- Email: [email protected]
Differential alcohol use patterns and personality traits among three alcoholics anonymous attendance level groups: Further considerations of the affiliation profile
Drug and Alcohol Dependence, 5 (1980) 135 - 144 @ Elsevier Sequoia S.A., Lausanne --- Printed in the Netherlands
135
DIFFERENTIAL ALCOHOL USE PATTERNS AND PERSONALITY TRAITS AMONG THREE ALCOHOLICS ANONYMOUS ATTENDANCE LEVEL GROUPS: FURTHER CONSIDERATIONS OF THE AFFILIATION PROFILE* MICHAEL R. O’LEARY Department of Psychiatry and Behavioral Sciences, University of Washington, Medicine, and Veterans Administration Medical Center, Seattle, Washington
School of (U.S.A.)
DONALD A. CALSYN Department of Educational Psychology, Administration Medical Center, Seattle,
University Washington
of Washington, (U.S.A.)
and Veterans
DENNIS L. HADDOCK Veterans
Administration
Medical
Center,
Seattle,
Washington
(U.S.A.)
CHARLES W. FREEMAN Pain Clinic,
Swedish
Hospital
Medical
Center,
Seattle,
Washington
(U.S.A.)
(Received March 5,197Q)
Summary The present study was designed to identify personality and drinking behavior correlates associated with membership in Alcoholics Anonymous (A.A.). Affiliation profiles were developed on the basis of frequency of A.A. attendance; subtypes were defined as non, low, and high affiliates, respectively. An initial canonical correlation analysis assessed the amount of common variance between personality and drinking characteristics. Subsequent discriminant function analyses on the second-order scales of the drinking and personality variables sets identified those variables differentiating the A.A. affiliation groups. High levels of affiliation were associated with a greater degree of anxiety, a tendency to be more affected by feelings rather than intellect, and more deterioration in the physical, psychological, and social areas of functioning. Regular attenders of A.A. also exhibited significant differences in perceived benefits, style, and consequences of alcohol usage. The potential therapeutic utility of A.A. affiliation profiles was discussed.
--*Address reprint requests to Michael R. O’Leary Ph. D. (llGATP), Veterans Administration Medical Center, 4435 Beacon Avenue South, Seattle, Washington 98108, U.S.A. This work was supported by the Veterans Administration Health Services Research and Development Service.
136
An explicit and integral part of many inpatient and many outpatient alcoholism treatment programs is Alcoholics Anonymous (A.A.) [ 11. This self-help organization, established in 1935, claims one of the highest rates of treatment success [ 21. However, surprisingly little is known about the interacting variables of drinking pattern and personality structure which would predict successful affiliation with A.A. Early research has identified some of the psychological characteristics of Alcoholics Anonymous members in an effort to maximize the therapeutic benefits of this treatment modality [2 41 . Personal and sociologic characteristics such as mental status, religion, social class, sex, ethnicity, age, and personality traits are not randomly distributed in A.A. [ 51. Other investigators suggest that A.A. affiliates differ significantly from non-affiliates in the degree to which they exhibit emotional dependency, proneness to guilt, high egocentricity, denial, high anxiety, and general alcoholic deterioration [4,6 - lo]. Recent attempts to assess structural conflicts experienced by some alcoholic patients concurrently engaged in group therapy and A.A. further point to the need of classifiying A.A. members in terms of known psychodiagnostic entities. Fontana et al. [ 111 suggest that substantial conflict arises out of the patient’s preferred ways of understanding the world. They found that individuals orientated toward understanding the world in terms of definite categories and typological labels (formistic thinking) found A.A. to be more appealing than group theory. Although this approach to subtype identification of A.A. affiliates is meaningful, it remains a controversial area. In an attempt to replicate previously found alcoholic MMPI personality subtypes, Donovan et al. [12] stated that “psychologically meaningful personality subtypes could be identified within alcoholic or alcohol-abusing populations, although investigators differ as to the number and appropriate composition of such subtypes”. Similarly, others have noted the difficulties encountered in identifying relationships between personality characteristics and drinking measures [131. A number of studies have attempted to generate alcoholism typologies from the MMPI [ 141 the 16 Personality Factor Questionnaire (16-PF) [15], and the Clinical Analysis Questionnaire [ 161 . These findings lend support to the contention of Wanberg et al. [ 171 that the unitary trait model of alcoholism is of limited heuristic value. Wanberg et al. suggest that alcohol abuse is a multiple syndrome disorder exhibiting a great degree of variability within and across subjects with respect to symptom nature and manifestation. Similarly, a multi-dimensional approach to the explanation of alcoholism provides a greater amount of information concerning inter- and intrasubject differences in the perceived benefits which motivate drinking, the nature of drinking patterns, and the specific consequences resulting from alcohol abuse [ 121. The primary purpose of the present study was to determine the interactive effects of personality, psychopathology and drinking pattern correlates with successful affiliation with A.A. It was thought that identification
137
of those variables which discriminated between patients’ level of affiliation with A.A. would provide treatment personnel with a more informed basis for treatment planning. Specifically the following questions were addressed: (1) What is the relationship between drinking style, personality, and psychopathology? (2) If the common variance between these variables is accounted for, do they distinguish between groups on the basis of their A.A. attendance? Method Subjects. The sample consisted of 76 male veterans participating in an inpatient alcoholism rehabilitation program at the Seattle Veterans Administration Medical Center. None of the subjects exhibited physical or cognitive residuals of acute alcohol intoxication at the time of testing. The sample had a mean age of 42.9 years (S.D. = 12.5) and a mean educational level of 12.5 years (S.D. = 2.2). The mean socioeconomic status of the sample wassowermiddle class. The reported age of first drink was approximately 15 (X = 15; S.D. = 4.8); problematic drinking began at age 29 (_%= 29.4; S.D. = 11.3). Fifty-eight percent of the subjects had previously been involved in alcoholism treatment, with the average number of treatments for the entire sample equalling 2.5 (SD. = 3.7; range = 0 - 20). None of the subjects had primary diagnoses of organic brain syndrome or psychiatric disorders. All subjects provided informed consent to participate in the present research after the nature of the project was fully explained. Materials. The subjects were administered the Alcohol Use Inventory (AUI) [ 171, to assess the level of drinking chronicity. The AUI, which has evolved from continued psychometric refinement of the Drinking History Questionnaire [ 181, consists of factor analytically derived scales (16 firstorder, 5 second-order, and 1 third-order) that assess symptoms related to drinking, drinking-related behavior, and perceived benefits derived from drinking. The scales of the AU1 have demonstrated high levels of internal consistency and test-retest reliability based on a normative group of over 2000 subjects. Similarly, the AU1 has demonstrated a high degree of construct and predictive validity with respect to treatment outcome and clinical judgments concerning a variety of drinking-related variables. Personality trait assessment was made by administering the Clinical Analysis Questionnaire (CAQ) [ 191 to the present sample. The CAQ is a psychodiagnostic questionnaire that includes the 16 primary personality source traits of the 16-PF, and a 12-factor psychopathology supplement. Serving as a clinical diagnostic instrument, the CAQ allows assessment of psychopathology dimensions that are measured separately from normal personality structures. From both the AU1 and CAQ second-order factors [ 201 are obtained which are linear combinations of the primary scales. In an effort to increase variable stability, all analyses were carried out on second-order factors.
