Explanatory Style as a Predictor of College Performance In Students with Physical Disabilities.

Attempts to modify the school environment to enhance the educational experience for students with physical disabilities have thus far been largely unsuccessful. For example, placing such students in a mainstreamed setting (Education Response Centre of the Alberta Department of Education, 1990) and the introduction of cooperative learning techniques into the curricula (Tateyama-Sniezek, 1990) have yielded inconclusive results and unimpressive success. Perhaps increased focus should be placed on more basic psychological characteristics of the individual, such as explanatory style.

Explanatory style refers to the manner in which an individual habitually explains to himself or herself the occurrence of bad events. Whether an individual perceives these occurrences in an adaptive or maladaptive manner can profoundly influence the general quality of his or her life (Seligman, 1990). More specifically, the reformulated model of learned helplessness contended that when an individual is faced with a bad or aversive situation, he or she will tend to attribute that situation to some cause (Abramson, Seligman, & Teasdale, 1978). Abramson and her colleagues stated that such causal attributions lie along three polar dimensions: 1) internal (it happened because of me) versus external (it happened because of something or someone else); 2) stable (this condition is permanent) versus unstable (this condition is temporary); and 3) global (this will interfere with everything I do) versus specific (this pertains only to this situation). The individual who habitually implements the internal-stable-global ends of the poles, or the pessimistic explanatory style, is at a greater risk for exhibiting depression-like deficits in cognition, mood, and motivation (Peterson & Seligman, 1984). Furthermore, research has shown that a pessimistic explanatory style may lead to a poorer quality of life in many different domains, which will be addressed below.

For example, more pessimistic explanatory styles have been shown to correlate with higher rates of illness (Peterson, Seligman, & Vaillant, 1988; Peterson, 1988) and impairment in immunological functioning (Kamen-Siegel, Rodin, Seligman, & Dwyer, 1991). Additionally, Seligman, and Schulman (1986) examined insurance sales agents and found that agents with more pessimistic explanatory styles had higher rates of quitting and earned less in sales commissions than did their counterparts employing an optimistic explanatory style (external-unstable-specific). Phelps and Waskel (1993) also found that a pessimistic explanatory style was associated with lower levels of job satisfaction. Another life domain that is affected by explanatory style is athletic performance. Seligman, Nolen-Hoeksema, Thornton, and Thornton, (1990) found that world-class swimmers with pessimistic explanatory styles experienced diminished performance during swims following an artificially induced failure condition; whereas their more optimistic teammates either showed no such decline or showed an improvement in time.

Finally, some research directly related to explanatory style and academic success (the topic of this article) has been conducted. Peterson and Barrett (1987) examined the explanatory style of university freshmen. These results showed that freshmen with a pessimistic explanatory style tended to achieve lower grade point averages in their first year of college. Furthermore, these results held even after controlling for the effects of ability and depression. Peterson and Barrett also found that freshmen with a pessimistic explanatory style were less likely to have specific academic goals and less likely to make use of academic advising. However, contrary to what Peterson and Barrett expected to find, explanatory style was not associated with goal efficacy (the degree to which students thought that they could achieve their academic goals).

The present study was designed to examine the role of explanatory style in the academic performance of college students with physical disabilities. Specifically, the following research hypotheses were addressed: (a) explanatory style would predict college grade point averages (GPA); (b) the prediction of GPA would hold even when the variance due to depression is statistically removed; and (c) explanatory style would predict academic goal specificity. Finally, one exploratory research question was examined: Is there an association between explanatory style and goal efficacy?

Method

Participants

Thirty-eight persons with physical disabilities (PWPD) and 32 matched persons not physically disabled (PNPD) were tested. The two groups were matched on gender and age (a criterion of 4 years). Because it became increasingly difficult to match older PWPD individuals, they were left unmatched. However, because two-tailed _t tests showed that this discrepancy did not result in any significant differences on the various demographic variables examined, such as age, years of school completed, semester credit hours completed, grade point average (GPA), and socioeconomic status (Hollingshead, 1975), the older PWPD individuals were left in the analyses (see Table 1). The sample consisted of both graduate and undergraduate students: freshmen n = 2; sophomores n = 7; juniors n = 15; seniors n = 32; and graduates n = 14. Female participants made up 61.40% of the total sample.

