An Attempt to Improve Self-Esteem by Modifying

Specific Irrational Beliefs

by

David M. Nielsen

John J. Horan

Bethanne Keen

Carolyn Cox St. Peter

Sherry Dyche Ceperich

Deborah Ostlund

Division of Psychology in Education

Arizona State University

RUNNING HEAD: Irrational Beliefs

The research described in this paper was supported by grants from the Sally M. Berridge Foundation and Southwest Gas Corporation. The authors are also indebted to Harold Slemmer and Richard Baniszewski who facilitated the conduct of this study in their school district.

(1996) Journal of Cognitive Psychotherapy, 10, 137-149.

Abstract

In a double-blind study, subjects with low to moderate self-esteem received either of two cognitive restructuring interventions. The experimental treatment addressed specific irrational beliefs previously found to be correlated with low self-esteem; the control treatment focused on irrational beliefs not empirically related to self-esteem. Each intervention produced appropriate changes on targeted irrationality measures. The pattern of changes on targeted and nontargeted irrationality scales, however, suggests that improvements in specific rationality readily generalize, a phenomenon which may have obscured posttest differences between the two interventions on a battery of self-esteem measures. Nevertheless, the self-esteem improvements of subjects within each treatment were consistently related to changes on the previously linked beliefs, and conversely, only sporadically related to changes on the non-linked beliefs.

An Attempt to Improve Self-Esteem by Modifying

Specific Irrational Beliefs

Many ex post facto studies have reported significant relationships between measures of irrationality and a wide array of psychological problems including anger (Zwemer & Deffenbacher, 1984; Rohsenow & Smith, 1982), anxiety (Tobacyk & Downs, 1986; Zwemer & Deffenbacher, 1984), depression (Cook & Peterson, 1986; Hyer, Jacobsen, & Harrison, 1985; Rohsenow & Smith, 1982; Van Den Bout, 1986; Vestre, 1984), poor problem-solving (Heppner, Reeder, & Larson, 1983; Tobacyk & Milford, 1982), and even schizophrenia (Newmark & Whitt, 1983). The presumption that irrational beliefs play a prominent role in the etiology of these dysfunctions is implicit in the reported correlations.

A subset of this research topic examining whether specific irrational beliefs are differentiallyrelated to these clinical problems is particularly intriguing (e.g., Cash, 1984; Deffenbacher, Zwemer, Whisman, Hill, & Sloan, 1986). Daly and Burton (1983), for instance, used the Irrational Beliefs Test (IBT, Jones, 1969) to chronicle the kinds of irrationality related to the development of low self-esteem; they found that four IBT subscales were significant predictors of low self-esteem; the other subscales were not. McLennan (1987) confirmed that these same irrational beliefs were associated with low self-esteem assessed with a different measure.

The Daly and Burton (1983) and McLennan (1987) studies provided confirmatory evidence in favor of the relationship between specific irrational beliefs and low self-esteem. The opportunity for disconfirmatory evidence to appear was subsequently explored by Erickson, Horan, and Hackett (1991); they attempted to replicate the Daly and Burton (1983) and McLennan (1987) investigations with a younger population, control measures of varying theoretical relevance to self-esteem, and alternative assessment methods.

Erickson and her associates (1991) found that two of these irrational beliefs (Demand for Approval and Anxious Overconcern) continued to predict low self-esteem with this younger population. Furthermore, theoretically appropriate divergent relationships appeared as well. For example, Demand for Approval was unique to low self-esteem, and Anxious Overconcern was shared only with depression, a problem conceptually and psychometrically linked to low self-esteem (Beck, 1990a). Moreover, neither belief was related to theoretically remote control measures (extraversion, facilitative anxiety, and grade-point-average); conversely, other irrational beliefs correlated with these control measures were not associated with low self-esteem.

Although three studies have now connected specific irrational beliefs to low self-esteem, current research has not established which causes the other or whether both are co-effects of yet another underlying variable. The matter is of considerable import to the science and practice of clinical and counseling psychology. Low self-esteem is ubiquitous among clients, and is inextricably involved with other psychological dysfunctions (Beck, 1967; Lorr & Wunderlich, 1988; Miller, 1988; Rosenberg, 1962). If specific irrational beliefs are causally linked to low self-esteem, then efficient therapeutic interventions could readily be developed and applied.