138
Procedure. The subjects were administered the CAQ and the AU1 along with other self-report inventories. Social and demographic status of subjects prior to admission was recorded and included educational, occupational, and familial items, as well as drinking history, and adjustment prior to hospitalization. Individual affiliation profiles were constructed for each subject on the basis of their response to statements describing past experience with A.A. Subjects acknowledging some experience with A.A. were regarded as affiliates, and those patients who had no prior experience were characterized as non-affiliates. Further, subjects having attended a few times (less than once every 6 months), or only sporadically in the past (once or twice every 3 months), were regarded as low affiliates. Finally, high affiliates were designated as those patients who had attended regularly most of the time (at least twice a month over the past year) or who demonstrated this pattern in years past but not recently. The three A.A. attendance groups did not differ significantly on any of the following demographic variables: age (F = 2.74; d.f. = 2, 76), education (F = 1.82, d.f. = 2,76), socioeconomic status (F = 1.33, d.f. = 2, 74), estimated WAIS Full Scale IQ from the Shipley Institute of Living Scale (F = 0.11, d.f. = 2,75), marital status (x2 = 1.85, d.f. = 2), employment status (x2 = 0.96, d.f. = 4), court referral (x2 = 1.85, d.f. = 2), age of first drink (F = 0.78, d.f. = 2,76), and age patient felt he became alcoholic (F = 2.6, d.f. = 2,73). Non A.A. affiliates were less likely to have had prior alcoholism treatment (x2 = 14.8, d.f. = 2, p < 0.001).
Results The analysis was designed to answer the two questions put forth in the introduction. A canonical correlation analysis [ 211, using the secondorder factors from the CAQ and AU1 as variable sets,was used to address the question of relationship between drinking characteristics, personality and psychopathology. The summary table from the analysis is presented in Table 1. Only the first canonical variate is significant, yielding a canonical correlation of 0.74. Listed in Table 2 are the coefficients for the first canonical variate. High loadings on factor Qn from the CAQ and Factor C from the AU1 suggest that the first canonical variate might best be thought of as measuring anxiety. These findings suggest that the CAQ and AU1 overlap in measuring anxiety, which accounts for about 55% of the variance between the two instruments. Outside of anxiety, however, the two measures do not appear to overlap at a significant level. A two-step process was used to answer the question of differential drinking and personality characteristics among the three patient groups when the common variance between the CAQ and AU1 is accounted for. The first canonical variate score for each subject was calculated using the coefficients listed in Table 2. Two separate stepdown discriminant function analyses
139 TABLE 1 Summary table of cananical correlation for Clinical Analysis Questionnaire with Alcohol Use Inventory secondorder factors Canonical variates
Canonical correlation
First Second Third Fourth Fifth
0.74 0.52 0.41 0.23 0.20
sz
-.. ---__
-_-.-.-
95.41 41 .oo 19.38 6.68 2.67
d.f.
P
45 32 21 12 5
0.000 0.132 0.561 0.878 0.750
TABLE 2 Coefficients for the first canonical variate from the secondorder Analysis Questionnaire and Alcohol Use Inventory Variables CAQ secondorder scales Qr (Extroversion)
&II (Anxiety)
Qm (Tough poise) Qrv (Independence) &VI (Superego) &vu (Socialization) Qvm (Depression) QIx (Psychoticism) Qx (Neuroticism) AU1 secondorder scales AU1 A (Self enhancing drinking) AU1 B (Obsessivesustained drinking) AU1 C (Anxiety) AU1 Dl (Alcoholic deterioration-l) AU1 D2 (Alcoholic deterioration-2)
scales of the Clinical
Coefficients
-0.28 0.99 m-O.32 0.01 -0.35 0.33 0.03 0.06 m-O.52 0.08 0.16 0.88 0.17 0.00
were then performed [ 211. In both analyses A.A. affiliation level served as the dependent classification grouping. The first canonical variate score was forced first into each discriminate function analysis along with AU1 and CAQ second-order factors respectively. Listed in Table 3 are the means, standard deviations and univariate F values for all of the variables in the two analyses. The summary tables from the two discriminant function analyses are presented in Tables 4 and 5. The canonical variate anxiety score contributed significantly to both discriminant function equations. In both analyses a significant contribution to the equation was added when additional variables were added. In the CAQ analysis, Socialization contributed significantly to the discriminant function. In the AU1 analysis, Alcoholic Deterioration-l contributed significantly to the
140 TABLE
3
Means, standard deviations, and results of analyses of variance for the first canonical variate and second-order scales of the Alcohol Use Inventory and Clinical Analysis Questionnaire for three A.A. attendance profiles Variables
First canonical
variate
AU1 secondorder scales AU1 A (Selfenhancing drinking) AU1 B (Obsessivesustained drinking) AU1 C (Anxiety) AU1 Dl (Alcoholic deterioration-l) AU1 D2 (Alcoholic deterioration-2) CAQ second-order scales &I (Extroversion) Qu (Anxiety) Qm (Tough poise) QIv (Independence) &VI (Superego) &VII (Socialization) QvIrr (Depression) Qrx (Psychoticism) Qx (Neuroticism)
F
Non Mean
High
S.D.