Table 1
Demographic Information

                                 Group
Variable
                         PWPD      PNPD      COMB

AGE              M      33.37     30.66     32.13
                SD      10.63     10.43     10.55

YRS              M       4.13      4.47      4.28
                SD       3.02      1.76      2.50

HRS              M      95.52     94.20     94.89
                SD      65.50     44.26     55.96

GPA              M       3.12      3.06      3.09
                SD        .47       .50       .48

SES              M      37.72     39.63     38.57
                SD      13.87     13.57     13.66

SAT              M     1073.75    878.31    952.76
                SD      179.84    175.84    198.31

                                     Group
Variable
                  df                   t                   p

AGE               68                 1.07                 .29

YRS               63                 -.54                 .59

HRS               57                  .09                 .93

GPA               62                  .43                 .67

SES               59                  .54                 .59

SAT        ([double dagger])   ([double dagger])   ([double dagger])
           ([double dagger])   ([double dagger])   ([double dagger])

NOTE. PWPD = persons with physical disabilities; PNPD = persons not physically disabled; COMB = combined group; YRS = years enrolled; HRS = hours completed; GPA = grade point average; SES = socioeconomic status (Hollingshead rating);

([double dagger]) = data not available.

PWPD participants were recruited from the student populations using four methods. They were recruited: (a) as they visited the Office of Disability Accommodations; (b) while attending a university orientation for new students with physical disabilities; (c) during presentations to psychology and rehabilitation classes (which also recruited PNPD participants); and (d) via a posted bulletin announcing the study. Eligible participants in either group (those enrolled in either psychology or rehabilitation classes) received two points of extra credit towards their final grade. One hundred questionnaires were distributed to potential PWPD participants in order to secure the 38 participants. Two hundred questionnaires were distributed to potential PNPD participants from which the 32 members were selected.

The disabilities represented in the present study included visual impairment (n = 10), hearing impairment (n = 7), hidden (n = 5), and motor/skeletal (n = 16). The mean duration of disabilities in the present sample was 16.99 years (SD = 12.68).

Materials

Demographic Information. A seven-item questionnaire gathered identifying information, disability status, educational background, and parental education/occupation.

Academic Goals Questionnaire. This instrument was used to assess the specificity of academic goals (Peterson & Barrett, 1987). Participants were asked to generate as many as five academic goals. Independent raters then determined the specificity of each goal from 4 (specific) to 1 (nonspecific). Goals with a concrete accomplishment (e.g., obtain a 3.0 GPA) were rated as 4. Goals without a specific outcome (e.g., better myself) were rated as 1. The goal specificity score was derived by averaging these ratings.

The participants were also asked to rate the degree to which they felt confident that they would meet each goal. These self-ratings ranged from 0 (not at all confident) to 100 (totally confident). The goal efficacy score was derived by averaging these ratings.

Academic Attributional Style Questionnaire. The AASQ was used to measure explanatory style (Peterson & Barrett, 1987). The AASQ is a 36 item questionnaire describing 12 negative school-related events. It requires approximately 20 minutes to complete under normal conditions and approximately 30 minutes if administered orally. The AASQ asks the participant to state a single cause for each negative event. Each participant is then asked to rate each cause along each of the three dimensions on a seven-point scale. The causes that are attributed to external, unstable, and specific (optimistic) factors receive low ratings (i.e., toward 1); the remaining three attributions receive high ratings (i.e., toward 7). The 36 possible responses are averaged which yields a composite explanatory style score. The higher the score, the more pessimistic it is said to be. In the present study, a mean AASQ score of 4.25 (SD = .76) was obtained. Also, an alpha internal consistency coefficient of .69 was obtained.

