Evaluating the effects of self-esteem on substance abuse among homeless men.

Abstract

Associations between self-esteem and abuse of alcohol and psychoactive substances have been documented in empirical studies involving high school and college students. No research exists that addresses whether this association generalizes to adult homeless substance users. The current study uses secondary data analysis methodology to evaluate an experimental design study involving 305 homeless men, assigned randomly to the treatment or control group. Control subjects were referred to community-based services. Experimental subjects were exposed to individual and group interventions, life-skills, and relapse prevention training while residing in a 24-hour shelter, for three months. Trained graduate students collected data using standardized questions to interview subjects. Three hypotheses were tested. Hypothesis I that the interventions would contribute toward increased self-esteem at T2, T3, T4 and T5 was not supported. The preponderance of findings pertaining to Hypothesis II, that higher self esteem would be associated with lower alcohol and drug use in treatment subjects, and Hypothesis III, that these associations would be greater among treatment than control subjects over time, were not confirmed, although a few results were consistent with these hypotheses. Overall, results indicated that self-esteem was not increased in treatment subjects despite decreases in alcohol and drug use. The role of serf-esteem in this population appears different from its importance in high school and college students. Possible reasons for this apparent difference are explored.

Key words: substance abuse; self-esteem; homeless males

INTRODUCTION

Self-esteem has long been believed to play an important role in the use of alcohol and psychoactive substances (Charalampous, Ford, and Skinner; 1976; Donnelly, 2000). Several researchers have argued that low self-esteem poses high risk for substance abuse in some populations, including adolescents, college students (Mitic, 1980; Yanish, and Battle 1985), young females (Beckman, 1978; Engs, and Hanson 1989) and African Americans (Grills, and Longshore 1996). Results of these studies have led researchers to promote the theory that if levels of self-esteem can be determined, it may be possible to predict, change, or improve the lives of some people (Gross 1970; Jessor, and Jessor 1977; Laflin et al., 1994). Yet no corresponding studies could be found that had rigorously investigated the relationship between levels of self-esteem and substance abuse among the homeless. Most of what is "known about the relationship between self-esteem and substance abuse is based on studies that involved alcohol use by students in high school or college, thereby severely limiting the generalizability of findings to more mature populations, such as the chronically homeless substance-abusing men living in urban settings (Segal, Rhenberg, and Sterling 1975). Greater knowledge about inverse relationships between self-esteem and substance use disorder in more mature and older adults is needed so that budgets are not wasted on implementing costly, ineffective interventions that have little or no proven long-lasting results (Watson, 1991).

Homeless substance abusing men are of special interest because studies have found that they consume twice as much alcohol as women. They also tend to have different reasons for abusing drugs and alcohol and exhibit different course and consequences of addiction than do women (Allen, 1969; Kaplan, 1996). In addition, Nielsen and Scarpitti (1997) found that man), homeless substance abusers had low self-esteem, low self-worth and were lacking in self-confidence. Research has also found that homeless individuals often suffer from lethal, abusive, or destructive personalities, self-defeating behaviors and underdeveloped personal skills (Break, 1987).

Homeless individuals pose special challenges for treatment providers because many suffer with disaffiliation, a mistrust of institutions, have multiple problems and tend to be highly mobile (Breaky, 1987). Leaf (1993) found also that single men comprise the majority in the homeless population and that significant numbers of them are poor, suffer disproportionately with chronic alcoholism, co-occurring drug use disorder, psychiatric illnesses, victimization and stigmatization. Research in self-esteem and its effects on substance use disorder has become an important area again because since 1997 American federal drug policy shifted its focus from supply reduction toward demand reduction. Consequently, greater attention and more resources will be directed toward prevention and treatment (ONDCP, 1998a). This shift in policy has also rekindled interest in therapeutic behavior model programs, social model and self-help programs, which were popular during the early 1990s. Many of these programs and interventions strongly emphasized the role play by factors such as self-esteem, self-efficacy and social support.

THEORETICAL BACKGROUND

"Substance abuse" refers to substance use disorder as defined by the fourth edition of the Diagnostic and Statistical Manual of Mental Disorders IV (American Psychiatric Association, 1994) and includes alcohol and illicit drugs. Self-esteem has been historically linked to theories about self and self-will, and has emerged as an important tool for understanding human behavior and for treating negative thoughts, inner feelings of incompleteness, emptiness, self-doubt and self hatred (Adler et al., 1992; Crocker et al., 1994). Self-esteem is viewed as the extent to which one's self-evaluations are favorable or unfavorable (Coppersmith, 1967). It is also referred to as "the cumulative product of socialization experiences" distributed across different social sectors and interpersonal associations (Kaplan, 1996).