Mean
S.D.
Mean
S.D.
11.89
8.83
14.62
6.09
19.58
5.81
4.99**
7.22
3.68
8.32
3.01
8.38
3.38
0.91
7.70
5.18
10.55
5.77
10.38
4.59
2.15
11.26 11.87
7.34 6.27
13.63 16.63
5.10 6.91
16.54 21.62
4.25 8.97
3.53* 8.07**
6.13
3.12
7.63
4.59
8.38
3.20
1.62
4.13 6.87 6.59 4.77 6.24 6.60 7.78 7.29 5.75
1.56 1.83 1.61 1.75 1.81 1.34 0.77 0.69 2.21
4.30 6.67 5.74 5.47 5.25 5.63 7.63 7.17 5.56
1.42 1.53 1.76 1.66 1.89 1.50 0.72 0.65 1.74
3.98 7.85 5.14 4.47 4.78 6.66 7.90 6.99 6.53
1.12 1.54 1.44 1.41 1.63 1.56 0.90 0.61 1.84
0.27 2.58 3.52* 2.37 3.28* 4.28* 0.67 0.88 1.24
LOW
Note: d.f. = 2,71. *p < 0.05; **p < 0.01.
discriminant function. Over-all these data indicate that regular A.A. attenders tend to be more anxious and display a more deteriorated drinking pattern than low or non-attenders. Low affiliates are more anxious than nonaffiliates. Low affiliates also demonstrated lower socialization than both high and non-attending groups. Discussion The results of the discriminant function analyses performed on the drinking and personality variable sets identified those variables differentiating the A.A. affiliation groups. The second-order Alcoholic Deterioration Scale (Dl), characterized by loss of behavioral control when drinking, social-role maladaptation, psychoperceptual and psychophysical withdrawal, was one factor contributing significantly to the discriminant function equation. The
TABLE 4
First canonical variate AU1 Dl (Alcoholic deterioration-l) AU1 C (Anxiety) AU1 A (Self-enhancing drinking) AU1 D2 (Alcoholic deterioration-z) AU1 B (Obsessivesuatained drinking)
1 2 3 4 5 6
0.88 0.80 0.76 0.74 0.73 0.72
Wilks’ lambda
0.009 0.003 0.004 0.008 0.019 0.037
Significance of discriminant function equation 9.99 17.62 21.26 23.35 24.47 25.54
Raos V
9.99 7.64 3.64 2.09 1.11 1.08
0.007 0.022 0.162 0.351 0.574 0.584
Significance of adding variable to equation
Use Inventory for three
Change in Raos V
scales of the Alcohol
2 3 4 5 6 7 8 9 10
1
Step number
&VI @we-w)
Qx (Neuroticism0 Qr (Extroversion) QIx (Psychoticism)
&II (Anxiety)
0.88
First canonical variate Qvn (Socialization) Qvnl (Depression) Qm (Tough poise) Qvl (Independence) 0.76 0.71 0.69 0.66 0.65 0.62 0.61 0.60 0.59
Wilks’ lambda
Variables entered
0 .OOl 0.001 0.001 0.002 0.003 0.004 0.006 0.011 0.020
0.009
Significance of discriminant function equation
9.99
11.47 4.77 2.82 3.87 1.89 3.31 2.28 1.50 1.32
9.99
Change in Raos V
21.45 26.23 29.04 32.91 34.81 38.11 40.38 41.89 43.20
Raos V
0.007 0.003 0.092 0.244 0.145 0.388 0.191 0.320 0.474 0.518
Significance of adding variable to equation
Summary of discriminant function analysis on first canonical variate, and second-order scales of the Clinical Analysis Questionnaire for three A.A. attendance profiles --.