Revised Beck Depression Inventory. The revised Beck Depression Inventory (BDI; Beck & Steer, 1987) was used as a measure of depression. The BDI is a 21 item questionnaire intended to measure depression. Scores are said to denote severity of the depression. It requires 5 to 10 minutes to complete under normal self-administered conditions and may take up to 15 minutes if orally administered. In the present study, a mean BDI score of 8.79 (SD = 6.52) was obtained. Also, an alpha internal consistency coefficient of .84 was obtained.

The scoring of the BDI required some modification. An error in item 20 resulted in its reading "0-I am more worded about my health than usual," rather than "0-I am not more worried about my health than usual." Therefore, this item was eliminated from the scoring of the BDI and the obtained results reflected this modification. A t-test showed no significant difference with or without the omitted item: t = .34, 12 p = .74.

Procedures

Questionnaire packets were distributed to complete on a take-home basis. The materials were included in the questionnaire packet in the following order: Informed Consent form, Release of Information form (to obtain objective information concerning participants' academic performance), Demographic Information questionnaire, Academic Goals Questionnaire, AASQ, and the BDI. All participants signed an Informed Consent form. Four participants (two in the PWPD group and two in the PNPD group) did not sign a Release of Information form which did not allow the investigator access to their academic records. In these instances, the participants' self-report data (from the Demographic Information questionnaire) were used for statistical analyses.

The participants were asked to complete the questionnaire packet in a medium with which they were most comfortable. This procedure primarily affected the participants with physical disabilities. Participants with visual impairments used a variety of media: (a) three completed the packet using large print text, (b) two listened to an audio recorded version of the questionnaires, (c) one returned responses via audio cassette, and (d) one returned responses in a braille format. Others may have required the use of a reader, a scribe, or both. Because packets were completed on a take-home basis, the exact number of those using this latter method was not known.

Results

A hierarchical multiple regression analysis was used to test the relation between explanatory style (the independent variable) and college performance as determined by the participants' GPA (the dependant variable). The questions under investigation were: does a pessimistic explanatory style predict a lower GPA among the individuals tested; and, would this relation remain even after the effects of depression (as measured by the modified BDI) were held statistically constant? The results of this multiple regression equation (summarized in Table 2) indicated that for the PWPD group, there was an inverse relation (e.g., as explanatory styles grew more pessimistic, GPAs decreased) and, as such, explanatory style by itself was a valid predictor of GPA: R = .42, [R.sup.2] = .18; F(1, 32) = 6.84, p = .007. Furthermore, explanatory style was a good predictor of GPA over and above the contributions made by the modified BDI scores: t = -2.22, p = .02.

Table 2
Hierarchical Regression of ES and BDI on GPA

Variable     GPA(DV)        BDI         ES             B

BDI         -.40(1)(*)                               -.02(1)
            -.07(2)                                  -.01(2)
            -.23(3)       (.84)(3)                   -.01(3)

ES          -.42(1)        .22(1)                    -.18(1)
            -.38(2)        .28(2)                    -.32(2)
            -.40(3)(**)    .23(3)     (.42)(3)       -.23(3)

M           3.12(1)       9.76(1)     4.23(1)
            3.06(2)       7.63(2)     4.26(2)
            3.09(3)       8.79(3)     4.25(3)

SD           .47(1)       6.75(1)      .87(1)
             .50(2)       6.13(2)      .63(2)
             .48(3)       6.52(3)      .76(3)

Variable                   [Beta]            t           p

BDI                        -.32(1)       -2.00(1)     .03(1)
                            .06(2)         .34(2)     .37(2)
                           -.13(3)       -1.11(3)     .14(3)

ES                         -.35(1)       -2.22(1)     .02(1)
                           -.40(2)       -2.15(2)     .02(2)
                           -.37(3)       -3.07(3)     .002(3)

M               [R.sup.2] = .27(1)    F = 5.74(1)     .004(1)
                          = .15(2)      = 2.37(2)     .06(2)
                          = .17(3)      = 6.46(3)     .001(3)

SD         Adj. [R.sup.2] = .22(1)
                          = .09(2)
                          = .15(3)

                       R  = .39(1)
                          = .06(2)
                          = .22(3)

Note. GPA = grade point average; BDI = Beck Depression Inventory; ES = explanatory style (AASQ); subscripts indicate group values: 1 = PWPD, 2 = PNPD, 3 = combined (COMB); Adj. [R.sup.2] = Adjusted [R.sup.2]; (*) indicates significance at the p [is less than or equal to] .05 level;

(**) indicates significance at the p .01 level.