It is also Kaplan's (1975) view that all individuals possess the "self-esteem motive," directing them toward minimizing negative self-attitudes and maximizing positive perceptions of the self. Low self-esteem and a lack of conformity were found to be high risk factors strongly correlated with the use of tobacco, alcohol and other drugs by adolescents and young adults (Ward, 2002). Several empirical studies found significant relationships between self-esteem and self-reported problem drinking (Beckman, 1978; Botvin et al., 1995b; Corbin, McNair, and Carter 1996; Glindemann, Scott, and Fortney, 1999; Maney, 1990; Parish, and Parish 1991; Schaeffer, Schuckit, and Morrissey 1976). Some studies reported contradictory and inconsistent relationships (Maney, 1990). For example, Mitic (1980) found that regular alcohol drinkers had greater scores for self-esteem compared with heavy drinkers and abstinent adolescents, and that although heavy drinking was associated with low self-esteem for females; the opposite was true for males. Corbin et al. (1996) found that a substantial number of alcoholics exhibited relatively high self-esteem, compared with non-alcoholics. Some studies also found that cocaine users in particular exhibit unusually high levels of self-esteem before an onset of drug abuse (Rickwood, and Braithwaite 1994; Shaffer, and Jones 1989). Other researchers have found individuals with high levels of self-esteem displayed lower levels of serious involvement with alcohol or illicit drugs and exhibited lower tendency to experiment with either alcohol or illicit drugs (Gorman, 1996; Schroeder, Laflin, and Weis, 1993). For example a study involving working-class males in the Metropolitan Los Angeles area found a moderate negative correlation between self-esteem (measured by the Rosenberg 10-item Scale) and a 6-item measure of drinking problems (Seeman, and Seeman, 1992).

Social influence and behavior modification programs were historically based on addictive behavior philosophy (Marlatt, and Gordon 1985), with the conceptual framework provided by social learning theory (Bandura, 1977). The core principle of addictive behavior philosophy has been that substance abuse is a "learned habit" which can also be "unlearned". Consequently, the goal of such interventions is to decrease participants' vulnerability to negative social influences by exposing them to skills training such as relapse prevention techniques, with the enhancement of self-esteem as a vital component (Watson, 1991). Behavior model treatment strongly emphasizes improvement of self-image and self-esteem.

In order to demonstrate the relationships between self-esteem and substance use disorder more evidence and reliable methods that demonstrate the lasting and cumulative effects are needed, especially from such challenging populations such as adult homeless male substance abusers and addicts living in urban settings. In the current study, answers were sought for two research questions: (a) Did the intervention contribute to higher self-esteem? (b) Were changes in self-esteem associated with changes in levels of alcohol use and drug use? Based on these questions the following hypotheses were formulated: (1) The experimental intervention would be associated with a higher level of self-esteem on the part of the treatment group compared with the control group at the 3-month (T2), 6-month (T3), 9-month (T4), and 15-month (T5) follow-up. (2) Higher self-esteem would be associated with lower alcohol use and drug use within the treatment group at T2, T3, T4, and T5. (3) Higher self-esteem in the experimental treatment group would be associated with lower alcohol use and drug use compared with such an association within the control group at T2, T3, T4, and T5.

METHOD

The study uses methodology for secondary analysis of clinical data originally collected by the Grant Street Partnership Project (GSP) to test its hypotheses. The original research was part of a three-year (1991-1993) research demonstration field trial sponsored by the National Institute of Alcohol Abuse and Alcoholism (NIAAA), in collaboration with Yale and Johns Hopkins University Schools of Medicine, The City of New Haven and the Hill Health Corporation (Leaf, 1993). Trained research assistants who were graduate students in public health, sociology, and anthropology collected the data between 1991 and 1993 using the "client interview questionnaire" (CIQ) (Leaf, 1993).

SUBJECTS

Subjects were adult males recruited by a specially trained outreach worker from shelters for the homeless, at soup kitchens, hospital emergency rooms, and the criminal justice system. Eligibility criteria were male gender, age 18 or older, with recent history or current problem with alcohol or drug use, and currently homeless in the New Haven area (Lain et al., 1995). Of the estimated 1,350 homeless individuals, 540 men were recruited of which 305 (N) met the criteria for admission. Over the three-year research period, 193 male subjects were randomly assigned to the experimental treatment group, and 112 to the control group. Subjects were medically detoxified for alcohol and heroin abuse. Subjects were each examined by a medical physician, a psychiatrist, and completed a battery of psychological tests.

Control group subjects received no additional services once randomization was complete. Those subjects in need of immediate medical services were routinely referred to appropriate levels of care such as emergency room, medical detoxification, medication, or counseling, and were routinely informed in writing about an array of community, based services they could access on their own. Subjects in both groups were randomly tested for alcohol use and drug use with breathalyzers and urinalysis respectively (Lam et al., 1995).