Rational-emotive theory has, of course, undergone revision over the past decade, and the subscales used to assess specific irrational beliefs have not escaped conceptual and empirical scrutiny (e.g., see Burgess, 1990). Nevertheless, all of the foregoing coefficients emerged from "classic Ellis" hypothesis testing, and continued work within that framework is still highly relevant to the appropriate evolution of rational-emotive theory. Our study attempted to test the possible causal nature of the relationship between specific irrationality and low self-esteem in a double-blind experiment which contrasted two different cognitive restructuring interventions. The first addressed specific irrational beliefs consistently correlated with low self-esteem; the second focused on irrational beliefs which have not been related to self-esteem in any previous investigation. We expected that each treatment would produce changes in the targeted irrational beliefs, but that only the first would effect concomitant gains in self-esteem.

Method

Subjects

Eighty-five eleventh-grade students enrolled in introductory high-school psychology courses obtained parental permission and gave their own consent to participate. Their ages ranged from 16 to 18 with a mean of 16.67; 60% were young women, 76% were white. Hispanics (14%) comprised the largest minority.

To preclude the possibility of a posttest ceiling problem on the assessment devices, screening scores ( 15) were established for each of the four targeted irrational beliefs; also the pretest scores on all self-esteem measures were pooled, and only those in the bottom two-thirds of the distribution were treated. Two additional subjects were excluded because of incomplete pretest protocols, and two others exceeded the rationality criterion. The remaining 50 subjects, all with low to moderate self-esteem, were retained and randomly assigned to either of the two treatments.

We opted to use subjects at this particular level of functioning, rather than say, a clinical sample defined by the lowest 5%, because the original relationships between specific rationality and self-esteem were computed on normal samples with scores spanning the entire distributions. We would not expect the correlations undergirding our experimental hypotheses to hold within the attenuated ranges of a clinical sample, and the appropriate treatment of such subjects would require a more complex intervention protocol. Our findings would thus be more relevant to research efforts in prevention than remediation.

Measures

Rationality Measures. The Irrational Beliefs Test (IBT, Jones, 1969) consists of 100 Likert items arrayed on 10 subscales corresponding to each of Ellis' (1962) ten irrational beliefs. Sample items include "It is important to me that others approve of me" and "I hate to fail at anything". Jones labeled the 10 subscales as follows: Demand for Approval, High Self-Expectations, Blame Proneness, Frustration Reactivity, Emotional Irresponsibility, Anxious Overconcern, Problem Avoidance, Dependency, Helplessness, and Perfectionism.

Jones reported internal consistency estimates for the individual scales ranging from .45 to .72, a test-retest coefficient of .92, and concurrent validity r of .61 that involved relationships with psychiatric problems. Lohr and Bonge (1982) essentially replicated the factor structure reported by Jones; however, a frustration-reactivity factor was not supported.

In our study, the IBT subscales served two functions. First they permitted a manipulation check on the independent variable. Both cognitive restructuring treatments would be required to effect changes on the specific measures of irrationality which they were constructed to address. In fact, if the experimental treatment produced changes in self-esteem in the absence of changes on the targeted beliefs, the construct validity of the experiment would be challenged (see Cook & Campbell, 1979). Second, if changes on the targeted beliefs for each intervention did occur, then concomitant changes on nontargeted beliefs would suggest that the effects of either intervention had generalized.

Self-Esteem Measures. Three self-report measures of self-esteem were employed: 1) The Janis-Field Feelings of Inadequacy Scale (Eagly, 1967) is a widely used measure of self-esteem in social areas. It contains 20 items on five-point Likert scales (e.g., "How often do you have the feeling that there is nothing you can do well?"). Crandall (1973) reported split-half reliabilities ranging from .72 to .88, and convergent validity coefficients of .67 and .84 with other self-esteem measures. Hamilton (1971) noted an r of -.36 with dominance ratings, suggesting divergent validity. 2) The Rosenberg Self-Esteem Scale (Rosenberg, 1965) consists of 10 items on four-point Likert scales that measure self-worth or self-acceptance (e.g, "On the whole, I am satisfied with myself" and "At times I think I am no good at all"). Silber and Tippett (1965) found a two-week test-retest reliability of .85 and convergent validity coefficients ranging from .56 to .83 with similar measures of self-esteem. 3) The Piers-Harris Children's Self-Concept Scale (Piers, 1984) is an 80 item self-report inventory calling for "yes" or "no" responses to items such as "I can be trusted." It provides a global and specific measures of self-esteem; however, only global scores were used in this study. Chiu (1988) noted internal consistencies ranging from .88 to .93; concurrent validity, based on correlations with other self-concept measures, varied from .32 to .85.

