Predicting GPA from Explanatory Style Viliyana Maleva, Kathryn Westcott, Mark Mckellop*, Ronald Mclaughlin, and David Widman Juniata College

Optimism and College Grades: Predicting GPA From Explanatory Style Viliyana Maleva, Kathryn Westcott, Mark McKellop*, Ronald McLaughlin, and David Widman Juniata College

ABSTRACT. The current study attempted to predict grade point average (GPA) based on academic explanatory style in college students. Building our work on Barrett and Peterson’s study (1987), we hypothesized that college students with optimistic explanatory style would have significantly higher GPAs than college students with pessimistic explanatory style. We tested this hypothesis with 171 undergraduate students at a small liberal arts college by using the Academic Attributional Style Questionnaire (AASQ; Barrett & Peterson, 1987) and found a significant, but small, r = -.15, p = .44, negative correlation between academic explanatory style and GPA. Although our results pointed to the conclusion that college students with optimistic explanatory style have higher GPAs than college students with pessimistic explanatory style, explanatory style seems to have a weak predictive value for college GPA. Further research needs to replicate these findings and examine their practical significance. Alternative explanations and future research directions are discussed.

xplanatory style is a personality variable consistently offering global attributions believe referring to the habitual way in which that the negative event will impact other life areas, Eindividuals explain to themselves the cause whereas individuals making specific attributions of the negative events that befall them (Maier think that this event will be circumscribed. A & Seligman, 1976). It spans three dimensions: student who drops out of college may make a global internality, stability, and globality (Maier & attribution and believe that the negative event will Seligman, 1976). Individuals who typically make negatively impact his or her relationships, or make internal attributions ascribe the cause of the a specific attribution and think that the failure will negative event to an internal factor. In contrast, not influence other life areas negatively (Maier & individuals offering external attributions ascribe Seligman, 1976). the cause to an external factor. A student failing an The choice of dimensions of explanatory exam may make an internal attribution such as “I style was theoretically driven and reflected the am stupid” or an external attribution such as “The researchers’ questions of interest (Peterson, 1991). exam was too hard.” Individuals who consistently Explanatory style originated from the reformulated make stable attributions believe that the cause learned helplessness model and sought to explain of the negative event will last forever, whereas why individuals responded in different ways to individuals offering unstable attributions expect major, negative life events. Individuals who typically this cause to be transient. A student who is on offered internal, stable, and global attributions for academic probation may make a stable attribution negative events were more likely to behave passively FALL 2014 and believe that the factor responsible for this event in the face of challenge, were poor at problem-solv- will last forever or make an unstable attribution and ing, and eventually gave up. In contrast, individuals PSI CHI JOURNAL OF regard this cause as temporary. Finally, individuals who offered external, unstable, and specific causal PSYCHOLOGICAL RESEARCH

*Faculty mentor COPYRIGHT 2014 BY PSI CHI, THE INTERNATIONAL HONOR SOCIETY IN PSYCHOLOGY (VOL. 19, NO. 3/ISSN 2164-8204) 129 College GPA and Explanatory Style | Maleva, Westcott, McKellop, McLaughlin, and Widman