TABLE 5
Variables entered
Step number
Summary of discrhninant function analysis on first canonical variate, and secondorder A.A. attendance profiles
2 +
142
CAQ second-order factor of Socialization (QvII) also contributed significantly. Primary s:ales loading high on Socialization include threat-sensitive behavior, group adherence, and degree of submissiveness, unpretentiousness, and conservativism. Two coefficients for the first canonical variate CAQ second-order factor Qn (Anxiety) and AU1 secondorder factor C (Anxiety) also helped to distinguish the affiliation groups. The methodological procedures employed suggest that measurement of differential levels of anxiety overlap in the two test instruments; but in relation to the other variables found to be significant measures of personality and drinking characteristics, Lhe two Lests remain independent of one another. Substantial constm:t validity of the second-order factors from the CAQ [ 201 and AU1 1171 lend further tiupport to the contentions of independent measurement taken from these two variable sets. Symptom patterns manifested by the high affiliate group in this study remain consistent with past findings [ 2 - 4,12,22,2?] . The interacting characteristics of frequent attenders to A.A. illustrates their significantly greater degree of anxiety, which may be differentially associated with drinkingrelated behaviors. The higher level of guilt-proneness, evidenced by the high affiliation group, also distinguishes them from the low and non-attending groups. The history of “hitting bottom” [ 8; and initiating surrender to the disease of alcoholism proposed by AA. is d central role demand A.A. places on its members. The opportunity AL!. gives to express these transgressions requires members to hale an ability to stare guilt-related emotions spontaneously [ 21. Essential elements necessary I’or meethg these role demands certainly include individuals who are group orientated and who are more formistic in their thinking [ 111 . Profiles established from the AU1 indicate that members of the high affiliation grow p are more likely to exhibit greater disruption in the physical, psychological, and social areas of functioning as a direct or indirect result of their alcohol abuse p&em. Their drinking style was characterized by a greater loss of behavioral cantrtil with associated social-role maladaptations. Similarly, this group experienced signil’icantly greater degrees of psychoperceptual and psychophysical withchawal symptoms. In contrast, the non and low affiliation grcups did not exhibit these broad alcohol-related disruptions although Wanberg et 51. [ 171 are careful to note that 2ven a low to moderate score on this factor (Alcoholic Deterioration-l) indicates a noteworthy alcohol-related problem. Affective vulnerability and emotional disruption resulting from the use of alcohol is also significantly less in the non affiliation group, which perhaps characterizes best the differences in drinking chronicity among the three ;ubt.ypes. Given these combinations of psychological and drinking-related characteristics, appropriate therapeutic interventions for high affiliation patients may be long-term residential or hospital care, with an emphasis on developing insight towards th&Jir anxiety and depression. Communications training would be beneficial in helping these patients deal with stressful situations more appropriately. Individuals in the non and low affiliation
143
groups may benefit from less-structured psychotherapeutic techniques offered on an outpatient basis. Further research on affiliation subtypes which sought to identify interrelationships of drinking pattern, personality traits, and treatment outcome would seem warranted.
References 1 F. Baekeland, Evaluation of treatment methods in chronic alcoholism. In B. Kissin and H. Beglciter (eds.), Biology of Alcoholism, Vol. 5, Plenum Press, New York, 1977. 2 H. M. Trite, A study of the process of affiliation with Alcoholics Anonymous. Q. J. Stud. Alcohol, 18 (1957) 39 - 54. 3 H. M. Trite, The affiliation motive and readiness to join Alcoholics Anonymous. Q. J. Stud. Alcohol, 20 (1959) 313 - 320. 4 H. M. Trite and P. M. Roman, Sociopsychological predictors of affiliation with Alcoholics Anonymous. Sot. Psychiatry, 5 (19 10) 51 - 59. 5 M. Bean, Alcoholics Anonymous, Chapber VI: Problem. Psychiatr. Ann., 5 (1975) 92 -97. 6 W. Eckhardt, Alcoholic values and Alcoholics Anonymous. Q. J. Stud. Alcohol, 28 (1967) 277 - 287: 7 H. T. Blane, The Personality of the Alcoholic. Harper and Row, New York, 1968. 8 E. M. Blum and R. H. Blum, Alcoholism. Jossey-Bass, San Francisco, 1972. 9 C. E. Thu,le, Alcoholism and the archetypal past: A phenomenulogical perspective on Alcoholics Anonymous. J. Stud. Alcohol, 3e (1977) 75 - 88. 10 C. Hensman, A. Hawker and V. Williamson, Who goes to .\lcoholics Anonymous? Lancet, ii (1966) 382. 11 A. F. Fontana, B. N. Dowds and M. A. Bethel, A. A. and group therapy for alcoholics: An application of the World Hypothesis Scale. J. Stud. Alcohol, 37 (1976) 675 - 682. 12 D. M. Donovan, E. F. Chaney and M. R. O’Leary, Alcoholic MMPI subtypes: relationship to drinking styles, benefits, and consequences. J. Nerv. Ment. Dis., 166 (1978) 553 - 561. 13 P. R. Whitelock, J. E. Overall and J. H. Patrick, Personality patterns and alcohol abuse in a state hospital population. J. Abnorm. l’sychol., 78 (1971) 9 - 16. 14 S. G. Goldstein and J. D. Linden, Multivariate classification of alcoholics by means of the MMPI. J. Abnorm. Psychol., 7d (1969) 661- 669. 15 G. F. Lawlis and S. E. Rubin, 16P.F. study of personality patterns in z$coholics. Q. J. Stud. Alcohol, 32 (1971) 318 - 327. 16 R. M. Costello, G. F. Lawliss, K. R. Manders and J. F. Celistino, Empirical derivation of a partial personality typdlogj of alcoholics. J. Stud. Alcohol, 39 (1978) 1258 1266. 17 K. W. Wanberg, J. L. Horn and F. M. FQster, A differential assessment model for alcoholism: the scales of the Alcohol Use Invent,ory. J. Stud. Alcohol, 38 (1977) 512543. 18 J. L. Horn and K. W. Wanberg, Symptom patterns related to excessive use of alcohol. Q. J. Stud. Alcohol, 30 (lY69) 35 - 58. 19 K. H. Delhees and R. B. Cattell, Manual for the Clinical Analysis Questionnaire (CAQ). Institute for Personality and Ability Testing, Champaign, IL, 1975. 20 S. E. Krug and J. E. Laughlin, Second+rder factors among normal and pathological primary personality traits. J. Consult. Clin. Psychol., 45 (1977) 575 - 582. 21 N. H. Nie, C. H. Hull, J. G. Jenkins, K. Steinbrenner and D. H. Bent, Statistical Package for the Social Services. McGraw-Hill, New York, 1975.
144 22 C. D. Emrick, C. L. Lassen and M. T. Edwards, Non-professional peers as therapeutic agents. In: A. S. Gorman and A. M. Razin (Eds.), Effective Psychotherapy: A Handbook of Research. Pergamon Press, Elmsford, New York, 1977. 23 R. D. Bonacker, Alcoholism and the Alcoholics Anonymous viewed symptomatologically. Ment. Hyg., 42 (1958) 562 - 566.