The first two instruments were employed in prior research upon which this study is based (i.e., Daly & Burton, 1983; McLennan, 1987; Erickson, et al., 1991); the last in order to better "triangulate on the referent" (Cook & Campbell, 1979, p. 65). Finally, although self-esteem is almost always assessed through self-report alone, Chiu (1987) developed a self-esteem measure derived from Teacher Reports (which he labeled The Self-Esteem Rating Scale for Children); this measure, consisting of 12 items (e.g., "Hesitates to speak up in class" and "Becomes upset when being criticized or scolded"), was included on an exploratory basis to address the issue of mono-method bias (Cook & Campbell, 1979). Chiu (1987, 1988) reported internal consistency coefficients ranging from .70 to .91 (with a median of .80), a test-retest coefficient of .93, interrater reliabilities between .82 and .86, and concurrent validities from .48 to .58. Erickson and her associates (1991), however, failed to obtain satisfactory interrater reliability when high-school teachers provided the data.

Control Measure. A five-item scale designed to assess the possibility of differential demand characteristics inherent in the experiment was derived from the work of Borkovec and Nau (1972). Sample items (e.g., "How logical does this type of treatment seem to you?" and "How confident would you be in recommending this treatment to a friend who was experiencing problems with their feelings?) were rated on 7-point Likert scales. Differences on this measure would suggest that changes in self-esteem were confounded with placebo-related phenomena, and hence, were not expected.

Subjects completed the self-report measures of irrationality and self-esteem a week before and two days after completing treatment. Each subject's English teacher concurrently completed the Teacher Reports measure at these times. Subjects responded to the control measure at the end of session #2 and again at the end of session #4.

Procedures

The treatments were delivered in the context of small groups meeting for four 50-minute treatment sessions. Eight groups were formed (four experimental and four control) ranging in size from five to eight members. Each group was led by one of four women; three were advanced doctoral students in an APA-approved psychology program, the fourth had a masters degree in counseling. Each counselor led one experimental and one control group.

The counselors had no knowledge of any subject's treatment condition. Moreover, because the relationship between the specific measures of irrationality and self-esteem is empirical rather than logical, it was possible to keep the counselors completely blind to the experimental or control nature of the treatments they delivered. Post-experimental debriefing indicated that none of the four counselors could accurately guess the hypotheses under investigation.

Structured treatment protocols, for both the experimental and control conditions, were written for each treatment session. (Copies of the treatment manuals, totaling 80 pages, are available for the cost of duplication and postage.) Prior to the onset of the study, the counselors received three hours of didactic and role-play training. In order to standardize the sessions, counselors were instructed to follow the protocols verbatim during each session. Brief periods of unstructured discussion were scheduled during which subjects could generate personal examples of rational thinking. The counselors were monitored during each session to assure compliance with each treatment regimen.

The contents of the two treatments differed only in the specific irrational beliefs that were discussed. Subjects in the experimental condition were repeatedly presented with disconfirmatory evidence regarding the irrational beliefs of Demand for Approval and Anxious Overconcern; those in the control condition were similarly given disconfirmatory evidence regarding the irrational beliefs of Emotional Irresponsibility and Blame-Proneness. Issues related to self-esteem were not included in either protocol.