attributions to negative events were more likely to style, and between depression and GPA. persist after failure and keep trying to excel (Maier Although pessimistic explanatory style has & Seligman, 1976). been linked to low achievement performance Rather than a typology, explanatory style can in an academic context, there may be multiple be represented on a continuum between pessimis- definitions of academic performance. Thus, it tic and optimistic explanatory style. Individuals seems reasonable to focus on a single aspect of should not be categorized as pessimistic or optimistic. academic performance such as GPA. Examining Instead, one should regard them as more likely the relationship between explanatory style and to have an optimistic or pessimistic explanatory GPA is important because the former may emerge style. Individuals with pessimistic explanatory style as an unexpected predictor of and have important habitually make internal, stable, and global attri- implications for the latter (Peterson & Barrett, butions for bad events. These individuals expect 1987). Explanatory style may have a direct bearing negative events to happen consistently and believe on students’ approach toward studying. Students in that they cannot exert control over these events. any course typically encounter a number of difficul- Individuals with pessimistic explanatory style are ties throughout the semester, such as failing a quiz, more likely to become helpless and passive in the coming unprepared to class, or not being able to face of obstacles and exhibit the “whatever will be solve a set of homework problems. It is important will be” mentality. On the other hand, individuals how students explain to themselves the causes with optimistic explanatory style consistently offer of these events because this may affect their aca- external, unstable, and specific causal attributions demic performance. for bad events. Unlike individuals with pessimistic Individuals believe that internal factors are less explanatory style, individuals with optimistic likely to change than external factors (Metalsky et explanatory style typically possess a wider repertoire al., 1982). Thus, if a student thinks “I’m stupid,” the of responses to challenge. Instead of giving up in student is likely to give up and believe intelligence the face of challenge, optimists typically renew their is not likely to change. External circumstances, on efforts (Maier & Seligman, 1976). the other hand, are perceived as more mutable. If Many studies have documented the relation a student assigns the blame to an external factor, between pessimistic explanatory style and depres- he or she is likely to keep trying to excel. Thus, stu- sion, and have consistently found that individuals dents with optimistic explanatory style are expected offering internal, stable, and global causal attribu- to persist in the face of challenge and renew tions for negative events were at a greater risk their efforts after a negative academic outcome for depression (Metalsky, Abramson, Seligman, (Peterson & Barrett, 1987). Students who eventu- Semmel, & Peterson, 1982; Miller & Seligman, ally achieve a higher GPA are perhaps less likely 1975). Fazio and Palm (1998) found a significant to behave passively because they offer external correlation between depression and explana- attributions (“The professor gave a very hard tory style, and between depression and GPA. exam”), unstable attributions (“It will be short- Besides depression, a large body of research has lived”), and specific attributions (“It will not supported the relation between explanatory affect the other areas of my life”). On the con- style and achievement in an academic context trary, students who end up with a low GPA may (Metalsky et al., 1982; Peterson & Barrett, 1987; be more likely to behave passively because they Nolen-Hoeksema, Girgus, & Seligman, 1986). use internal (“I’m stupid”), stable (“It’s going Specifically, Nolen-Hoeksema et al. (1986) to last forever”), and global (“It will undermine demonstrated that, compared to schoolchildren everything I do”) causal attributions for their with optimistic explanatory style, schoolchildren academic failures. These students tend to be pas- with pessimistic explanatory style showed higher sive when they encounter obstacles and eventually depression levels and were more likely to experi- give up (Peterson & Barrett, 1987). Yet, it must ence achievement problems. Similarly, Metalsky et be noted that the relationship between attribu- al. (1982) found that the more internal and global tional style and GPA seems to be moderated by attributions university first-year students offered for different variables such as students’ ability level FALL 2014 negative events, the more severely depressed they (Gibb, Zhu, Alloy, & Abramson, 2002; Houston, became after the receipt of a low midterm-exam 1994), major (Satterfield, Monahan, & Seligman, PSI CHI JOURNAL OF grade. Fazio and Palm (1998) found a significant 1997), and time of the semester (Yee, Pierce, PSYCHOLOGICAL correlation between depression and explanatory Ptacek, & Modzelesky, 2003). RESEARCH

