This article was downloaded by: [University of Toronto Libraries] On: 02 February 2015, At: 09:22 Publisher: Routledge Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

The Journal of Psychology: Interdisciplinary and Applied Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/vjrl20

A Further Test of the InvertedU Hypothesis Relating Achievement Anxiety and Academic Test Performance a

a

David C. Munz , Christine T. Costello & Karen Korabik

a

a

Saint Louis University , USA Published online: 02 Jul 2010.

To cite this article: David C. Munz , Christine T. Costello & Karen Korabik (1975) A Further Test of the Inverted-U Hypothesis Relating Achievement Anxiety and Academic Test Performance, The Journal of Psychology: Interdisciplinary and Applied, 89:1, 39-47, DOI: 10.1080/00223980.1975.9923904 To link to this article: http://dx.doi.org/10.1080/00223980.1975.9923904

PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or

indirectly in connection with, in relation to or arising out of the use of the Content.

Downloaded by [University of Toronto Libraries] at 09:22 02 February 2015

This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://www.tandfonline.com/page/terms-and-conditions

Published as a separate and in The Journal of Psychology, 1975, 89, 39-47

A FURTHER TEST O F THE INVERTED-U HYPOTHESIS RELATING ACHIEVEMENT ANXIETY AND ACADEMIC TEST PERFORMANCE* Saint Louis University

Downloaded by [University of Toronto Libraries] at 09:22 02 February 2015

DAVID c . MUNZ,CHRISTINE T. COSTELLO,

AND KAREN

KORABIK

SUMMARY The assumption that the inverted-U hypothesis, which shows performance as a function of activation level, mediates the relationship between achievement anxiety and academic test performance was tested by comparing Achievement Anxiety Test scores of 75 male and female college students with a self-report measure of activation taken prior to a classroom examination. Results supported the predicted relationship between achievement anxiety reaction type and academic performance (p < .05), but only partially supported the inverted-U hypothesis posited to account for this relationship. Results were further interpreted as suggesting that examinees experience two general types of arousal in the testing situation-ne type that enhances performance and one that impedes performance. Further implications of the results were discussed. A. INTRODUCTION The relationship between achievement anxiety, as operationalized by Alpert and Haber’s (1) Achievement Anxiety Test (AAT), and academic performance has been convincingly demonstrated (e. g . , 2, 4, 11, 20). Because Alpert and Haber conceived of test-anxiety as a two-dimensional construct-an anxiety process that debilitates test performance and a separate anxiety process that facilitates performance-they developed a selfreport inventory to measure both types of anxiety. The AAT consists of a Facilitation (AAT+) scale and a Debilitation (AAT-) scale. Scores on the AAT+ scale have been found to correlate positively with academic performance, while scores on the AAT - scale correlate negatively. Moreover, the legitimacy of a two dimensional approach to test anxiety has been

* Received in the Editorial Office on September 16, 1974, and published immediately at Provincetown, Massachusetts. Copyright by The Journal Press. 39

Downloaded by [University of Toronto Libraries] at 09:22 02 February 2015

40

JOURNAL O F PSYCHOLOGY

established, since the predictive efficiency of the AAT is significantly increased by using the two scales in combination to predict academic performance (1, 4). While some researchers have concentrated their efforts on delineatiiig the structural relationships of the achievement anxiety construct (3, 10, 18, 19), others have studied the interaction effects on test performance of this construct and such situational factors as instructor achievement cues (6, 7 ) , item-difficulty arrangement format (8, 9, 131, and instructional set (12). However, no attempt has been made to investigate in the testing environment specific organismic reactions that may mediate or actually account for the relationship between AAT scores and achievement performance. The fact that academic achievement can be predicted from a self-report inventory that presumes to assess an individual’s awareness of how he has generally reacted in the past to the test-taking situation suggests that the mediating causal factors for this relationship may be more easily identified in the actual testing environment. Since the theory underlying the AAT presupposes that “an individual may possess a large amount of both anxieties or of one but not the other, or of none of either” (1, p. 213), four basic anxiety-reaction types can be operationalized by the AAT: f a c i l i t a t o r s 4 s scoring relatively high on the AAT+ and relatively low on the AAT-; d e 6 i l i t a t o r s - S ~scoring relatively low on the AAT+ and relatively high on the AAT-; n o n a f f e c t e d s S s scoring relatively low on both AAT+ and AAT-; and high-affectedslSs scoring relatively high on both AAT scales. One study (9) has suggested that the differential academic performances of the basic reaction types can be predicted by the inverted-U hypothesis curve which explains performance as a function of organismic “arousal” or “activation” in the testing situation. This explanation assumes that under tyfiical testing conditions the test-reaction types have a characteristic position on the inverted-U curve. Given this position as depicted in Figure 1, the different activation levels of each of the three anxiety types contribute directly to the differential group performances. In an attempt to support this explanation Sweeney et al. (12) administered a final examination under conditions induced by instructional set of high, normal, and low arousal. Within each arousal condition, the hypothesized relationships among the reaction types’ performance as generated from the inverted-U curve were either supported or in the predicted direction. The purpose of the present study is to explore further and clarify this relationship by attempting to assess in a more direct manner the activation levels of Ss in the testing situation.