The small groups were conducted during regularly scheduled psychology classes. The interventions followed a typical cognitive restructuring format, identical for experimental and control conditions. During the first two sessions, all subjects were introduced to the rationale for evaluating, challenging and replacing internal, irrational beliefs based on Ellis and Harper's (1975) "A-B-C" theory of personality. Experimental subjects reviewed the domain represented by the Demand for Approval subscale items; control subjects similarly focused on the domain represented by the Emotional Irresponsibility subscale items. All subjects were taught how to challenge these beliefs using a "Disputing Irrational Beliefs (DIB) Instruction Sheet" (Ellis & Harper, 1975).

Sessions three and four followed the same format; counselors taught the subjects how to refute the second of the two beliefs relevant to their respective treatments. Experimental subjects focused on Anxious Overconcern; controls on Blame Proneness. Make-up sessions were offered to subjects who missed their first or second sessions. Scheduling constraints, however, precluded make-ups for subjects missing the third or fourth session. Subjects who missed more than one session without attending a make-up session were dropped from the analysis.

Results

Preliminary Analyses

Attrition. Complete pretest and post-test data was collected from 44 (24 experimental and 20 control) of the original 50 subjects. One subject in each condition had missed two treatment sessions; four others were absent from school during post-test data collection. Attrition was equivalent across both conditions (x² (1) = .82, n.s.).

Seven additional subjects (five experimental and two control) missed, and did not make-up, a single treatment session. Analyses conducted with and without including them yielded similar outcomes; these seven subjects are included in all data reported below.

Pretest Equivalence. A MANOVA (using Wilks's lambda) conducted on the four pretest measures of self-esteem was not significant, multivariate F (4,45) = .177, indicating that the experimental and control conditions did not differ in self-esteem prior to treatment.

Control Analysis. A 2 x 2 (Treatment by Repeated Measures) ANOVA on the demand measure administered early and late in treatment revealed no significant main effects or interactions, indicating that placebo-related phenomena were equivalent over the course of the experimental and control conditions. This equivalence was not a function of ceiling effects; mean values per condition and time reflected moderate levels of experimental demand.

Counselor Effects. A 2 x 4 (Treatment by Counselor) MANOVA (using Wilks's lambda) conducted on the four self-esteem posttest measures yielded no significant main effects or interactions. Thus, counselor effects were controlled and evenly disbursed.

Psychometric Analyses. Internal consistency reliability coefficients (s) were calculated on the pretest measures of all subjects in this study. The targeted IBT subscale s for Demand for Approval (.81), Anxious Overconcern (.47), Emotional Irresponsibility (.72), and Blame-Proneness (.46) were variable but consistent with those reported by Jones (1969). In contrast, the reliabilities of all self-esteem dependent measures were uniformly high, namely .90, .91, .94, .78 for the Janis-Field, Rosenberg, Piers-Harris, and Teacher Reports, respectively.

Although intercorrelations among all three self-report self-esteem measures were at least moderate and highly significant (rs of .70, .62, .48 were obtained, all ps < .001), no concurrent validity coefficient involving the Teacher Reports measure was significant or exceeded .18. Finally, a statistical power analysis adapted from Stevens (1986) indicated sufficient strength (i.e., .8) to detect large effects at an of .05. Table 1 presents the means and standard deviations for experimental and control subjects on both testing occasions.

Insert Table 1 about here.

Treatment Effects

Effects on Irrationality. Multivariate and univariate ANOVAs (two treatments by two repeated measures) were conducted on the targeted and nontargeted irrational belief scales. The initial MANOVA (using Wilks's lambda) on the four targeted beliefs yielded a significant interaction, multivariate F(4, 39) = 3.77, p < .01, and a large main effect for repeated measures, multivariate F(4, 39) = 13.70, p < .001. Follow-up univariate ANOVAs showed significant interactions on one targeted experimental belief [Demand for Approval F (1, 42) = 5.3, p < .03] favoring the experimental treatment and on one targeted control belief [Emotional Irresponsibility F(1, 42) = 4.54, p <. 04] favoring the control treatment. Repeated measures effects reflecting improvement over time occurred on both targeted experimental beliefs [Demand for Approval F(1, 42) = 32.49, p <. 001; Anxious Overconcern F (1, 42) = 33.33, p <. 001] and on both targeted control beliefs [Blame Proneness F(1, 42) = 24.62, p <. 001; Emotional Irresponsibility F(1, 42) = 3.85, p <. 056].