Demographic and social characteristics from the sample are presented in Table 1. African Americans comprised 59% of the baseline sample (n = 179), Whites 22% (n = 67), and Hispanics 19% (n = 59). Mean age was 32.5 years with a range of 19 to 63 years. The mean years of schooling were 11.6, with the range of 3 to 17 years. Sixty percent (n = 200) of subjects had completed high school. About 40% of the men had experienced prolonged homelessness during the past five years. All had been homeless during the past 60 days, and had spent a mean of 15 nights in a shelter. Sixty-five percent of the sample (n = 197) reported lifetime alcohol or drug abuse of ten years or more. Seventy-nine percent of the sample (n = 240) reported using alcohol during the previous month. Eighty-nine percent of subjects (n = 272) report cocaine use, 38% (n = 116) reported marijuana use and 30% (n = 92) reported heroin use during the previous month. Eighty-two percent (n = 250) reported combination drug use of mostly marijuana and another drug (Lam et al., 1995).

EXPERIMENTAL CONDITIONS

Program services were delivered exclusively to subjects in the experimental group while they resided at the Grant Street Partnership facility (GSP), which was operated as a 24-hour monitored rehabilitation facility for alcohol and drug dependency. Women were excluded from the study because of the lack of adequate sleeping and child-care facilities. Program offerings included congregate dining, opportunities for sports and recreation, and semi-private living and sleeping areas. Experienced and trained professionals in psychiatry, medicine, psychology, and drug and alcohol abuse and dependency delivered the clinical program. Treated subjects resided at the facility for up to three months and were offered on average 60 sessions of hour long individual therapy, 35 sessions of two-hour long group interaction, 75 sessions of hour long meetings with case managers, and 45 sessions of hour-long exposure to a "life skills" curriculum that included self-esteem building and relapse prevention techniques.

Sessions in self-esteem focused on several topics such as, "the sense of belonging", "learning", "contributing", "courage", "responsibility," and "cooperation". Other core elements included additional teaching and interactive learning sessions (20) and weekly attendance at Alcoholic Anonymous (AA) and Narcotics Anonymous (NA) meetings for two months. Overlapping services included training in conducting job searches, art and music therapy, strategizing and planning for subjects' post-treatment re-entry into the community (Lam et al. 1995; Leaf, 1993). Subjects who successfully completed the three-months experimental treatment phase became eligible for the "Aftercare Program", which provided limited case management for six months with other support services in obtaining and maintaining private housing, employment, daily-life skills, social support and training in relationship building skills.

MEASURES

This questionnaire contained factual closed-ended questions, and a battery of instruments required by the national (NIAAA) evaluation, such as the Rosenberg Self-Esteem Inventory, the Alcohol Severity Index (ASI), and the Quality of Life Interview (Lain et al. 1995; Orwin, 1995). The full version of the CIQ was administered for the baseline, and modified versions were administered at T2, T3, T4, and T5 follow-ups. Interviews lasted for about 2 1/2 hours, and subjects were paid an average of $25 for a completed interview.

INSTRUMENTS

Alcohol and Drug Use

Dependent variables were self-reported Alcohol Use and self-reported Drug Use, collected with the standardized Alcohol Severity Index. Questions asked by interviewers were: (1) "How many days in the past 30 did you drink alcohol?" and (2), "How many days in the past 30 did you use drugs?"

Self-Esteem

Self-reported Self-Esteem was both a dependent and independent variable in the present study. Self-Esteem was measured using the standardized Self-Esteem Scale (Rosenberg 1965). This instrument was embedded in a standardized client interview questionnaire (CIQ). The Self-Esteem Inventory (Rosenberg 1965) contains 10-item measures and a response set to four levels (1 represented strongly agree, 2 agree, 3 disagree, and 4 strongly disagree).

Respondents were asked to rate the following items: 1) "I feel that I'm a person of worth, at least on an equal plane with others". (2) "I feel that I have a number of good qualifies". (3) All in all, I am inclined to feel that I am a failure". (4) "I am able to do things as well as most other people". (5) "I feel that I do not have much to be proud of ". (6) "I take a positive attitude toward myself". (7) On the whole, I am satisfied with myself". (8) "I wish I could have more respect for myself". (9) I certainly feel useless at times". (10) "At times I think I am no good at all".