135
DIFFERENTIAL ALCOHOL USE PATTERNS AND PERSONALITY TRAITS AMONG THREE ALCOHOLICS ANONYMOUS ATTENDANCE LEVEL GROUPS: FURTHER CONSIDERATIONS OF THE AFFILIATION PROFILE* MICHAEL R. O’LEARY Department of Psychiatry and Behavioral Sciences, University of Washington, Medicine, and Veterans Administration Medical Center, Seattle, Washington
School of (U.S.A.)
DONALD A. CALSYN Department of Educational Psychology, Administration Medical Center, Seattle,
University Washington
of Washington, (U.S.A.)
and Veterans
DENNIS L. HADDOCK Veterans
Administration
Medical
Center,
Seattle,
Washington
(U.S.A.)
CHARLES W. FREEMAN Pain Clinic,
Swedish
Hospital
Medical
Center,
Seattle,
Washington
(U.S.A.)
(Received March 5,197Q)
Summary The present study was designed to identify personality and drinking behavior correlates associated with membership in Alcoholics Anonymous (A.A.). Affiliation profiles were developed on the basis of frequency of A.A. attendance; subtypes were defined as non, low, and high affiliates, respectively. An initial canonical correlation analysis assessed the amount of common variance between personality and drinking characteristics. Subsequent discriminant function analyses on the second-order scales of the drinking and personality variables sets identified those variables differentiating the A.A. affiliation groups. High levels of affiliation were associated with a greater degree of anxiety, a tendency to be more affected by feelings rather than intellect, and more deterioration in the physical, psychological, and social areas of functioning. Regular attenders of A.A. also exhibited significant differences in perceived benefits, style, and consequences of alcohol usage. The potential therapeutic utility of A.A. affiliation profiles was discussed.
--*Address reprint requests to Michael R. O’Leary Ph. D. (llGATP), Veterans Administration Medical Center, 4435 Beacon Avenue South, Seattle, Washington 98108, U.S.A. This work was supported by the Veterans Administration Health Services Research and Development Service.
136
An explicit and integral part of many inpatient and many outpatient alcoholism treatment programs is Alcoholics Anonymous (A.A.) [ 11. This self-help organization, established in 1935, claims one of the highest rates of treatment success [ 21. However, surprisingly little is known about the interacting variables of drinking pattern and personality structure which would predict successful affiliation with A.A. Early research has identified some of the psychological characteristics of Alcoholics Anonymous members in an effort to maximize the therapeutic benefits of this treatment modality [2 41 . Personal and sociologic characteristics such as mental status, religion, social class, sex, ethnicity, age, and personality traits are not randomly distributed in A.A. [ 51. Other investigators suggest that A.A. affiliates differ significantly from non-affiliates in the degree to which they exhibit emotional dependency, proneness to guilt, high egocentricity, denial, high anxiety, and general alcoholic deterioration [4,6 - lo]. Recent attempts to assess structural conflicts experienced by some alcoholic patients concurrently engaged in group therapy and A.A. further point to the need of classifiying A.A. members in terms of known psychodiagnostic entities. Fontana et al. [ 111 suggest that substantial conflict arises out of the patient’s preferred ways of understanding the world. They found that individuals orientated toward understanding the world in terms of definite categories and typological labels (formistic thinking) found A.A. to be more appealing than group theory. Although this approach to subtype identification of A.A. affiliates is meaningful, it remains a controversial area. In an attempt to replicate previously found alcoholic MMPI personality subtypes, Donovan et al. [12] stated that “psychologically meaningful personality subtypes could be identified within alcoholic or alcohol-abusing populations, although investigators differ as to the number and appropriate composition of such subtypes”. Similarly, others have noted the difficulties encountered in identifying relationships between personality characteristics and drinking measures [131. A number of studies have attempted to generate alcoholism typologies from the MMPI [ 141 the 16 Personality Factor Questionnaire (16-PF) [15], and the Clinical Analysis Questionnaire [ 161 . These findings lend support to the contention of Wanberg et al. [ 171 that the unitary trait model of alcoholism is of limited heuristic value. Wanberg et al. suggest that alcohol abuse is a multiple syndrome disorder exhibiting a great degree of variability within and across subjects with respect to symptom nature and manifestation. Similarly, a multi-dimensional approach to the explanation of alcoholism provides a greater amount of information concerning inter- and intrasubject differences in the perceived benefits which motivate drinking, the nature of drinking patterns, and the specific consequences resulting from alcohol abuse [ 121. The primary purpose of the present study was to determine the interactive effects of personality, psychopathology and drinking pattern correlates with successful affiliation with A.A. It was thought that identification
137
of those variables which discriminated between patients’ level of affiliation with A.A. would provide treatment personnel with a more informed basis for treatment planning. Specifically the following questions were addressed: (1) What is the relationship between drinking style, personality, and psychopathology? (2) If the common variance between these variables is accounted for, do they distinguish between groups on the basis of their A.A. attendance? Method Subjects. The sample consisted of 76 male veterans participating in an inpatient alcoholism rehabilitation program at the Seattle Veterans Administration Medical Center. None of the subjects exhibited physical or cognitive residuals of acute alcohol intoxication at the time of testing. The sample had a mean age of 42.9 years (S.D. = 12.5) and a mean educational level of 12.5 years (S.D. = 2.2). The mean socioeconomic status of the sample wassowermiddle class. The reported age of first drink was approximately 15 (X = 15; S.D. = 4.8); problematic drinking began at age 29 (_%= 29.4; S.D. = 11.3). Fifty-eight percent of the subjects had previously been involved in alcoholism treatment, with the average number of treatments for the entire sample equalling 2.5 (SD. = 3.7; range = 0 - 20). None of the subjects had primary diagnoses of organic brain syndrome or psychiatric disorders. All subjects provided informed consent to participate in the present research after the nature of the project was fully explained. Materials. The subjects were administered the Alcohol Use Inventory (AUI) [ 171, to assess the level of drinking chronicity. The AUI, which has evolved from continued psychometric refinement of the Drinking History Questionnaire [ 181, consists of factor analytically derived scales (16 firstorder, 5 second-order, and 1 third-order) that assess symptoms related to drinking, drinking-related behavior, and perceived benefits derived from drinking. The scales of the AU1 have demonstrated high levels of internal consistency and test-retest reliability based on a normative group of over 2000 subjects. Similarly, the AU1 has demonstrated a high degree of construct and predictive validity with respect to treatment outcome and clinical judgments concerning a variety of drinking-related variables. Personality trait assessment was made by administering the Clinical Analysis Questionnaire (CAQ) [ 191 to the present sample. The CAQ is a psychodiagnostic questionnaire that includes the 16 primary personality source traits of the 16-PF, and a 12-factor psychopathology supplement. Serving as a clinical diagnostic instrument, the CAQ allows assessment of psychopathology dimensions that are measured separately from normal personality structures. From both the AU1 and CAQ second-order factors [ 201 are obtained which are linear combinations of the primary scales. In an effort to increase variable stability, all analyses were carried out on second-order factors.
