . opmscmuflsu we menEmN po>Nhoa :oNumNmppou osfim>comNu :oNuocsm .mxaflz m¢0wuo::m Hmoflcocmu pcmcfiefiuomflo Amuoomnsm onzv 5>moz .oumpmpoz .ucmNN .cNmumn< "soNumonNmmmNu wcwchpa Now :oncouxm pew :oNpmNucmwamNo mo mopsmmoz 5Nmecooom can 5NmENNm mo mNm5ch< ucchENhomNa ma mqma62 NNN ON.N Nm.m Nm.N NN.NN NN.N om.w 0N.m 55.N oumpoeoz NN 50.N No.4 mo.m NN.ON Nm.m mN.m NN.N NN.N DENNN ON ON.ON N5.¢ 55.N om.ON ms.m oo.w mo.m NN.N :Naumn< am new: om cam: mm :mm: mm cam: m m m a z coflmcouxm :oNumNpcopowwflm cofimcouxm :oflumflucohmmmfla .mmmau maNxeNNa 5>moz .muwpovoz .uszN .cflmumn< “:oNumonNmmmNu mcflxcfium pom :oncouxm mam :oNumNucopmwwN: mo mmopsmmoz 5vacoo6w paw 5NmENpm so monEom mo mcofipmfl>om ppmecmum ecu mcmoz 5H mqmtho :oNumNoppou ost>cowNm :oNuocsm .meNz mcoNuocsm NmoNcocmu NcchENpomNa Nmuoonnsm onEomv 5>moz .oumuowoz .ucmNN .cNmNmn< "coNumoNuNmmmNu chchNo Now :oncouxm can :oNumNucouommN: No mopsmmoz 5vacooom wcw 5NmENNm No mmech< acchENpomN: wN mqm