A subsequent MANOVA (using Wilks's lambda) on the remaining six scales of the IBT(not targeted for either the experimental or control condition) yielded no treatment or interaction effects, but again, a large repeated measures effect emerged, multivariate F(6, 37) = 5.59, p < .001. Univariate follow-up ANOVAs showed improvements for both experimental and control subjects on three of the six measures: High Self-Expectations F(1, 42) = 18.53, p < .001]; Frustration Reactivity F(1, 42) = 15.41, p < .001]; and Helplessness F(1. 42) = 10.45, p <. 002.

In sum, two of the four anticipated interactions were significant indicating that the experimental and control treatments were each successful in improving one of the two beliefs targeted by their protocols. As might be expected these interactions emerged on the two subscales with the highest internal consistencies. Moreover, repeated measures main effects on both targeted and nontargeted irrational beliefs were plentiful (i.e., they occurred on seven of the ten scales) indicating that the experimental and control subjects improved over time on a wide array of specific irrationality scales. This change pattern suggests that the effects of both treatments generalized well beyond their targeted domains.

Effects on Self-Esteem. Since interactions emerged on two of the appropriate irrationality scales, subsequent differences on the self-esteem measures in favor of the experimental treatment would be consistent with the requirements of experimental construct validity (cf, Cook & Campbell, 1979). That is, improvements in self-esteem could logically be attributed to experimentally-produced changes in specific irrational beliefs.

Thus, similar 2 x 2 (treatment by repeated measures) multivariate and univariate ANOVAs were conducted on the battery of self-esteem measures. The MANOVA (using Wilks's lambda) yielded no treatment or interaction effects; however, a significant main effect for repeated measures occurred, multivariate F(4, 39) = 5.33, p < .002]. Follow-up univariate ANOVAs showed consistent main effects for repeated measures: Janis-Fields F(1, 42) = 11.39, p <. 002; Rosenberg F(1, 42) = 8.82, p < .005; Piers-Harris F(1, 42) = 18.53, p < .001]; and Teacher Ratings F (1, 42) = 3.75, p < .06]. In sum, these analyses reflect improvements in self-esteem, over time, for subjects in both the experimental and control treatments.

Ex post facto Analysis

The pattern of changes on the irrationality scales suggests that one can't teach subjects to think more rationally in one area without producing concomitant changes in other areas. This diffusion phenomenon may have obscured differences between the two interventions on the self-esteem battery. For example, the control treatment's success in promoting change on one of the control beliefs may have generalized rather quickly to the targeted experimental beliefs and ultimately then to the self-esteem measures.

Such a scenario, though plausible, was not confirmed by the foregoing analyses. Wholesale repeated measures effects merely suggest that both experimental and control subjects improved in rationality and self-esteem over the course of treatment. If such treatment gains do exist, they would be confounded by any of the well-known sources of internal invalidity, particularly "testing" (c.f., Campbell & Stanley, 1963; Cook & Campbell, 1979). Moreover, the possibility of a regression artifact is particularly salient in the self-esteem data, as these subjects occupied the bottom two-thirds of the pooled distribution of self-esteem measures.

We, therefore, attempted an ex post facto correlational analysis which focused on the relationships between changes in the specific beliefs targeted by the experimental and control treatments and overall gains in self-esteem displayed by all subjects. Such an analysis, of course, is relatively weak and cannot be substituted for the MANOVA tests of causality. On the other hand, it did offer the possibility of illuminating whether change processes occurring within the experiment were at least consistent with theory. Such information would be highly relevant to any decision about continuing research in this area.

Gain scores for all subjects on the measures linked and not linked to self-esteem were correlated with gain scores on the self-esteem measures. These coefficients are presented in Table 2.

Insert Table 2 about here.

All six of the gain-score correlations involving the two linked beliefs and the three self-report self-esteem devices were significant. Only two of the six correlations involving the two nonlinked beliefs and the three self-esteem measures were significant. Essentially, then, the self-esteem of all subjects reliably improved in the presence of changes in the experimental irrational beliefs; conversely, changes in control beliefs were infrequently accompanied by changes in self-esteem.