Peterson & Barrett (1987) examined the Questionnaire (ASQ; Cooligan, Offord, Malinchoc, relation between explanatory style and academic Schulman & Seligman, 1994). The ASQ asked performance in 87 university first-year students participants about various life events, mainly and found that first-year students with pessimistic interpersonal situations (Cooligan et al., 1994), explanatory style received lower grades during their whereas the AASQ presented participants only first year and were more likely to show passivity in with academic situations (Peterson & Barrett, the face of challenge. These findings held true 1987). Because the nonacademic items were after the researchers controlled for confounding irrelevant to academic outcomes, the AASQ variables including initial depression, measured included 12 hypothetical negative academic by the Beck Depression Inventory (Beck & Steer, situations such as “You fail a final examination” 1984), and aptitude test scores, measured by the (Peterson & Barrett, 1987). Scholastic Aptitude Test (Peterson & Barrett, 1987). In the original study, participants wrote short Building our work on Peterson & Barrett’s (1987) answers to the academic situations, after which study, we hypothesized that students with optimistic trained evaluators rated participants’ answers on explanatory style who offered external, unstable, internality, stability, and globality (Peterson & and specific attributions to negative academic Barrett, 1987). We altered the way the scale was outcomes would have a higher GPA than students administered because (a) our method was much with pessimistic explanatory style who offered less time-consuming, (b) participants provided internal, stable, and global attributions to negative their answers in a quantitative rather than a academic situations. qualitative way, and (c) evaluator bias perhaps decreased. Participants rated the cause for each of Methods the 12 academic events on the three dimensions Participants of explanatory style: internality (“It’s me”) versus Participants were 235 college students from a small externality (“It’s somebody else”), stability (“It will liberal arts college and were at least 18 years old. last forever”) versus instability (“It will be short- Missing values necessitated the exclusion of 64 lived”), and globality (“It will affect everything I participants, leaving 171 participants for the final do”) versus specificity (“It won’t affect anything I analyses. The major reason for this large exclu- do”). The rating used a 7-point Likert scale, with sion was that participants received class credit for low scores indicating optimistic explanatory style providing informed consent in the survey. As they and high scores indicating pessimistic explanatory could not be required to complete the whole survey style. A sample survey item is the Appendix. For because of ethical considerations, many of them each participant, we obtained a composite score answered only several questions and left the rest ranging from 1 to 7 by averaging participants’ blank. The mean GPA was 3.16 (SD = 0.51) with scores across the dimensions of explanatory style a range of 1.72 to 4.00. There were 123 women and across the 12 negative academic situations (72%) and 48 men (28%). The majority of partici- (Peterson & Barrett, 1987). pants were 88 first-year students (51%), followed by 30 sophomores (18%), 21 juniors (12%), and 32 Procedure seniors (19%). Participants included both domestic We obtained institutional review board approval (n = 158) and international students (n = 13). prior to beginning the study. All participants signed an electronic informed consent before taking the Measures survey. We employed several contact methods to Demographic questionnaire. Participants recruit participants. First, college students received reported demographic information including sex, an e-mail with a link to the survey. Second, intro- age, year in college, overall GPA, major, and duction to psychology students received credit whether they were a transfer or an international for completing the survey. Third, students were student. We did not collect ethnicity information availed of the survey through a link in the college because we believed that it was irrelevant to our announcement e-mail. All participants completed major hypotheses. the survey electronically via LimeSurvey®. Explanatory style. Participants com- FALL 2014 pleted the Academic Attributional Style Ques- Results PSI CHI tionnaire (AASQ; Peterson & Barrett, 1987), Our main hypothesis was supported. There was JOURNAL OF which is patterned after the Attributional Style a significant correlation between explanatory PSYCHOLOGICAL RESEARCH

