Running Head: Internet COGNITIVE Restructuring

The Effects of Internet-Based Cognitive Restructuring

On Adolescent Self-Esteem

Sandra M. Dannenbaum-Daubney and John J. Horan

Division of Psychology in Education

Arizona State University

Paper presented at the Annual Meeting of the American Psychological Association, Boston, August 23, 1999

10/25/99

http://horan.ed.asu.edu/d-sd-apa.htm

Abstract

Previous research has linked specific irrational beliefs to low self-esteem; a cognitive-restructuring computer program focused on changing these beliefs produced higher levels of rationality and self-esteem among normatively-low high-school juniors and seniors. Presently, that program was improved, adapted for Internet delivery, and evaluated on a younger but normal sample. Sixty-five seventh- and eighth-grade students were blocked (on gender, ethnicity, and grade-level), randomly assigned to experimental and control conditions, and assessed before and after treatment on a battery of devices reflecting irrationality and self-esteem. The primary outcome analyses did not confirm the program's earlier overall success with the new sample and conditions. Exploratory analyses, however, indicated that self-esteem benefits were evident among eighth grade but not seventh grade participants.

The Effects of Internet-Based Cognitive Restructuring

On Adolescent Self-Esteem

Many ex post facto studies have linked specific irrational beliefs to various psychological dysfunctions such as anger, anxiety, depression, and low self-esteem (see Horan, 1996; Nielsen et al, 1996). For example, Daly and Burton (1983) found that among college students four subscales of the Irrational Beliefs Test (IBT, Jones, 1996) were significant predictors of low self-esteem; the other subscales were not. McLennan (1987) also reported that these same irrational beliefs forecasted low self-esteem assessed by a different measure.

Erickson, Horan, and Hackett (1991) noted that two of these irrational beliefs continued to predict low self-esteem with a younger population. From a construct-validity standpoint, logical divergent relationships appeared as well. For example, 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.

Subsequently, Horan (1996) developed an interactive multimedia computer program to foster rational thinking in belief areas previously associated with low self-esteem. In contrast to a control condition, computer-based cognitive restructuring produced improvements on a battery of devices assessing both rationality and self-esteem.

Several challenges remain in this line of research. In the first place, the relationship between specific irrationality and low self-esteem along the life span remains unclear. The Daly and Burton and McLennan studies involved college students; the Erickson, et al and Nielsen et al projects were conducted on high school students. The relationships between rationality and self-esteem were not perfectly consistent between these populations; it may well be that with younger subjects further differences might arise that would undermine the logical basis of cognitive restructuring for these particular beliefs. Although, Horan's (1996) program proved efficacious with high school juniors and seniors, it is an open question whether similar results might occur with seventh- and eighth-grade adolescents.

Secondly, the Horan (1996) study was remedial in nature; only subjects with low self-esteem were included. It thus remains to be seen whether cognitive restructuring for these specific beliefs could foster improvements in rationality and self esteem among normal populations.

Finally, despite their potential utility, computer-based treatment programs are imperfectly exportable. Assuming such interventions survive empirical scrutiny, how might they be best distributed to those who would benefit? Horan's (1996) program occupies 121 MB of hard-drive space. Even sharing this program with students and colleagues requires Zip drives or the burning of a CD-ROMs. Widespread dissemination of multimedia computer treatments involves serious capital expenditures for producers and consumers alike. Internet-based interventions are much more exportable. Anyone with access to the Internet is a potential client, and anyone with access to a server is a potential publisher.

The present study, therefore, adapted Horan's (1996) program for Internet delivery and experimental evaluation on a younger audience with levels of self-esteem in the normal range of development. The conversion to an Internet format presented several opportunities for improving it. For example, we increased the program's visual appeal by adding benevolent graphical "mentors" (Jones, Valdez, Nowakowski, & Rasmussen, 1995) and visual metaphors (Fenk, 1994; Mayer, 1994; Gyselinck & Tardieu, 1994; McDaniel & Waddill, 1994) to aid in the cognitive restructuring. Moreover, we incorporated the script text into the display graphics to enhance learning and/or allow for hearing impaired participants (Weidenmann, 1994). Finally, we used three-dimensional figures and added both auditory and visual reinforcers for rational responses prior to exiting a module. Consistent with Horan's (1996) earlier hypotheses, we expected that relative to a control treatment, Internet-based cognitive restructuring would produce improvements on a battery of devices reflecting rational thinking and self-esteem.