ANALYTICAL STRATEGY

The current investigator used Pearson Chi-Square for nonparametric variables, and the F-test in one-way analysis of variance (ANOVA) for continuous variables to determine comparability between groups at baseline (T1), and at T2, T3, T4, and T5. Mixed between-within subjects' multivariate analysis of variance (MANOVA) in repeated-measures was used for investigating the impact of the intervention on self-esteem of experimental subjects, and the difference between groups at T2, T3, T4 and T5. For this analysis, Self-Esteem was entered as the within-subject factor, and Group (treatment versus control) was entered as the fixed factor. Repeated-measures multivariate analysis of covariance (MANCOVA) was used to test Hypotheses II, that higher Self-Esteem would be associated with lower Alcohol Use and Drug Use at T2, T3, T4 and T5. Repeated-measures multivariate analysis of covariance (MANCOVA) was also used to test Hypothesis III, that the association between higher Self-Esteem and Lower Alcohol Use and Drug Use would be greater among experimental than among control subjects at T2, T3, T4, and T5. Between-subjects effects, which show effect by the independent variable on the dependent variable, and were assessed to show the unique contribution by Self-Esteem to differences in Alcohol Use and Drug Use by group (treatment versus control). For this analysis, Self-Esteem and Group (treatment versus control) were entered as the between-subjects or independent variables, and Alcohol Use and Drugs Use were both entered as the within-subjects or dependent measures.

The significance levels of statistical analyses were set at alpha .05 for the t and F values. Estimated size of effects produced for the repeated-measures analyses of variance and covariance is indicated by the eta square ([[eta].sup.2]) statistic. This statistic indicates the proportion of variance in the dependent variable accounted for by the independent variable (George, and Mallery 2001).

ATTRITION AND MISSING DATA

The number of subjects the treatment and control groups varied, due to the effects of administrative discharges of treatment group subjects, and according to the numbers of subjects located by interviewers at each of the four follow-up points. Thus, of 305 (N) interviewed at Baseline, 222 were located and interviewed at T2, 229 were interviewed at T3, 244 at T4, and 158 at T5. The study was designed as a controlled experimental field trial and researchers had planned equal distribution of subjects between the two groups. However, this did not occur because several subjects who were randomly assigned to the control condition refused to participate further upon realizing they would not receive any direct services, resulting in uneven sample sizes. The study also encountered missing data due to non-responses by subjects, skipped questions, subjects' inability to understand or respond, and premature attrition. Some attrition was attributed to treatment subjects leaving voluntarily, and to administrative discharges for violent behavior, and other serious program rule infractions. For purposes of this current evaluation, rather than dropping subjects (Cohen, and Cohen 1983), this investigator used "Linear-trend at point" procedures to replace missing data statistically (George, and Mallery, 2001).

RESULTS

Hypothesis I

The hypothesis that the intervention would be associated with higher Self-Esteem in the treatment group compared with the control group at T2 (posttest), and T3, T4 and T5, was not supported. Results of the between-subjects effect in Table 2, show that Self-Esteem by Group interaction effect, produced a non-significant eta square of .08, indicating an estimated main effect of 0.8% in lower Self-Esteem effect by the treatment group compared with the control group. The within-subjects effects produced a significant eta square of .018, indicating that the intervention had an estimated main effect of 1.8% decrease in total Self-Esteem for the two groups between baseline and exit (F = 5.412, df 4,303, p < .01). Results in Tables 2 and 3 show that from baseline to T5, self-esteem measures decreased for both groups, with divergence between groups increasing significantly from T3 to T5 (F = 5.412, df 4,303, p < .01), (see Table 3 for changes in mean self-esteem scores). Over time, treatment subjects' self-esteem ratings declined more that did those of the control subjects. This finding runs contrary to the hypothesis that the treatment group would experience increases in self-esteem measures.

Results of the descriptive statistics presented in Table 3, show that except for T2, when the mean was slightly higher, mean Self-Esteem for the treatment group was lower compared to the control group, and was significantly lower at T5 with a mean of 32.5 ([+ or -] 3.3) compared with 33.5 (+3.1) for control group (t = -2.416, p < .01). Standard deviations for mean Self-Esteem were larger for the treatment group ranging from 5.2 at baseline to 3.3 at exit, compared with the control group range between 3.9 at baseline and 3.1 at exit.

Hypothesis II

The second hypothesis that higher Self-Esteem would be associated with lower Alcohol Use and lower Drug Use at T2, T3, T4 and T5 was not fully supported (see Table 4). Self-Esteem was actually lower in the treatment group. However, results presented in Table 4, show at T5 there was a significant interaction effect between Self-Esteem and Alcohol Use of .023, indicating that Self-Esteem accounted for 2.3% of the difference in Alcohol Use at the 15 months follow-up (F = 6.967, df 1,298, p < .01). Likewise, results were significant for the interaction effect between Self-Esteem and Drug Use at T5, producing main effect of .027, indicating that Self-Esteem accounted for 2.7% of the difference in Drug Use at the 15 months follow-up (F= 8.113, df 1,298,p < .01).

Hypothesis III

The hypothesis that associations between higher Self-Esteem and lower Alcohol Use and Drug Use would be greater among experimental than among contrail subjects at T2, T3, T4, and T5 was not supported. Results presented in Table 4, show a significant three-way interaction effect at T5 for the Self-Esteem * Alcohol Use * Group, producing 0.014 in main effect indicating that at TS, when Self-Esteem was significantly lower, it had accounted for 1.4% lower Alcohol Use in the experimental treatment group [F (1.298) = 4.104, p < .05].