Discussion

Despite the gradually accumulating bank of evidence that low self-esteem is related to specific irrational beliefs, particularly Demand for Approval and Anxious Overconcern (e.g., Daly & Burton, 1983; Erickson, et al., 1991; Mclennan, 1987), we were not able to demonstrate conclusively, that the relationship is causal. We did find that cognitive restructuring can produce changes in specific rationality. Moreover, the overall patten of changes suggests that the teaching of specific rationality may result in increased general rationality. For example, although subjects in both the experimental and control conditions were taught to become more rational on only two of the ten IBT subscales, they showed increased rationality over time on seven of these subscales. Apparently, as the subjects learned how to challenge specific irrational beliefs and replace them with more reasonable cognitions, they rather quickly began applying the rules of rationality to a broader spectrum of their thinking.

If so, then practitioners need not necessarily target the specific irrational beliefs linked by previous research to self-esteem. Generalization processes, evolving from the successful alteration of other irrational beliefs, would indirectly benefit their clients' self-esteem. Nevertheless, our ex post facto analyses suggest that cognitive restructuring will be successful here only to the extent that it directly or indirectly changes the irrational beliefs that are empirically related to self-esteem.

Although our study was designed to provide unequivocal experimental evidence of the causal relationship between certain specific irrational beliefs and low self-esteem, the foregoing suppositions are presented with the lesser level of certainty inherent in ex post facto data. At the outset of the study we had the option of randomly assigning the available subjects to a third condition that involved a nonrational treatment or no treatment at all. Statistical power considerations led us to prefer comparing two conditions rather than three. Moreover, the rare opportunity to conduct a double-blind treatment study with sophisticated experimenters could not be preserved if an additional qualitatively-different intervention were added. Future research may capitalize on our hindsight. An experimental comparison involving equally credible rational and nonrational interventions would satisfactorily resolve the question.

Finally, the problem of mono-method bias (Cook & Campbell, 1979) continues to plague self-esteem research, at least that subset which involves adolescents and adults. The field is sorely lacking a psychometrically sound measure not based on self-report. Currently, the alternatives are few. For example, none of the three teacher-report self-esteem scales reviewed by Chiu (1988) were developed for high school students. Researchers who study the self-esteem of adolescents are thus left with the problematic choice of declining to employ an other-report measure and incurring the cost of mono-method bias, or using an other-report measure intended for a younger population. We opted for the latter alternative, and although we obtained adequate reliability, no evidence of concurrent validity appeared whatsoever.

References

Beck, A.T. (1967). Depression: Causes and treatment. Philadelphia: University of Pennsylvania Press.

Beck, A.T. (1990). Beck self-concept test. Psychological Assessment, 2, 191-197.

Borkovec, T.D. & Nau, S.D. (1972). Credibility of analogue therapy rationales. Journal of Behavior Therapy and Experimental Psychiatry, 3, 257-260.

Burgess, P. M. (1990). Toward resolution of conceptual issues in the assessment of belief systems in rational-emotive therapy. Journal of Cognitive Psychotherapy: An international Quarterly, 4, 171-184.

Campbell, D.T., & Stanley, J.C. (1963). Experimental and quasi-experimental designs for research. Chicago: Rand McNally.

Chiu, L.H. (1987). Development of the self-esteem rating scale for children (revised). Measurement and Evaluation in Counseling and Development, 20, 36-41.

Chiu, L.H. (1988). Measures of self-esteem in school-age children. Journal of Counseling and Development, 66, 298-301.

Cook, T.D., & Campbell, D.T. (1979). Quasi-experimentation: Design and analysis issues for field settings. Chicago: Rand McNally.

Cook, M.L., & Peterson, C. (1986). Depressive irrationality. Cognitive Therapy and Research, 10, 293-298.

Crandall, R. (1973). The measurement of self-esteem and related constructs. In J. Robinson, & P. Shaver (Eds.) Measures of social psychological attitudes. Ann Arbor MI: Institute for Social Research.

Daly, M.J., & Burton, R.L. (1983). Self-esteem and irrational beliefs: An exploratory investigation with implications for counseling. Journal of Counseling Psychology, 30, 361-366.