Method

Sixty-five seventh and eighth-grade adolescents from an urban school were blocked on gender, ethnicity, and grade-level, and randomly assigned to experimental and control conditions.

Two rationality measures were employed: The Irrational Beliefs Test (IBT) (Jones, 1969), a 100 item Likert-type scale comprised of 10 subscales, has displayed adequate reliability and validity coefficients in the literature. For example, Jones (1969) reported (individual scale) internal consistencies ranging from .45 to .72. Jones also found a test-retest coefficient of .92. Given the length of the assessment battery, 44 items not content valid with respect to the curriculum were excluded. The attenuation yielded a .45 internal consistency coefficient using pretest scores of the participants in this study.

The Rational Beliefs Scale (RBS) is an experimenter-made device composed of 13 items scored on a 5-point Likert scale. The items were drawn directly from the introductions to the 13 curriculum modules, e.g., "I feel bad when other people don't like how I look." As such, this instrument might be construed as an independent variable manipulation check as well as an outcome device. An internal consistency coefficient of .68 was obtained from this study's pretest scores.

Three self-esteem measures were also employed, namely, a modified version of the Piers-Harris Children's Self-Concept Scale (Piers, 1984), The Rosenberg Self-Esteem Scale (Rosenberg, 1965), and The Janis-Field Feelings of Inadequacy Scale (Eagly, 1967).

The Piers-Harris is an 80-item yes/no scale; as with the IBT described above, we discarded 23 items not relevant to the treatment contents. Reliabilities reported by other researchers are good, e.g., internal consistencies of .88 to .93 (Chui, 1988). Moreover, it is a well-known and respected instrument in clinical use (Epstein, 1985; Jeske, 1985). Pretest internal consistency for the shortened version in our study was .91.

The Rosenberg is a reliable and valid measure of global or "unidimensional" self-esteem (Gray-Little, Williams, & Hancock, 1996). The scale consists of ten items in a four-point Likert-type format. Silber and Tippett (1965) reported a two-week test-retest r of .85 as well as convergent validity coefficients ranging from .56 to .83. Pretest internal consistency in our study was .83.

The Janis-Field is a 20 item five-point Likert device. Crandall (1973) published split-half reliabilities ranging from .72 to .88 and convergent validity coefficients ranging from .76 to .84. Pretest internal consistency in this study was .83.

In addition to rationality and self-esteem measures, a demand measure was developed from the work of Borkovec and Nau (1972); it contained five-items in a 7-point Likert format. A difference on this control measure, administered at the end of the first treatment session, would suggest that changes in self-esteem were confounded by placebo-related phenomena.

The assessment battery, requiring approximately 50 minutes to complete, was administered in classroom settings on separate days before and after treatment. To minimize the reactivity of the experiment, all students in each of three classrooms received either the experimental or control treatments; however, we decided a priori to exclude data from learning-disabled special education students mainstreamed in those classes.

The experimental and control treatments were deployed over two days in labs equipped with Pentium class computers. Each student was treated at his or her own workstation. To minimize distraction, audio portions of the intervention were delivered via earphones. Experimental participants were asked to complete the 13 modules of a cognitive restructuring program titled "Thoughts and Feelings" in a manner similar to that described in Horan (1996). Complete type scripts for that program, as well as the program itself, are available from the authors over the Internet. Control participants interacted with a computer-based geography program.

Results

Univariate ANOVAs revealed no significant pretest differences between the experimental and control treatments on any dependant variable. An ANOVA on the demand measure, however, indicated differential placebo effects favoring the experimental condition after the first session [F(1,64) = 11.195, p<.001]. Pretest correlations among the self-esteem measures were moderate (r's ranging from .62 to .75) and inversely related to the irrationality measures (r's ranging from -.29 to -.39). Table 1 summarizes all assessment data.

Treatment by grade by gender by ethnicity ANCOVAs, with pretests as covariates, were conducted on each of the five outcome measures. No violations of the equal-slopes assumption occurred; however, neither did main effects for treatment appear on any measure. Thus, the primary outcome analyses did not confirm the program's earlier success with the new sample and conditions used in this study.