As indicated by the descriptive statistics presented in Table 5, the treatment group had significantly lower Alcohol Use at T2 IT(l, 298) = -3.818, p < .001], and at T4 [T(1,298) = -3.564, p < .001]. Mean Drug Use was also significantly lower for the treatment group at T2 [T(1,298) = -3.366,p < .01]. At T5, the treatment group also showed a slightly higher mean Drug Use of 5.1 (+/- 7.1) vs. the control group mean of 4.5 (+/-5.7).

Limitations

Results of the study are limited to homeless adult male substance abusers and are based on analyses of data that were collected between 1991 and 1993. The experimental subjects participated in a newly designed shelter-based treatment program for substance abuse that was established as part of a federal research demonstration project. As such, various elements in the program may not have reflected the typical residential treatment programs for homeless substance abuser. Subject attrition and missing data were problematic for the study. Attrition was attributed to voluntary withdrawal and involuntary dismissals of subjects. Despite the attrition problem the program evaluators found that 74% of subjects had maintained a mean length of stay of 57 days and that the mean for treatment "completers" at 100.2 days (Orwin, et al. 1995). Primarily subjects' non-response, skipping or inability to understand some of the questions caused the problem of missing data. As was stated earlier the researchers determined that the level of missing data was not severe to affect the power of these data. In the present study, whenever necessary, contrast coding (Cohen and Cohen, 1983) was used for addressing missing data.

DISCUSSION

Results indicated that the experimental treatment was not associated with increase in Self-Esteem, and that contrary to the predicted results, the experimental treatment interventions were associated with the lowering of Self-Esteem within treatment subjects compared with control subjects. Some of the variability in Self-Esteem may be accounted for by the effects of regressing towards the mean between T1 and T2. However, for reasons unexplained, Self-Esteem was higher in control group, and results showed that it was significantly higher in control subjects at T5. Other researchers have found similar a discrepancy between theory and research findings and concluded that self-esteem appears to behave differently with different populations. For example, Kaplan (1996) found that "help seekers" often reported considerably lower self-esteem than did "help rejecters." Given that in the current study 85% of subjects used cocaine, and treatment subjects can be viewed in the role as "help seekers," it is easy to see why they reported lower levels of Self-Esteem.

Another reason for lower Self-Esteem by the treatment group may stem from the way the Rosenberg Inventory was rated. While interviewers administered the tool it was based on subjects' self-reports and the assumption that individuals would be accurate in reporting their true attitudes and behaviors. The analysis shows a greater diminishing pattern in Standard Deviation scores in Self-Esteem for treatment subjects between T3 and TS, while these scores remained almost constant for control subjects. This implies that treatment subjects were being more conscientious, and probably more honest and realistic in rating their Self-Esteem Inventories, consequently their lower scores at T3, T4 and T5. This result is therefore not necessarily negative. It may be an indication treatment effects on the experimental subjects who were encouraged to be open and honest about their feelings, share in group sessions and attend and participate in AA and NA meetings, where these qualities were reinforced.

The results of testing Hypothesis I (i.e., that Self-Esteem would show an inverse relationship with Alcohol Use and Drug Use at T2, T3, T4 and T5) indicated that the hypothesis was not fully supported. With respect to Hypothesis III, repeated-measures MANCOVA indicated that inverse relationships between Self-Esteem and Alcohol Use and Drug Use would be greater in the treatment group than the control group was supported for Alcohol Use only at the 15-months follow-up.

The pattern of results indicated that despite support in the literature, the overwhelming findings are that the experimental treatment was associated with general lessening self-esteem, that at the end of the study self-esteem was significantly lower in the treatment than the control group, and that this was associated with significantly higher alcohol use by treatment subjects. Similar disagreements between theory and findings by empirical research have been reported. For example, after reviewing several empirical studies CSAP (1997) concluded that cocaine users exhibit unusually high levels of self-esteem before use onset. Alder et al (1992) also concluded that the consensus of research findings suggests that as a theory, self-esteem does have intuition on its side, but not a monopoly on convincing empirical and universally collaborated evidence. After reviewing available evidence CSAP (1997) concluded that increased self-esteem probably should not be used either as a measure of the effectiveness of a substance abuse prevention effort or as an objective of prevention efforts.

IMPLICATIONS

Results of this evaluation have important implications for education, practice, research, and policy. Educators and researchers who use theories about self-esteem, need to exercise caution lest they fall prey to the observation made by Krohn et al. (1996) that many current theories informing substance abuse prevention are incomplete and misleading because often they have emphasized unidirectional relationships between psychosocial variables and abuse. In concluding, self esteem appears to be more a complicated and changeable psychological measure than generally realized. It has been noted that during periods of inflated ego, due to occasional positive life events, some individuals will consume more alcohol than usual. The same is true of some whom also consume more alcohol during periods of depressing or stressful events.