style and GPA (see Figure 1). However, this cor- was a surprising finding because students with pes- relation was relatively small, r = -.15, p = .044. simistic explanatory style are expected to attribute The correlation between internality (M = 4.48, the cause of the negative event to internal factors SD = 0.92) and GPA did not reach statistical such as their character and abilities, as opposed to significance, r = .008, p = .92. Stability (M = 3.55, external factors such as other people or circum- SD = 0.88), r = -.16, p = .04, and globality (M = 4.611, stances (Peterson & Barrett, 1987). Yet, our results SD = 0.74), r = -.21, p = .07, each correlated with did not support this assumption. Internality might GPA significantly. Regression analyses indicated have failed to reach statistical significance because that explanatory style and stability respectively students who offer internal attributions for nega- accounted for 2.4% and 2.5% of the variance in tive academic events may take more responsibility, GPA. Finally, a multiple regression was run using renew their efforts, and so achieve a higher GPA stability and globality as predictors and GPA as or because students may believe that internal the criterion variable. We found a nonsignificant factors and external factors are equally likely to regression coefficient for stability, but a significant change. Yet, these explanations violate a main regression coefficient for globality,r = -.14, p = .027. assumption of learned helplessness theory (Maier Globality predicted 5.3% of the variance in GPA. & Seligman, 1976), and if they were true, we should have found a significant positive correlation Discussion between internality and GPA. Our main hypothesis was supported. There was a Another interesting finding was that stability significant correlation between explanatory style was no longer significant when we simultaneously and GPA. Students with optimistic explanatory style predicted GPA based on stability and globality. tended to have a higher GPA than students with Stability might have been nonsignificant in the pessimistic explanatory style. There seemed to be a multiple regression because stability and globality significant, but small, relation between the habitual were significantly correlated. Stability and globality way in which students explained to themselves their correlate highly, perhaps because a factor has to last negative academic outcomes and their GPA. This long enough in order to influence many life areas is perhaps because there are multiple other factors (Peterson, 1991). influencing one’s GPA such as parents’ level of Our results were consistent with Peterson & education, socioeconomic status, IQ, personality Barrett’s results (1987). We found a significant traits, and abilities. correlation between explanatory style and GPA. Two explanatory style dimensions, stability and However, there were several important differ- globality, each correlated with GPA significantly. ences between the present study and Peterson & Stability, the extent to which individuals believe the Barrett’s (1987) study. First, the correlation factor causing the negative outcome will persist, between explanatory style and GPA was smaller significantly predicted GPA. This might be because in the present study. Second, unlike Peterson and students who believe that the factors causing Barrett, we broke down explanatory style by its academic negative events will persist become dimensions and examined their individual cor- temporarily depressed and withdraw their efforts relation with GPA. Third, our sample was larger (Peterson & Barrett, 1987). It might also be that (171 participants as opposed to 87) and slightly students’ pessimistic explanatory style negatively more diverse as it included international students influences important life areas such as romantic and students from all four years in college. Fourth, relationships and friendships, which, in turn, Peterson and Barrett (1987) used measures that we negatively impacts GPA. Out of the three dimen- did not administer: a questionnaire for the specific- sions, globality had the largest predictive value for ity of students’ goals, a self-efficacy questionnaire, GPA. It seems that students believing that a failing a questionnaire about students’ typical coping grade would ruin other life areas become hopeless, response to academic failure, and the degree to withdraw their efforts and end up with a low GPA. which students made use of advising throughout Although globality seemed to be the best predic- the semester. tor of GPA out of the three dimensions, further There were a number of limitations to the FALL 2014 research needs to replicate our findings. current study. First, we only analyzed participants’ However, the correlation between the third self-ratings. Second, we altered the format of the PSI CHI JOURNAL OF dimension of explanatory style, internality, and AASQ because (a) reading such a large number of PSYCHOLOGICAL GPA did not reach statistical significance. This short answers was considerably time-consuming, RESEARCH