Deffenbacher, J.L., Zwemer, W.A., Whisman, M.A., Hill, R.A., & Sloan, R.D. (1986). Irrational beliefs and anxiety. Cognitive Therapy and Research, 10, 281-292.

Eagly, A.H. (1967). Involvement as a determinant of response to favorable information. Journal of Personality and Social Psychology, 7, 1-15.

Ellis, A. (1962). Reason and emotion in psychotherapy. New York: Lyle Stuart.

Ellis, A., & Harper, R.A. (1975). A new guide to rational living. North Hollywood, CA: Wilshire Book Company.

Erickson, C.D., Horan, J.J., & Hackett, G. (1991, August). On thinking and feeling bad: Do client problems derive from a common irrationality or specific irrational belief? Paper presented at the annual meeting of the American Psychological Association, San Francisco, CA.

Hamilton, D. (1971). A comparative study of five methods of assessing self-esteem, dominance and dogmatism. Educational and Psychological Measurement, 31, 441-452.

Heppner, P., Reeder, B.L., & Larson, L.M. (1983). Cognitive variables associated with personal problem-solving appraisal: Implications for counseling. Journal of Counseling Psychology, 30, 537-545.

Hyer, L.A., Jacobsen, R., & Harrison, W.R. (1985). Irrational ideas: Older vs. younger inpatients. Journal of Nervous and Mental Disease, 173, 232-235.

Jones, R.G. (1969). A factored measure of Ellis' irrational belief system, with personality and maladjustment correlates (Doctoral Dissertation, University of Missouri, Columbia, 1968). Dissertation Abstracts International, 29(11B), 4379-4380.

Lohr, J.M., & Bonge, D. (1982). The factorial validity of the irrational beliefs test: A psychometric investigation. Cognitive Therapy and Research, 6, 225-230.

Lorr, M., & Wunderlich, R. (1988). Self-esteem and negative affect. Journal of Clinical Psychology, 44, 36-39.

Mclennan, J.P. (1987). Irrational beliefs in relation to self-esteem and depression. Journal of Clinical Psychology, 43, 89-91.

Miller, R.L. (1988). Positive self-esteem and alcohol/drug related attitudes among school children. Journal of Alcohol and Drug Education, 33, 26-31.

Newmark, C.S., and Whitt, J.K. (1983). Endorsements of Ellis' irrational beliefs as a function of DSM-III psychotic diagnoses. Journal of Clinical Psychology, 39, 820-823.

Piers, E.V. (1984). Revised manual for the Piers-Harris Children's Self-Concept Scale. Los Angeles: Western Psychological Services.

Rosenberg, M. (1962). The association between self-esteem and anxiety. Journal of Psychiatric Research, 1, 135-152.

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

Rohsenow, D.J., & Smith, R.E. (1982). Irrational beliefs as predictors of negative affective states. Motivation and Emotion, 6, 299-314.

Silber, E. & Tippett, J.S. (1965). Self-esteem: Clinical assessment and measurement validation. Psychological Reports, 16, 1017-1071.

Stevens, J. (1986). Applied multivariate statistics for the social sciences. Hillsdale, NJ: Lawrence Erlbaum Associates.

Tobacyk, J.J., & Downs, A. (1986). Personal construct threat and irrational beliefs as cognitive predictors of increases in musical performance anxiety. Journal of Personality and Social Psychology, 51, 779-782.

Tobacyk, J.J. & Milford, G. (1982). Criterion validity of Ellis' irrational beliefs: Dogmatism and uncritical inferences. Journal of Clinical Psychology, 38, 605-607.

Van Den Bout, J. (1986). Attributional and irrational cognitions. Psychological Reports, 59, 951-954.

Vestre, N.D. (1984). Irrational beliefs and self-reported depressed mood. Journal of Abnormal Psychology, 93, 239-241.

Zwemer, W.A., & Deffenbacher, J.L. (1984). Irrational beliefs, anger, and anxiety. Journal of Counseling Psychology, 31, 391-393.

TABLE 1

Means and Standard Deviations for Experimental and Control

Subjects on Both Testing Occasions.

TABLE 2

Correlations Between Self-Esteem Gain Scores and Targeted

Irrational Belief Gain Scores

^b Targeted for change in the control condition.

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