138
Procedure. The subjects were administered the CAQ and the AU1 along with other self-report inventories. Social and demographic status of subjects prior to admission was recorded and included educational, occupational, and familial items, as well as drinking history, and adjustment prior to hospitalization. Individual affiliation profiles were constructed for each subject on the basis of their response to statements describing past experience with A.A. Subjects acknowledging some experience with A.A. were regarded as affiliates, and those patients who had no prior experience were characterized as non-affiliates. Further, subjects having attended a few times (less than once every 6 months), or only sporadically in the past (once or twice every 3 months), were regarded as low affiliates. Finally, high affiliates were designated as those patients who had attended regularly most of the time (at least twice a month over the past year) or who demonstrated this pattern in years past but not recently. The three A.A. attendance groups did not differ significantly on any of the following demographic variables: age (F = 2.74; d.f. = 2, 76), education (F = 1.82, d.f. = 2,76), socioeconomic status (F = 1.33, d.f. = 2, 74), estimated WAIS Full Scale IQ from the Shipley Institute of Living Scale (F = 0.11, d.f. = 2,75), marital status (x2 = 1.85, d.f. = 2), employment status (x2 = 0.96, d.f. = 4), court referral (x2 = 1.85, d.f. = 2), age of first drink (F = 0.78, d.f. = 2,76), and age patient felt he became alcoholic (F = 2.6, d.f. = 2,73). Non A.A. affiliates were less likely to have had prior alcoholism treatment (x2 = 14.8, d.f. = 2, p < 0.001).
Results The analysis was designed to answer the two questions put forth in the introduction. A canonical correlation analysis [ 211, using the secondorder factors from the CAQ and AU1 as variable sets,was used to address the question of relationship between drinking characteristics, personality and psychopathology. The summary table from the analysis is presented in Table 1. Only the first canonical variate is significant, yielding a canonical correlation of 0.74. Listed in Table 2 are the coefficients for the first canonical variate. High loadings on factor Qn from the CAQ and Factor C from the AU1 suggest that the first canonical variate might best be thought of as measuring anxiety. These findings suggest that the CAQ and AU1 overlap in measuring anxiety, which accounts for about 55% of the variance between the two instruments. Outside of anxiety, however, the two measures do not appear to overlap at a significant level. A two-step process was used to answer the question of differential drinking and personality characteristics among the three patient groups when the common variance between the CAQ and AU1 is accounted for. The first canonical variate score for each subject was calculated using the coefficients listed in Table 2. Two separate stepdown discriminant function analyses
139 TABLE 1 Summary table of cananical correlation for Clinical Analysis Questionnaire with Alcohol Use Inventory secondorder factors Canonical variates
Canonical correlation
First Second Third Fourth Fifth
0.74 0.52 0.41 0.23 0.20
sz
-.. ---__
-_-.-.-
95.41 41 .oo 19.38 6.68 2.67
d.f.
P
45 32 21 12 5
0.000 0.132 0.561 0.878 0.750
TABLE 2 Coefficients for the first canonical variate from the secondorder Analysis Questionnaire and Alcohol Use Inventory Variables CAQ secondorder scales Qr (Extroversion)
&II (Anxiety)
Qm (Tough poise) Qrv (Independence) &VI (Superego) &vu (Socialization) Qvm (Depression) QIx (Psychoticism) Qx (Neuroticism) AU1 secondorder scales AU1 A (Self enhancing drinking) AU1 B (Obsessivesustained drinking) AU1 C (Anxiety) AU1 Dl (Alcoholic deterioration-l) AU1 D2 (Alcoholic deterioration-2)
scales of the Clinical
Coefficients
-0.28 0.99 m-O.32 0.01 -0.35 0.33 0.03 0.06 m-O.52 0.08 0.16 0.88 0.17 0.00
were then performed [ 211. In both analyses A.A. affiliation level served as the dependent classification grouping. The first canonical variate score was forced first into each discriminate function analysis along with AU1 and CAQ second-order factors respectively. Listed in Table 3 are the means, standard deviations and univariate F values for all of the variables in the two analyses. The summary tables from the two discriminant function analyses are presented in Tables 4 and 5. The canonical variate anxiety score contributed significantly to both discriminant function equations. In both analyses a significant contribution to the equation was added when additional variables were added. In the CAQ analysis, Socialization contributed significantly to the discriminant function. In the AU1 analysis, Alcoholic Deterioration-l contributed significantly to the
140 TABLE
3
Means, standard deviations, and results of analyses of variance for the first canonical variate and second-order scales of the Alcohol Use Inventory and Clinical Analysis Questionnaire for three A.A. attendance profiles Variables
First canonical
variate
AU1 secondorder scales AU1 A (Selfenhancing drinking) AU1 B (Obsessivesustained drinking) AU1 C (Anxiety) AU1 Dl (Alcoholic deterioration-l) AU1 D2 (Alcoholic deterioration-2) CAQ second-order scales &I (Extroversion) Qu (Anxiety) Qm (Tough poise) QIv (Independence) &VI (Superego) &VII (Socialization) QvIrr (Depression) Qrx (Psychoticism) Qx (Neuroticism)
F
Non Mean
High
S.D.