A similar pattern of results was found for the PNPD group. An inverse association existed between explanatory style and GPA. This relation suggested that a more pessimistic explanatory style was a good predictor of a lower GPA: R = .38, [R.sup.2] = .15; F(1, 28) = 4.78, p = .02. Also, explanatory style contributed significantly to the regression equation: t = -2.15, p = .02.

A hierarchical multiple regression was also used to test how well explanatory style predicted students' academic goal specificity. For the PWPD group, explanatory style was a valid predictor of goal specificity: R = .36, [R.sup.2] = .13; F(1, 36) = 5.24, p = .01. Although the association was significant, it was not in the predicted direction: [Beta] = .33, t = 2.06, p = .02. That is, those individuals in the PWPD group with a more pessimistic explanatory style tended to have their academic goal specificity rated as more specific by the independent raters.

For the PNPD group, the association between explanatory style and goal specificity was not significant: R = .15, [R.sup.2] = .02; F(1, 30) = .66, p = .22. Although this relation was not significant, it too was not in the predicted direction: [Beta] = .24, t = 1.33, p = .10. Data for these analyses can be found in Table 3.

Table 3
Hierarchical Regression of ES and BDI on GS

Variable      GS(DV)         BDI        ES             B

BDI          .19(1)                                  .01(1)
            -.27(2)                                 -.02(2)
             .00(3)      (.84)(3)                   -.00(3)

ES           .36(1)(*)    .22(1)                     .18(1)
             .15(2)       .28(2)                     .13(2)
             .293(*)      .23(3)     (.42)(3)        .17(3)

M           2.82(1)      9.76(1)     4.23(1)
            2.94(2)      7.63(2)     4.26(2)
            2.88(3)      8.79(3)     4.25(3)

SD           .48(1)      6.75(1)      .87(1)
             .34(2)      6.13(2)      .63(2)
             .43(3)      6.52(3)      .76(3)

Variable                   [Beta]              t           p

BDI                        .14(1)           .87(1)      .20(1)
                           .01(2)         -1.88(2)      .04(2)
                          -.05(3)          -.45(3)      .33(3)

ES                         .33(1)          2.06(1)      .02(1)
                           .10(2)          1.33(2)      .10(2)
                           .30(3)          2.55(3)      .007(3)

M               [R.sup.2] = .151       F = 2.98(1)      .03(1)
                          = .132         = 2.12(2)      .07(2)
                          = .093         = 3.25(3)      .02(3)

SD         Adj. [R.sup.2] = .10(1)
                          = .07(2)
                          = .06(3)

                        R = .381
                          = .362
                          = .303

Note. GS = goal specificity; BDI = Beck Depression Inventory; ES = explanatory style (AASQ); subscripts indicate group values: 1 = PWPD, 2 = PNPD, 3 = combined (COMB); Adj. [R.sup.2] = Adjusted [R.sup.2]; (*) indicates significance at the p [is less than or equal to] .05 level.

Similar analyses were used to test the relation between explanatory style and goal efficacy. The relation between

explanatory style and goal efficacy was significant: R = .62, [R.sup.2] = .37; F(2, 35) = 10.44, p = .0002, even after holding the variance due to depression constant. Explanatory style contributed significantly to the relation: [Beta] = -.50, t = -3.68, p = .0004. This relation was in the negative direction; that is, individuals with higher pessimistic explanatory style scores demonstrated lower goal efficacy scores. No such relation existed for the PNPD group. Data for these analyses are in table 4.

To further investigate GPA with regards to explanatory style, an exploratory one-way ANOVA was used, splitting the sample at the median explanatory style score (4.08). The group with the more pessimistic explanatory style (n = 33) had a mean GPA of 2.93, SD = .46; the group with the more optimistic explanatory style (n = 31) had a mean GPA of 3.26, SD = .45. This difference was statistically significant, F(1, 62) 8.15, p = .006. Finally, the results of a two-tale t-test showed that there was no statistically significant difference between the explanatory style scores of PWPD and PNPD individuals: t = .08, p = ..94.