Educators, researchers and practitioners need to also be aware that other psychological measures that may be more useful in the treatment of substance use disorder include changes in such areas as future orientation, family conflict, or self-perceived social competence, because improving self-esteem seems more difficult to achieve. Educators, researchers, practitioners and public policy makers must reconsider whether enhancing self-esteem should remain an active goal for behavior modification interventions programs that serve adult homeless male substance abusers in urban settings. The findings of this evaluation seem to suggest that treatment interventions that serve this population need to moderate and help participants develop a more realistic and authentic self-image, rather than seeking to heighten their sense of self.

Table 1
Demographic and Social Characteristics of the
Study Population at Baseline

                                        Treatment     Control
                                        Group         Group

                                        N       %     N      (%)
                                        193           112
African American                        111    57     69     61.6
Caucasian                                40    21     27     24
Hispanic                                 43    22     16     13.4
Employed (during past 30-days)           63    33     36     32
Homeless (during past 60 days)           71    37     47     42.3
Used Alcohol (during past 30 days)      154    80     67     75
Used Cocaine (during past 30 days)      172    89     100    89.3
Used Marijuana (during past 30 days)     77    40     35     31.3
Used Heroin (during past 30 days)        68    35.2   24     21.4
Used Combination Drugs (past 30 days)   159    82.3   91     81.1

                                        M      SD     M      SD
Mean Age                                32.6   6.7    32.5   7.5
Mean # Years of Education               11.4   2.2    11.4   1.9
Mean Homeless Nights (past 60 days)     16.0   20.9   15.6   21.2
Mean # Years of Lifetime Alcohol Use    11.5   8.4    10.8   9.0
Mean Alcohol Use (past 30 days)         12.0   11.2   11.0   11.4
Mean # Years of Lifetime Drug Use       12.9   7.0    11.6   6.8
Mean Drug Use (past 30 days)            18.1   11.1   16.3   11.0

Adapted from "Assessing the value of a short-term residential drug
treatment program for homeless men," by Lam et al. 1995, Journal of
Addictive Diseases, 14, (4). 212-39.

Table 2

Multivariate Analysis of Variance for
Treatment Effects on Self-Esteem

Source           df       F      Eta Square

                      Between subjects

GP                1   2.500      .008
  error         303   (6879.6)

                      Within subjects

Time            4     5.412 **   .018
Time x GP       4     .787       .003
  error (Time)  1212  (17059.5)

Note: Eta square ([eta].sup.2] indicates the estimated size of the
treatment effect on the dependent variable (George, and Mallery 2001).

Values enclosed in parentheses represent mean square errors.

* p < .05. ** p < .01. *** p < .001

Table 3
Mean Self-Esteem at T1, T2, T3, T4, and T5

         Treatment Group  Control Group

             N            N
             193 (63%)    112 (37%)

Variable        M    SD    M    SD       T

Self-Esteem1  33.6   5.2  33.9  3.9  -0.568
Self-Esteem2  33.1   3.5  32.9  4.5   0.218
Self-Esteem3  32.4   3.9  32.6  4.7  -0.444
Self-Esteem4  32.3   3.8  32.8  3.1  -1.452
Self-Esteem5  32.5   3.3  33.5  3.1  -2.416 **

* p < .05.

** p < .01.

Table 4
Multivariate Analysis of Covariance for Self-Esteem and
Alcohol Use and Drug Use

Source                           df        F         Eta Square

Between subjects

Self-Esteem1 x Alcohol Use       1          .048           .000
Self-Esteeml x Drug Use          1         3.588           .012
Self-Esteem2 x Alcohol Use       1          .037           .000
Self-Esteem2 x Drug Use          1         1.074           .004
Self-Esteem3 x Alcohol Use       1          .260           .001
Self-Esteem3 x Drug Use          1          .024           .000
Self-Esteem4 x Alcohol Use       1         3.014           .010
Self-Esteem4 x Drug Use          1          .936           .003
Self-Esteem5 x Alcohol Use       1         6.967 **        .023
Self-Esteem5 x Drug Use          1         8.113 **        .027
Group x Alcohol                  1         4.104 *         .014
Group x Drugs                    1         1.390           .005
       error Alcohol        298  (6892.8)
             Drugs          298  (6990.8)

Note: Eta square ([[eta].sup.2] indicates the estimated effect size of
the total variance explained by the independent variable (Georgeand
Mallery 2001).

Values enclosed in parentheses represent mean square errors.

* p < .05. ** p < .01.