Although the participants were blocked on gender, ethnicity, and grade level, it would not have been cost-beneficial to design this study with sufficient power to detect higher order interactions. Moreover, given the cell sizes and our use of multiple measures, any interactions that might occur would need to be carefully scrutinized from the standpoint of fishing and the error-rate problem (c.f., Cook & Campbell, 1979). Nevertheless, in the spirit of exploratory data analysis (Behrens & Smith, 1997), we felt it appropriate to examine interactions that involve treatment and the blocking variables.

No such interactions occurred on either of the two rationality measures. Given that three self-esteem outcome variables were employed, a Bonferroni alpha correction requires an obtained p of .0125 before an alpha of .05 can be safely claimed on any of the self-esteem measures. Only one occurred, namely, treatment and grade-level on the Janis-Field [F(1,18) = 10.856, p < .005], indicating that self-esteem benefits were evident among eighth grade but not seventh grade participants. See Table 2 for details.

Discussion

Previous research has linked specific irrational beliefs to low self-esteem (e.g., Daly & Burton, 1984; Erickson, Horan, & Hackett, 1991). A cognitive-restructuring computer program focused on changing these beliefs produced higher levels of rationality and self-esteem among normatively-low high-school juniors and seniors (Horan, 1996). Presently, that program was improved, adapted for Internet delivery, and evaluated on a younger sample with normative levels of self-esteem. Our multiply modified replication did not produce unequivocal evidence supporting the treatment; however, there is reason to suspect that at least one of the changes may be responsible for undermining the expected benefits.

This research project was designed with sufficient power to detect large effects favoring the experimental treatment with all participants regardless of gender, ethnicity, and grade level. In retrospect, however, some studies suggest that the experimental treatment might prove to be differentially effective with specific subsets of our sample. For example, Horan (1996) found that males were more favorably influenced than females. Moreover, Kovalski and Horan (1999) reported differential benefits for Caucasians relative to ethnic minorities (but this effect was more likely an artifact of socio-economic status than race). Finally, although the relationships between irrationality and certain self-esteem measures were consistent in two studies involving college students (Daley & Burton, 1983; McLennan, 1987), pattern differences occur when college students are compared to high-school juniors and seniors (Erickson, Horan, & Hackett, 1991); further variations may exist among junior-high aged populations.

Although we blocked participants on gender, ethnicity, and grade level, we were seeking to control such influences procedurally rather than to isolate possible effects of organismic variables not linked to our hypotheses. Exploratory analyses indicated that self-esteem benefits were evident among eighth grade but not seventh grade participants. Perhaps if we had restricted the age range of our sample to high school students, the positive effects observed by Horan (1996) might have endured despite our additional modifications to the format of the program and the severity of the audience's clinical problem. On the other hand it is really not clear whether age per se is the culprit or intra-session history variables commonly associated with younger audiences. Anecdotally, for example, we noted that in each seventh grade class one or more role models prematurely finished their program which may well have bred a disposition for others to take their work less seriously. In any event it would appear additional replications involving diverse populations of older high school students are clearly warranted.

References

Authorware (1995), version 2, Macromedia, Inc., 600 Townsent St., San Francisco, CA 94103.

Adobe Acrobat Reader (1998), Adobe Systems Incorporated 345 Park Avenue, San Jose, CA 95110-2704.

Behrens, J. T., & Smith, M. L. (1997). Data and data analysis. In D. Berliner & Calfee (Eds.) Handbook of Educational Research. New York: Macmillan.

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

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

Crandall, R., (1973). The measurement of self-esteem and related constructs. In J. Robinson, & P. Shaver (Eds). Measures of 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.

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

Epstein, H. (1985). [Review of the Piers-Harris Children’s Self-Concept Scale (The way I feel about myself)]. In J.V. Mitchell, Jr. (Ed), Ninth mental measurements yearbook (Vol.1, pp.1168-1169). Lincoln: University of Nebraska, Buros Institute of Mental Measurement.

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.

Fenk, A.(1994). Spatial Metaphor and Logical Pictures. In W. Schnotz & R.W. Kulhavy (Eds), Comprehension of Graphics.(pp. 43-62). Amsterdam, The Netherlands: North Holland.