Discussion

The primary research hypothesis of the present study was upheld. A pessimistic explanatory style was negatively associated with college GPA; the more pessimistic the student, the lower his or her grades tended to be. Additionally, this relationship remained when the initial effects of depression were controlled. These results, together with Peterson and Barrett's (1987) examination of college freshmen, support the conclusion that pessimism is an academic risk factor.

The most important aspect in which the present study differed from Peterson and Barrett's (1987) study was the inclusion of the group of students with physical disabilities. Not only did this difference allow for the direct examination of students with physical disabilities, but also it allowed for the comparison of that group to the sample of individuals without physical disabilities. The results of the present study showed that the relation between explanatory style and GPA existed for the overall sample, regardless of the presence or absence of a physical disability. This similarity has been forwarded by authors such as Vash (1981), Beaty, (1994), and Martinez and Sewell (1996). Thus, with respect to explanatory style as well as other factors studied, college students with and without physical disabilities appear more alike than different.

The investigation of academic goals yielded interesting results. The results of the present study revealed, for only the PWPD group, a significant relationship between explanatory style and goal specificity. Furthermore, this relationship was not in the predicted direction. Specifically, those individuals with physical disabilities who implemented a more pessimistic explanatory style tended to have goals that were rated as more specific. These results were contrary to what Peterson and Barrett (1987) found. It is unclear why these studies yielded contradictory results. Perhaps the presence of environmental and societal obstacles facing individuals with physical disabilities might have contributed to the results in the present study. Rather than goal specificity being a positive trait that connotes focus and direction, it might be more of a hindrance to those with disabilities who make use of a pessimistic explanatory style. It is possible that they view their environment in a more rigid and restrictive way. Along this line, Vash (1981) reported that those individuals with physical disabilities who hold a unifocal goal will be more profoundly influenced by the presence of a disability. Therefore, perhaps high goal specificity in the present study is analogous to unifocal goals among the PWPD group.

The measure of goal efficacy also yielded results that merit attention because they differed from previous findings (Peterson & Barrett, 1987). There was a significant negative relation between explanatory style and goal efficacy for both groups. That is, the more pessimistic explanatory style an individual held, the less likely he or she felt that his or her goals could be accomplished. Given the characteristics of learned helplessness, Peterson and Barrett had originally expected this negative relation between explanatory style and goal efficacy; but no such result emerged. Again, these differences in results beg explanation. First of all, the sample in the present study was much older than Peterson and Barrett's (1987) sample. Thus age, rather than disability status, might be contributing to the present results. Second, as with goal specificity, perhaps the individuals in the present study view their ability to attain their stated goals as being more closely tied to societal barricades. Therefore, they might perceive these societal barricades as significant hindrances to the accomplishment of their academic goals. Vash (1981) stated that individuals with physical disabilities may tend to experience powerlessness when attempting to navigate through societal barriers. She added that this powerlessness can result in learned helplessness. Eventually, as hypothesized in the present study, the sense of efficacy for the individuals with more pessimistic explanatory style scores may have been diminished. These dynamics might be extended to the PNPD participants because as Healey (1993) reported, women with disabilities and elderly women share many similar characteristics. Perhaps the age of the present sample is sufficiently high to evoke the use of such reasoning.

The results of the present study have important implications for individuals with physical disabilities. Primarily, the results should inform both professors and students that they need not assume the presence of a physical disability will make classroom requirements insurmountable. This implication is particularly important in light of previous findings (Fichten, Goodricck, Tagalakis, Amsel, & Libman, 1990). Fichten et al. reported that both professors and students with physical disabilities tended to place more attention on nonacademic aspects of education than necessary. These types of dynamics can lead to a less productive educational experience for all parties involved (Stewart, 1990). However, should it be determined that a student possesses a pessimistic explanatory style, referral to a university counseling center for cognitive remediation could be made. For persons with disabilities, such an assessment can be readily made during initial intake procedures at the university's office that handles accommodations for students with physical disabilities. For non-disabled students, identification of pessimism presents a screening problem commensurate to that in the detection of most other subtle learning problems.