Previous research with the high-affected reaction type has yielded inconsistent and incomprehensible results, and consequently, this type was excluded from !he investigation.

DAVID C. MUNZ et

A

Downloaded by [University of Toronto Libraries] at 09:22 02 February 2015

41 0 NON AFFECTEOS

I

I

d.

A x

FAClL ITATORS DEBILITATORS

LOW

HIGiI L E V E L OF ACTIVATION FIGURE 1 INVERTED-UCURVESHOWINGACHIEVEMENT TEST PERFORMANCE AS A FUNCTIONOF ACTIVATION LEVEL

The superimposed plottings on the theoretical curve show mean performances of the three anxiety reaction types in a typical testing atmosphere.

Because of the difficulty in using physiological measures in a field setting, the Activation-Deactivation Adjective Check List (AD-ACL), a state measure of transitory activation (14), was employed in this study. Following Duffy ( S ) , the AD-ACL was conceived as a measure of total organismic energy release and purports to assess through the use of self-descriptive adjectives an individual’s perception of his feelings of arousal. Factor analysis of these adjectives identified four orthogonal activation dimensions: High Activation (H-Atjittery, intense, stirred-up, fearful, clutched-up; General Activation (G-A)-lively, active, full-of-pep, energetic, peppy, vigorous, activated; General Deactivation (G-D)-at rest, still, leisurely, quiescent, quiet, calm, placid; and Deactivation-Sleep (D-S)-sleepy, tired, drowsy. Thayer (16) has suggested that use L‘of the AD-ACL does not necessarily imply acceptance of the assumption that a single activation continuum underlies all behavior” (p. 87). Recent research (16) has suggested that the G-A scale assesses moderate activation feelings which are not elicited by high anxiety or stress conditions. The D-S scale has been found to represent best the low range of activation feelings in nonstress situations. On the other hand, the H-A scale is most representa-

Downloaded by [University of Toronto Libraries] at 09:22 02 February 2015

42

JOURNAL O F PSYCHOLOGY

tive of feelings associated with anxiety or stress, while the G-D scale appears to measure the low range of activation feelings in stressful situations. The research reported here was designed to test the following specific hypotheses: Hypothesis 1. Differential reactions to test-taking anxiety, as measured by the AAT, significantly affect academic performance scores, with debilitators and nonaffecteds performing at a significantly lower level than facilitators. Hypothesis 2. The above relationships are mediated by differential types and amounts of arousal; i. e., prior to taking the test, debilitators report experiencing significantly more H-A than facilitators and nonaffecteds. Moreover, facilitators report experiencing significantly more G-A than the nonaffecteds and debilitators. And finally, the nonaffecteds report experiencing significantly more G-D than the facilitators and debilitators. No predictions were made concerning anxiety-reaction type and D-S activation, since it seems highly unlikely that this activation scale is relevant to the testing situation. Hypothesis 3. With use of all AAT data (the anxiety reaction types axe operationally constructed by selecting extreme scores on the AAT distributions), directional differences in relationships, as well as significant amounts, are envisioned between the Facilitation scale and Debilitation scale and the different activation scales. Since it is assumed that H-A scores reflect experienced stress or anxiety arousal, the Facilitation scale should correlate negatively with these testing experiences while the Debilitation scores correlate positively. The reverse should hold between G-A scores and the AAT+ and AAT- scale scores.

B.

METHOD Subjects The Ss for this study were 75 members (38 males, 37 females) of a first semester introductory psychology class taught at Saint Louis University. This lower-division course is comprised of freshman and first-semester sophomores. 2 . Instruments The AAT was designed to assess the reported effects of anxiety experienced in the testing environment. To distinguish between different degrees of achievement anxiety that are reported by the respondent as either facilitating or debilitating to test performance, each type of anxiety is 1.

Downloaded by [University of Toronto Libraries] at 09:22 02 February 2015

DAVID C . MUNZ

et d.