(b) participants provided their responses quanti- variable. Second, explanatory style about academic tatively rather than qualitatively, and (c) evaluator events may be more strongly expressed in upper bias perhaps diminished. Third, the change of the division than in first-year students. Students in their scale might have compromised its internal validity first year may not have fully formed their academic because (a) it limited participants’ responses to a explanatory style yet, whereas upperclass students 7-point Likert scale instead of allowing participants may be more familiar with the negative academic to express subtler nuances of their responses, events in college. Third, individuals may offer (b) social demand characteristics might have been “spontaneous” attributions to hypothetical events active, and (c) participants might have been simply and do not necessarily think of the cause of the confused by the new wording of the situations. negative events that befall them unless explicitly Second, the external validity of our results was asked (Peterson, 1991). Students might have rumi- limited because our sample was not representative nated on the cause of the negative academic events and because of a potential restriction of range they encountered, but they might have found it issue (e.g., high average GPA). Third, social desir- difficult to determine the cause of hypothetical, ability characteristics might have been active. It imagined negative academic events. Fourth, there was not possible to determine whether participants may be cultural differences in explanatory style reported their GPAs accurately because we did (Peterson, 1991), and the international students not verify their GPAs with the college’s registrar. in our sample might not have conformed to the Participants might have also given more socially typical pattern of explanatory style attributions. desirable responses on the AASQ. Fourth, because All of these variables could have acted as potential the present study used a correlational rather than confounds and might have suppressed the correla- an experimental design, the direction of causality is tion between explanatory style and GPA. Partialing not clear. Explanatory style might have influenced out their variance could have revealed a greater GPA, GPA might have influenced explanatory style, correlation between explanatory style and GPA. or a third variable might have been responsible for It seems that the relationship between explana- both explanatory style and GPA. tory style and GPA is more complex. Specifically, Fifth, though many of our results were Gibb et al. (2002) found that students who had a statistically significant, they were not necessarily lower ability level (indicated by lower SAT scores) practically significant. Although the typical correla- and a pessimistic explanatory style (consisting of tion with explanatory style is in the range of .20 to internal and stable attributions) had a lower GPA .30 (Peterson, 1991), the size of our correlations compared to students who had a high ability level was relatively small. Globality, the best predictor (indicated by high SAT scores), suggesting that the out of the three dimensions, explained only 5%of correlation between attributional style and GPA the variance in GPA, which means that 95%of the is moderated by students’ ability level. Interest- variance in GPA was explained by factors other ingly, globality was not at all related to academic than explanatory style. Alternative explanations performance (Gibb et al., 2002). The expectation could include participants’ cultural background, IQ, SAT scores, personal traits and abilities (e.g., FIGURE 1 persistence, motivation, time-management skills, goal-directedness), depression, major, sex, age, Scatterplot for Explanatory Style and GPA parental education, inaccuracy in reporting GPA, social demand characteristics, and others. GPA versus Explanatory, Internality, Stability, Globality Explanatory Internality Sixth, it is not clear whether academic 4.0 explanatory style is a legitimate construct by itself. 3.5 3.0 Explanatory style in general may be a more strongly 2.5 established psychological construct and may have 2.0 a greater correlation with GPA than academic GPA 2 3 4 5 6 2 3 4 5 6 Stability Globality explanatory style. Academic attributional style is 4.0 3.5 only an example of attributional style. It is restricted 3.0 to academic situations only and may thus have a 2.5 FALL 2014 more limited general validity. 2.0 2 4 6 3 4 5 6 7 PSI CHI Our sample included many more women than JOURNAL OF men, and sex might have been a confounding PSYCHOLOGICAL RESEARCH