Furthermore, cognitive remediation of pessimistic explanatory style should be coupled with the direct teaching of concrete behaviors that can help students improve their academic performance (Kunnen, 1993). In this instance, the students receiving cognitive remediation can also be encouraged to attend university classes that teach behaviors such as time management and study skills. These recommendations could provide college students-explicitly including those with physical disabilities-with a college experience that is maximally positive.

References

Abramson, L. Y., Seligman, M. E. P., & Teasdale, J. D. (1978). Learned helplessness in humans: Critique and reformulation. Journal of Abnormal Psychology, 87, 49-74.

Beaty, L. A. (1994). Psychological factors and academic success of visually impaired college students. Psychological Review, 26. 131-139.

Beck, A. T. & Steer, R. A. (1987). Beck Depression Inventory. San Antonio, San Diego, New York, Chicago, Toronto: Psychological Corporation.

Education Response Centre, Alberta Education (1990). Integrating exceptional students into the mainstream: Literature review. ERIC Document 348822 (pp. 1-12).

Fichten, C. S., Goodrick, G., Tagalakis, V., Amsel, R., & Libman, E. (1990). Getting along in college: Recommendations for college students with disabilities and their professors. Rehabilitation Counseling Bulletin, 34(2), 103-125.

Hollingshead, A. B. (1975). Four factor index of social status. Unpublished manuscript, Yale University at New Haven.

Kamen-Siegel, L., Rodin, J., Seligman, M. E. P., & Dwyer, J. (1991). Explanatory style and cell-mediated immunity in elderly men and women. Health Psychology, 10, 229-235.

Kunnen, S. (1993). Attributions and perceived control over school failure in handicapped and non-handicapped children. International Journal of Behavioral Development, 16, 113-125.

Martinez, R., & Sewell, K. W. (1996). Self-concept of adults with visual impairments. Journal of Rehabilitation, 62, 55-58.

Peterson, C. (1988). Explanatory style as a risk factor for illness. Cognitive Therapy and Research, 12, 119-132.

Peterson, C., & Barrett, L. C. (1987). Explanatory style and academic performance among university freshmen. Journal of Personality and Social Psychology, 53, 603-607.

Peterson, C., & Seligman, M. E. P. (1984). Causal explanations as a risk factor for depression: Theory and evidence. Psychological Review, 91,347-374.

Peterson, C., Seligman, M. E. P., & Vaillant, G. E. (1988). Pessimistic explanatory style is a risk factor for physical illness: A Thirty-Five-year longitudinal study. Journal of Personality and Social Psychology, 55, 23-27.

Phelps, L., & Waskel, S. A. (1993). Work reinforcers and explanatory style for women age 40 to 75 years. Journal of Psychology, 128(4), 403-407.

Seligman, M. E. P. (1990). Learned Optimism. New York: Pocket Books.

Seligman, M. E. P., Nolen-Hoeksema, S., Thornton, N., & Thornton, K. M. (1990). Explanatory style as a mechanism of disappointing athletic performance. Psychological Science, 1,143-146.

Seligman, M. E. P., & Schulman, P. (1986). Explanatory style as a predictor of productivity and quitting among life insurance sales agents. Journal of Personality and Social Psychology, 50, 832-838.

Stewart, C. C. (1990). Effect of practica types in preservice adapted-physical education curriculum on attitudes towards disabled populations. Journal of Teaching in Physical Education, 10, 76-83.

Tateyama-Sniezek, K. M. (1990). Cooperative learning: Does it improve the academic achievement of students with handicaps? Exceptional Children, 56, 426-437.

Vash, C. (1981). Psychology of disability: Springer series on rehabilitation, Vol. 14. The New York: Springer Publishing Company.

Kenneth W. Sewell, Ph.D., Department of Psychology, Box 311280, University of North Texas, Denton, Texas 76203-1280.