Mean
S.D.
Mean
S.D.
11.89
8.83
14.62
6.09
19.58
5.81
4.99**
7.22
3.68
8.32
3.01
8.38
3.38
0.91
7.70
5.18
10.55
5.77
10.38
4.59
2.15
11.26 11.87
7.34 6.27
13.63 16.63
5.10 6.91
16.54 21.62
4.25 8.97
3.53* 8.07**
6.13
3.12
7.63
4.59
8.38
3.20
1.62
4.13 6.87 6.59 4.77 6.24 6.60 7.78 7.29 5.75
1.56 1.83 1.61 1.75 1.81 1.34 0.77 0.69 2.21
4.30 6.67 5.74 5.47 5.25 5.63 7.63 7.17 5.56
1.42 1.53 1.76 1.66 1.89 1.50 0.72 0.65 1.74
3.98 7.85 5.14 4.47 4.78 6.66 7.90 6.99 6.53
1.12 1.54 1.44 1.41 1.63 1.56 0.90 0.61 1.84
0.27 2.58 3.52* 2.37 3.28* 4.28* 0.67 0.88 1.24
LOW
Note: d.f. = 2,71. *p < 0.05; **p < 0.01.
discriminant function. Over-all these data indicate that regular A.A. attenders tend to be more anxious and display a more deteriorated drinking pattern than low or non-attenders. Low affiliates are more anxious than nonaffiliates. Low affiliates also demonstrated lower socialization than both high and non-attending groups. Discussion The results of the discriminant function analyses performed on the drinking and personality variable sets identified those variables differentiating the A.A. affiliation groups. The second-order Alcoholic Deterioration Scale (Dl), characterized by loss of behavioral control when drinking, social-role maladaptation, psychoperceptual and psychophysical withdrawal, was one factor contributing significantly to the discriminant function equation. The
TABLE 4
First canonical variate AU1 Dl (Alcoholic deterioration-l) AU1 C (Anxiety) AU1 A (Self-enhancing drinking) AU1 D2 (Alcoholic deterioration-z) AU1 B (Obsessivesuatained drinking)
1 2 3 4 5 6
0.88 0.80 0.76 0.74 0.73 0.72
Wilks’ lambda
0.009 0.003 0.004 0.008 0.019 0.037
Significance of discriminant function equation 9.99 17.62 21.26 23.35 24.47 25.54
Raos V
9.99 7.64 3.64 2.09 1.11 1.08
0.007 0.022 0.162 0.351 0.574 0.584
Significance of adding variable to equation
Use Inventory for three
Change in Raos V
scales of the Alcohol
2 3 4 5 6 7 8 9 10
1
Step number
&VI @we-w)
Qx (Neuroticism0 Qr (Extroversion) QIx (Psychoticism)
&II (Anxiety)
0.88
First canonical variate Qvn (Socialization) Qvnl (Depression) Qm (Tough poise) Qvl (Independence) 0.76 0.71 0.69 0.66 0.65 0.62 0.61 0.60 0.59
Wilks’ lambda
Variables entered
0 .OOl 0.001 0.001 0.002 0.003 0.004 0.006 0.011 0.020
0.009
Significance of discriminant function equation
9.99
11.47 4.77 2.82 3.87 1.89 3.31 2.28 1.50 1.32
9.99
Change in Raos V
21.45 26.23 29.04 32.91 34.81 38.11 40.38 41.89 43.20
Raos V
0.007 0.003 0.092 0.244 0.145 0.388 0.191 0.320 0.474 0.518
Significance of adding variable to equation
Summary of discriminant function analysis on first canonical variate, and second-order scales of the Clinical Analysis Questionnaire for three A.A. attendance profiles --.