that pessimistic attributional style is related to low investigate the role of explanatory style at differ- academic performance and optimistic attributional ent school levels (e.g., elementary school, middle style is related to high academic performance was school, high school, college, and graduate school). violated by other studies as well. Houston (1994) One has to examine the development of academic found that students who perform well make stable explanatory style over time through longitudinal and, to some extent, global attributions. Satterfield studies. Future researchers could study academic et al. (1997) demonstrated that law students with explanatory style for good events because it is pessimistic explanatory style (consisting of stable, important to understand how students explain internal, and global attributions for negative their academic success to themselves. Last but not events) had a higher GPA than law students with least, future researchers could refine the existing optimistic explanatory style. Further, Bridges academic explanatory style scale or develop a new (2001) indicated that SAT, but not attributional one that serves solely the purposes of academic style, was significantly related to later performance explanatory style. on course-related objective exams in college. Finally, in a study by Yee et al. (2003), pessimistic References Beck, A. T., & Steer, R. A. (1984). Internal consistencies of the original and explanatory style was not related to academic revised Beck Depression Inventory. Journal of Clinical Psychology, performance in the first part of the semester, but 40, 1365–1367. doi:10.1002/1097-4679(198411)40:6<1365::AID- was positively correlated with it during the second JCLP2270400615>3.0.CO;2-D Bridges, K. R. (2001). Using attributional style to predict academic performance: half of the semester. All of these findings that dif- How does it compare to traditional methods? Personality and Individual fer from the general attributional style hypothesis Differences, 31, 723–730. doi:10.1016/S0191-8869(00)00174-4 underscore the importance of further examining Cooligan, R. C., Offord, K. P., Malinchoc, M., Schulman, P., & Seligman, M. E. (1994). CAVEing the MMPI for an optimism-pessimism scale: the role of potential confounding variables in the Seligman's attributional model and the assessment of explanatory relationship between explanatory style and GPA. style. Journal of Clinical Psychology, 50, 71–75. doi:10.1002/1097- The numerous limitations of the present 4679(199401)50:1<71::aid-jclp2270500107>3.0.co;2-8 Fazio, N. M., & Palm, L. J. (1998). Attributional style, depression, and grade study point to a number of future research direc- point averages in college students. Psychological Reports, 83, 159–162. tions. Peterson and Barrett (1987) suggested that doi:10.2466/PR0.83.5.159-162 prevention programs be designed in order to Gibb, B. E., Zhu, L., Alloy, L. B., & Abramson, L. Y. (2002). Attributional styles and academic achievement in university students: A target students in need of academic attribution longitudinal investigation. Cognitive Therapy and Research, 26, therapy. Wilson and Linville (1982, 1985, as cited in 309–315. doi:10.1023/A:1016072810255 Peterson & Barrett, 1987) found that interventions Houston, D. M. (1994). Gloomy but smarter: The academic consequences of attributional style. British Journal of Social Psychology, 33, teaching students how to make more transient attri- 433–441. doi:10.1111/j.2044-8309.1994.tb01039.x butions helped students become more persistent Maier, S. F., & Seligman, M. E. (1976). Learned helplessness: Theory and and improved their grades in college. We suggest evidence. Journal of Experimental Psychology: General, 105, 3–46. doi:10.1037/0096-3445.105.1.3 that colleges design prevention and intervention Metalsky, G. I., Abramson, L. Y., Seligman, M. E., Semmel, A., & Peterson, C. programs in order to target students with highly (1982). Attributional styles and life events in the classroom: Vulnerability pessimistic explanatory style who are doing poorly and invulnerability to depressive mood reactions. Journal of Personality and Social Psychology, 43, 612–617. doi:10.1037//0022-3514.43.3.612 academically. The programs’ sections on explana- Miller, W. R., & Seligman, M. E. (1975). Depression and learned tory style should teach students that a low grade helplessness in man. Journal of Abnormal Psychology, 84, 228–238. does not necessarily affect the other areas of their doi:10.1037/h0076720 Nolen-Hoeksema, S., Girgus, J. S., & Seligman, M. E. (1986). Learned lives in a negative and long-lasting way. However, helplessness in children: A longitudinal study of depression, achievement, because the correlation between explanatory style and explanatory style. Journal of Personality and Social Psychology, 51, and GPA is relatively small, the former should be a 435–442. doi:10.1037/0022-3514.51.2.435 Peterson, C. (1991). The meaning and measurement of explanatory style. supplementary part of such programs rather than Psychological Inquiry, 2, 1–10. doi:10.1207/s15327965pli0201_1 their primary topic. Peterson, C., & Barrett, L. C. (1987). Explanatory style and academic Future researchers could examine the causal performance among university freshmen. Journal of Personality and Social Psychology, 53, 603–607. doi:10.1037//0022-3514.53.3.603 relation between explanatory style and GPA. Does Satterfield, J. M., Monahan, J., & Seligman, M. E. P. (1997). Law school the former cause the latter or vice versa, or does performance predicted by explanatory style. Behavioral Sciences and the a third variable cause both? In order to increase Law, 15, 95–105. doi:10.1002/(SICI)1099-0798(199724)15:1<95::AID- BSL255>3.0.CO;2-Y external validity, one could diversify the sample by Yee, P. L., Pierce, G. R., Ptacek, J. T., & Modzelesky, K. L. (2003). Learned FALL 2014 recruiting more participants from various cultural helplessness attributional style and examination performance: backgrounds, by recruiting more participants with Enhancement effects are not necessarily moderated by prior failure. Anxiety, PSI CHI Stress, and Coping, 16, 359–373. doi:10.1080/0003379031000140928 JOURNAL OF a low GPA, and by randomly selecting participants PSYCHOLOGICAL from the student body. Future researchers could Author Note. Viliyana Maleva, Kathryn Westcott, Mark McKellop, RESEARCH

Ronald McLaughlin, and David Widman, Department of Psychology, Juniata College, PA. APPENDIX Viliyana Maleva is now at Department of Psychology at Sample Survey Question Measuring Explanatory Hofstra University, NY. Style in Negative Academic Situations Correspondence concerning this article should be directed to Viliyana Maleva, Department of Psychology at Hofstra University, You fail a final examination. Hempstead, New York, 11549. E-mail: [email protected] It's somebody/something else It's me 1 2 3 4 5 6 7 It will be short-lived It will last forever 1 2 3 4 5 6 7 It won't affect what I do It will affect everything I do 1 2 3 4 5 6 7

FALL 2014

PSI CHI JOURNAL OF PSYCHOLOGICAL RESEARCH