Gray-Little, B., Williams, V., Hancock, T., (1997). An item response theory analysis of the Rosenberg Self-Esteem Scale. Personality and Social Psychology Bulletin, 23(5), 443-451.

Gyselinck, V. & Tardieu, H. (1994). Illustrations, Mental Models, and Comprehension of Instructional Text. In W. Schnotz & R.W. Kulhavy (Eds), Comprehension of Graphics.(pp. 139-152). Amsterdam, The Netherlands: North Holland.

Hagborg, W.J. (1996). Scores of middle-school-age students on the Rosenberg Self-Esteem Scale. Psychological Reports, 78, 1071-1074.

Hewson, C.M., Laurent, D., and Vogel, C.M., (1996). Proper methodologies for psychological and sociological studies conducted via the Internet, Behavior Research Methods, Instruments, & Computers, 28(2), 186-191.

Horan, J. J. (1996), The effects of computer-based cognitive restructuring on rationally mediated self-esteem. Journal of Counseling Psychology, 43, 371-382.

JavaScript, Sun Microsystems, Inc., 2550 Garcia Avenue Mountain View, CA 94043

Jeske, P. J. (1985), [Review of the Piers-Harris Children’s Self-Concept Scale (The way I feel about myself)]. In J.V. Mitchell, Jr. (Ed), Ninth mental measurements yearbook (Vol.1, pp.1169-1170). Lincoln: University of Nebraska, Buros Institute of Mental Measurement.

Jones, B.F., Valdez, G., Nowakowski, J., & Rasmussen, C.(1995). Plugging In: choosing and using educational technology. EdTalk [Online], pp.1-11. Available: http://www.ncrel.org.sdrs/edtalk/newtimes.htm.

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.

Lamke, L.K., Lujan, B.M., & Showalter, J.M., (1988). The case for modifying adolescents’ self-statements. Adolescence,23. 967-974.

Mayer, R.E.(1994). Visual Aids to Knowledge Construction: Building Mental Representations form Pictures and Words. In W. Schnotz & R.W. Kulhavy (Eds), Comprehension of Graphics.(pp. 125-138). Amsterdam, The Netherlands: North Holland.

McDaniel, M.A., & Waddill, P.J.(1994). The Mnemonic Benefit of Pictures in Test: Selective Enrichment for Differentially Skilled Readers. In W. Schnotz & R.W. Kulhavy (Eds), Comprehension of Graphics.(pp. 165-184). Amsterdam, The Netherlands: North Holland.

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

Netscape Communicator (1997), version 4.0, Netscape Communications Corporation, 501 East Middlefield Road, Mountain View, CA 94043

Nielsen, D.M., Horan, J.J., Keen, B., St. Peter, C.C., Ceperich, S.D., & Ostlund, D., (1996) An attempt to improve Self-esteem by modifying specific irrational beliefs. Journal of Cognitive Psychotherapy, 10, 137-149.

PowerPoint (1992), version, Microsoft Office: 1997, Microsoft, Corporation. 21919 20th Avenue Southeast, Bothell, WA 98021

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

Piers, E.V.(1969). The Way I feel About Myself: The Piers-Harris Children’s Self-concept Scale. Los Angeles: Western Psychological Services.

RealPlayer (1997), RealNetworks, Inc., 1111 Third Avenue, Suite 2900 Seattle, Washington 98101 USA

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

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

Warren, R., McLellarn, R., & Ponzohoa, C.(1988). Rational emotive therapy vs. general cognitive-behavior therapy in the treatment of low self-esteem and related emotional disturbance. Cognitive Therapy and Research, 12(1), 21-38.

Weidenmann, B.(1994). Codes of Instructional Pictures. In W. Schnotz & R.W. Kulhavy (Eds), Comprehension of Graphics.(pp. 29-42). Amsterdam, The Netherlands: North Holland.

Table 1:

Note: Higher scores on the Irrational Beliefs Test and the Rational Beliefs Scale reflect increased irrationality; higher scores on the Rosenberg Self-Esteem Scale, the Piers-Harris Children’s Self-Concept Scale, and the Janis-Field Feelings of Inadequacy Scale indicate increased self-esteem.