TABLE 5
Variables entered
Step number
Summary of discrhninant function analysis on first canonical variate, and secondorder A.A. attendance profiles
2 +
142
CAQ second-order factor of Socialization (QvII) also contributed significantly. Primary s:ales loading high on Socialization include threat-sensitive behavior, group adherence, and degree of submissiveness, unpretentiousness, and conservativism. Two coefficients for the first canonical variate CAQ second-order factor Qn (Anxiety) and AU1 secondorder factor C (Anxiety) also helped to distinguish the affiliation groups. The methodological procedures employed suggest that measurement of differential levels of anxiety overlap in the two test instruments; but in relation to the other variables found to be significant measures of personality and drinking characteristics, Lhe two Lests remain independent of one another. Substantial constm:t validity of the second-order factors from the CAQ [ 201 and AU1 1171 lend further tiupport to the contentions of independent measurement taken from these two variable sets. Symptom patterns manifested by the high affiliate group in this study remain consistent with past findings [ 2 - 4,12,22,2?] . The interacting characteristics of frequent attenders to A.A. illustrates their significantly greater degree of anxiety, which may be differentially associated with drinkingrelated behaviors. The higher level of guilt-proneness, evidenced by the high affiliation group, also distinguishes them from the low and non-attending groups. The history of “hitting bottom” [ 8; and initiating surrender to the disease of alcoholism proposed by AA. is d central role demand A.A. places on its members. The opportunity AL!. gives to express these transgressions requires members to hale an ability to stare guilt-related emotions spontaneously [ 21. Essential elements necessary I’or meethg these role demands certainly include individuals who are group orientated and who are more formistic in their thinking [ 111 . Profiles established from the AU1 indicate that members of the high affiliation grow p are more likely to exhibit greater disruption in the physical, psychological, and social areas of functioning as a direct or indirect result of their alcohol abuse p&em. Their drinking style was characterized by a greater loss of behavioral cantrtil with associated social-role maladaptations. Similarly, this group experienced signil’icantly greater degrees of psychoperceptual and psychophysical withchawal symptoms. In contrast, the non and low affiliation grcups did not exhibit these broad alcohol-related disruptions although Wanberg et 51. [ 171 are careful to note that 2ven a low to moderate score on this factor (Alcoholic Deterioration-l) indicates a noteworthy alcohol-related problem. Affective vulnerability and emotional disruption resulting from the use of alcohol is also significantly less in the non affiliation group, which perhaps characterizes best the differences in drinking chronicity among the three ;ubt.ypes. Given these combinations of psychological and drinking-related characteristics, appropriate therapeutic interventions for high affiliation patients may be long-term residential or hospital care, with an emphasis on developing insight towards th&Jir anxiety and depression. Communications training would be beneficial in helping these patients deal with stressful situations more appropriately. Individuals in the non and low affiliation
143
groups may benefit from less-structured psychotherapeutic techniques offered on an outpatient basis. Further research on affiliation subtypes which sought to identify interrelationships of drinking pattern, personality traits, and treatment outcome would seem warranted.
References 1 F. Baekeland, Evaluation of treatment methods in chronic alcoholism. In B. Kissin and H. Beglciter (eds.), Biology of Alcoholism, Vol. 5, Plenum Press, New York, 1977. 2 H. M. Trite, A study of the process of affiliation with Alcoholics Anonymous. Q. J. Stud. Alcohol, 18 (1957) 39 - 54. 3 H. M. Trite, The affiliation motive and readiness to join Alcoholics Anonymous. Q. J. Stud. Alcohol, 20 (1959) 313 - 320. 4 H. M. Trite and P. M. Roman, Sociopsychological predictors of affiliation with Alcoholics Anonymous. Sot. Psychiatry, 5 (19 10) 51 - 59. 5 M. Bean, Alcoholics Anonymous, Chapber VI: Problem. Psychiatr. Ann., 5 (1975) 92 -97. 6 W. Eckhardt, Alcoholic values and Alcoholics Anonymous. Q. J. Stud. Alcohol, 28 (1967) 277 - 287: 7 H. T. Blane, The Personality of the Alcoholic. Harper and Row, New York, 1968. 8 E. M. Blum and R. H. Blum, Alcoholism. Jossey-Bass, San Francisco, 1972. 9 C. E. Thu,le, Alcoholism and the archetypal past: A phenomenulogical perspective on Alcoholics Anonymous. J. Stud. Alcohol, 3e (1977) 75 - 88. 10 C. Hensman, A. Hawker and V. Williamson, Who goes to .\lcoholics Anonymous? Lancet, ii (1966) 382. 11 A. F. Fontana, B. N. Dowds and M. A. Bethel, A. A. and group therapy for alcoholics: An application of the World Hypothesis Scale. J. Stud. Alcohol, 37 (1976) 675 - 682. 12 D. M. Donovan, E. F. Chaney and M. R. O’Leary, Alcoholic MMPI subtypes: relationship to drinking styles, benefits, and consequences. J. Nerv. Ment. Dis., 166 (1978) 553 - 561. 13 P. R. Whitelock, J. E. Overall and J. H. Patrick, Personality patterns and alcohol abuse in a state hospital population. J. Abnorm. l’sychol., 78 (1971) 9 - 16. 14 S. G. Goldstein and J. D. Linden, Multivariate classification of alcoholics by means of the MMPI. J. Abnorm. Psychol., 7d (1969) 661- 669. 15 G. F. Lawlis and S. E. Rubin, 16P.F. study of personality patterns in z$coholics. Q. J. Stud. Alcohol, 32 (1971) 318 - 327. 16 R. M. Costello, G. F. Lawliss, K. R. Manders and J. F. Celistino, Empirical derivation of a partial personality typdlogj of alcoholics. J. Stud. Alcohol, 39 (1978) 1258 1266. 17 K. W. Wanberg, J. L. Horn and F. M. FQster, A differential assessment model for alcoholism: the scales of the Alcohol Use Invent,ory. J. Stud. Alcohol, 38 (1977) 512543. 18 J. L. Horn and K. W. Wanberg, Symptom patterns related to excessive use of alcohol. Q. J. Stud. Alcohol, 30 (lY69) 35 - 58. 19 K. H. Delhees and R. B. Cattell, Manual for the Clinical Analysis Questionnaire (CAQ). Institute for Personality and Ability Testing, Champaign, IL, 1975. 20 S. E. Krug and J. E. Laughlin, Second+rder factors among normal and pathological primary personality traits. J. Consult. Clin. Psychol., 45 (1977) 575 - 582. 21 N. H. Nie, C. H. Hull, J. G. Jenkins, K. Steinbrenner and D. H. Bent, Statistical Package for the Social Services. McGraw-Hill, New York, 1975.
144 22 C. D. Emrick, C. L. Lassen and M. T. Edwards, Non-professional peers as therapeutic agents. In: A. S. Gorman and A. M. Razin (Eds.), Effective Psychotherapy: A Handbook of Research. Pergamon Press, Elmsford, New York, 1977. 23 R. D. Bonacker, Alcoholism and the Alcoholics Anonymous viewed symptomatologically. Ment. Hyg., 42 (1958) 562 - 566.