Table 5
Mean Alcohol and Drug Use at T1, T2, T3, T4, and T5

          Treatment Group   Control Group

          M     SD          M     SD      T

Alcohol Use

T1      12.1  11.2          11.0  11.4        .675
T2       2.5   6.2           5.2   7.2      -3.818 ***
T3       3.7   6.5           4.4   7.0      -1.242
T4       3.3   6.1           5.9   8.3      -3.564 ***
T5       4.1   6.3           3.6   5.0        .235

Drug Use

T1      18.1  11.1          16.3  10.8       1.202
T2       3.6   7.4           6.5   8.1      -3.366 **
T3       3.7   6.7           5.1   8.0      -1.906
T4       4.8   8.2           5.2   8.0      -0.810
T5       5.1   7.1           4.5   5.7        .624

* p < .05. ** p < .01. *** p < .001.

References

Adler, J., Wingert, P., Houston, P., Manly, H., & Cohen, A. D. (1992). The curse of self-esteem: What's wrong with the feel-good movement. Newsweek Inc. NY: Feb. 17, 1992, 46-52.

Allen, L.R. (1969). Self-esteem of male alcoholics. Psychological Record, 19, 381-389.

American Psychiatric Association. (1994). Diagnostic and Statistical Manual of Mental Disorders (4th ed.). Washington, DC: Author

Bandura, A. (1977). Social Learning Theory. Englewood Cliffs, N.J.: Prentice Hall.

Beckman, L.J. (1978). Self-esteem of women alcoholics. Journal of Studies on Alcohol, 39, 491-498.

Botvin, G.J., Schinke, S.P., Epstein, J.A., Diaz, Z. T., & Botvin, E.M. (1995b). Effectiveness of culturally focused and generic skills training approaches to alcohol and drug abuse prevention among minority adolescents: Two-year follow-up results. Psychology of Addictive Behaviors, 9(3), 183-194.

Breaky, W (1987). Treating the homeless. Alcohol Health and Research World, 11, 42-47.

Charalampous, K.D., Ford, B.K., & Skinner, T.J. (1976). Self-esteem in alcoholics and nonalcoholics. Journal of Studies on Alcohol, 37, 990-994.

Coopersmith, S. (1967). The Antecedents of Self-Esteem. San Francisco, CA: W.G. Freedman.

Cohen, J., and Cohen, P. (1983). Applied Multiple Regression/Correlation Analysis for the Behavioral Sciences (Second edition). NJ: Lawrence Erlbraum Associates.

Corbin, W. R., McNair, L. D., & Carter, J. (1996). Self-esteem and problem drinking among male and female college students. Journal of Alcohol and Drug Education, 42, 1-14.

Crocker, J. R. Luthanen, B. Blaine, & S. Broadnax (1994). Collective self-esteem and psychological well being among White, Black, and Asian college students. Personality and Social Psychology Bulletin, 20(5), 503-513.

CSAP (1997) Center for Substance Abuse Prevention. Unproven Programs/Strategies: Self-Esteem Enhancement Programs. Western Regional Center for the Application of Prevention Technologies. In Selected Findings in Prevention: A Decade of Results from the Center for Substance Abuse Prevention.

Donnelly, J. (2000). Self-Esteem and its Relationship to Alcohol and Substance Abuse Prevention in Adolescents. Unpublished Final Report. Dept. of Health Professions/PERLS, Montclair State University. NJ: Upper Montclair.

Engs, R.C., & Hanson, D.J. (1989). Gender differences in drinking patterns and problems among college students: A review of the literature. Journal of Alcohol and Drug Education, 35(2), 36-47.

George, D., and Mallery, P. (2001). SPSS for Windows Step by Step. Boston, MA: Allyn, & Bacon.

Glindemann, K.E., Geller, E.S., & Fortney, J.N. (1999). Self-esteem and alcohol consumption: A study of college drinking behavior in a naturalistic setting. Journal of Alcohol and Drug Education, 45(1), 60-71.

Gorman, D. M. (1996). Etiological theories and the primary prevention of drug use. Journal of Drug Issues, 36, 505-20 Catalano et al. (1996).

Grills, C. & Longshore, D. (1996). Africentrism: Psychometric analyses of a self-report measure. Journal of Black Psychology, 22(1), 86-106.

Gross, W.F. & Adler, L.O. (1970). Aspects of alcoholics' self-concepts as measured by the Tennessee Self-Concept Scale. Psychological Reports, 27, 431-434

Jessor, R., & Jessor, S.L. (1977). Problem Behavior and Psychosocial Development: A Longitudinal Study of Youth. New York: Academic Press.

Kaplan, H.B. (1975c). Sequelae of self-derogation: Predicting from a general theory of deviant behavior. Youth and Society, 7, 171-197.

Kaplan, H.B. (1996) Empirical validation of the applicability of an integrative theory of deviant behavior to the study of drug use. Journal of Drug Issues, 26, 345-378.

Krohn, M.D., Lizotte, A.J., Thornberry, T.P., Smith, C., & McDowall, D. (1996). Reciprocal causal relationships among drug use, peers, and beliefs: A five-wave panel model. Journal of Drug Issues, 26(2), 405-428.

Laflin, M.T., Moore-Hirschl, S., Weis, D.L., & Hayes, B.E. (1994). Use of the theory of reasoned action to predict drug and alcohol use. International Journal of the Addictions. 29(7), 927-940.

Lam, j. A., Jekel, J. F., Thompson, K. S., Leaf, P. J. Hartwell, S. W., & Florio, L. (1995). Assessing the value of a short-term residential drug treatment program for homeless men. Journal of Addictive Diseases, 14(4), 212-39.

Leaf, P. J. (1993). Partnerships in recovery: Shelter-based services from homeless cocaine abusers. New Haven: Haworth Press. N/AAA, Final Report, Rockville, MD. 1994. Substance abuse treatment programs for homeless persons. Albuquerque, NM: Lovelace Institutes.

Maney, D.W. (1990). Predicting university students' use of alcoholic beverages. Journal of College Student Development, 31(1), 23-32.

Martlatt, G. A., & Gordon, J.R. (1985) Determinants of relapse: Implications for the maintenance of behavior change. In, Watson, L. (1991). Paradigms of recovery: Theoretical implications for relapse prevention in alcoholics. The Journal of Drug Issues, 21(4), 839-858.

Mitic, W.R. (1980). Alcohol use and self-esteem of adolescents. Journal of Drug Education, 10, 197-208.

Nielsen, A. L., & Scarpitti, F. R. (1997). Changing the behavior of substance abusers: Factors influencing the effectiveness of therapeutic communities. Journal of Drug Issues, 27, 279-98.

ONDCP (1998a). The 1998 Report of the Director, Office of National Drug Control and Prevention. Available at http://www.health.org.

Orwin, R.G., Sonnefeld, L. J., Jacobs, M., Oram, S., Garrison-Morgen, R., & Blasinky, M., (1995). Cross-site synthesis of intervention designs and implementation analyses from the NIAAA Cooperative Agreement Program for Homeless Persons with ATOD Problems. Journal or the American Psyhological Association.

Parish, J.G., & Parish, T.S. (1991). Support systems functionality, self-concepts, and alcohol use. Journal of Drug Issues, 26, 345-77.

Rickwood, D.J. & V.A. Braithwaite (1994). Social-psychological factors affecting help-seeking for emotional problems. Social Science and Medicine 39(4), 563-572.

Rosenberg, M. (1965). Society and the adolescent self-image. New Jersey: Princeton University Press.

Shaffer, H.J. & Jones, S.B. (1989). Quitting Cocaine: The Struggle Against Impulse. Lexington, Kentucky: Lexington Books.

Schaeffer, G.M., Schuckit, M.A., & Morrissey, E.R. (1976). Correlation between two measures of self-esteem and drug use in a college sample. Psychological Reports, 39, 915-919.

Schroeder, D.S., Laflin, M.T., & Weis, D.L. (1993). The relationship between self-esteem and drug use: Methodological and statistical limitations of the research. Journal of Drug Issues, 23(4), 645-665.

Seeman, M., & Seeman, A.Z. (1992). Life strains, alienation, and drinking behavior. Alcoholism: Clinical and Experimental Research, 16(2), 199-205.

Segal, B., Rhenberg, G., & Sterling, S. (1975). Self-concept and drug and alcohol use in female college students. Journal of Alcohol and Drug Education, 20, 17-22.

Ward, S. (1995). Filling the World with Self-Esteem: A Social History of Truth-Making. In "The Making of Serious Speech: A Social Theory of Professional Discourse." Current Perspectives in Social Theory, 15, 63-81.

Watson, L. (1991) Paradigms of recovery: Theoretical implications for relapse prevention in alcoholics. The Journal of Drug Issues, 21(4), 839-858.

Yanish, D.L., & Battle, J. (1985). Relationship between self-esteem, depression, and alcohol consumption among adolescents. Psychological Reports, 57, 331-33.

Barris P. Malcolm

University of Connecticut School of Social Work

West Hartford, Connecticut 

Author Note

Baris P. Malcolm, University of Connecticut School of Social Work, West Hartford, Connecticut.

Quality data produced by experimental-design studies of homeless individuals are so rare that even dated information is a treasure trove, therefore much appreciation goes to Dr. James Jekel, and Dr. Phillip Leaf, co-principal investigators of the Grant Street Partnership project for permission to use their data. The author also thanks Dr. David Cournoyer, Dr. Kay Davidson, Dr. Harriette Johnson, and Dr. Michie Hesselbrock for proofreading drafts of the manuscript.