43

measured by a separate subtest of items (AAT+ scale and AAT- scale) together comprising a 19-item questionnaire. Ss answer each item on a five-point scale, indicating the degree to which the item applies to them. The AD-ACL, a self-report measure of transitory levels of activation, consists of 22 activation-descriptive adjectives which assess four independent activation dimensions. In addition to these items, the AD-ACL contains a number of other mood-descriptive adjectives which are used to disguise the purpose of the instrument. Validity studies indicate that the AD-ACL (a) correlates substantially with composites of physiological variables, (b) reflects diurnal changes in activation, (c) is predictive of activation changes as a function of an impending college examination, and ( d ) can function as a measure of drive in tests of predictions drawn from Hull-Spence theory (14, 17). With use of a four-point rating scale, Ss are instructed to respond to each adjective “on the basis of how well the adjective describes their feelings at the moment (14,p. 665).” 3. Procedure and Design

Three weeks into the course the AAT was administered at the beginning of class, ostensibly as part of a Master’s thesis to develop two new questionnaires. The students were informed that their participation was needed in the final phase of this research project in order that normative data could be procured for the instruments. Moreover, the graduate student explained that she had been unable to complete the second instrument (AD-ACL) in time for this class testing. However, since the second questionnaire was very short, they would be asked to complete it at a later date. The graduate student emphasized the fact that she was attempting to meet a graduate school deadline on her thesis and that as soon as she had completed the other questionnaire she would again solicit their cooperation. The students were also told that they would receive extra course credit for answering the two questionnaires and that the results of these instruments would be used for research purposes only. One week later, on the class period of the first course examination the graduate student appeared with the instructor. She explained that it had taken her longer than she had envisioned to prepare the second questionnaire and that in order for her to meet her deadline she must collect at this time the rest of her normative data. She apologized for using examination time but explained that the instructor had consented to give them extra time to compensate for the time taken to answer the questionnaire. The AD-ACL was then administered.

Downloaded by [University of Toronto Libraries] at 09:22 02 February 2015

44

JOURNAL OF PSYCHOLOGY

The course examination was administered immediately upon completion of the AD-ACL by each student. The test consisted of 60 randomly presented multiple-choice items. Students were given as much time as they needed to finish the exam. In order to test the first two hypotheses, the three achievement anxiety types were generated by selecting Ss according to the procedure of Munz and Smouse (9). For each student (N = 75) the AAT- score was subtracted from the AAT+ score, and the differences ranked. The top 15 Ss in this distribution were defined as facilitators, and the bottom 15 as debilitators. Scores for the remaining Ss were summed. The 15 scores with the smallest sums became the nonaffecteds. To test the first hypothesis, the examinations were scored for those Ss within each reaction type, and the performance data (total number of items answered correctly) were subjected to a one-way analysis of variance. To test the second hypothesis, the AD-ACL was scored, yielding four factor scores for each S within each anxietyreaction group. Each set of activation data (H-A, G-A, G-D, and D-S) was subjected to a one-way analysis of variance. In order to test the last hypothesis, Pearson product-moment correlation coefficients were computed for all data (N = 75) between each of the anxiety measures (AAT+ and AAT-) and (a) the performance scores and (b) the four activation factor scores.

C. RESULTS The first hypothesis which predicted a relationship between anxiety reaction type and academic performance was supported (F = 4.43, df = 2/42, p < .05). Group mean comparisons within this analysis further substantiated the specific predictions, in that the debilitators’ and nonaffecteds’ mean performances (respectively: = 41.40, S D = 7.01, p < .05; X = 44.73, S D = 7.94, p < .06) were significantly lower than the facilitators’ mean performance @ = 48.93, SD = 5.54). The debilitators’ and nonaffecteds’ mean performance were not significantly different

x

(p

> .05).

Table 1 presents the means and standard deviations for the AD-ACL data. As evidenced in Table 1 only partial support was found for the specific predictions under hypothesis 2 . The ANOVA performed on the H-A data was significant (F = 6.33, df = 2/42, p < .01). Mean comparisons within this analysis confirmed the prediction that debilitators experience in the test situation significantly more H-A than facilitators and nonaffecteds (p < .05). Moreover, nonaffecteds experience higher H-A

DAVID C. MUNZ MEANSAND SDS

TABLE 1 ACTIVATION-DEACTIVATION ADJECTIVE CHECKLIST DATA

FOR THE

Debilitators M SD

Dimension High Activation (H-A) General Activation (G-A) General Deactivation (G-D) Deactivation-Sleep (D-S)

Downloaded by [University of Toronto Libraries] at 09:22 02 February 2015

45

et d .

16.47 16.60 12.67 6.80

Anxiety reaction type Facilitators Nonaffecteds M SD M SD

1.51 7.07 3.72 3.67

12.07 19.07 13.13 6.27

3.75 5.22 3.66 4.23

14.40 18.87 14.20 6.40

4.26 5.00 3.86 2.77

levels than facilitators (p < .05). However, none of the ANOVAs performed on the G-A, G-D, and D-S data yielded a significant effect for anxiety reaction type (p > .05). Yet, as evidenced in Table 1, the means were in the predicted direction within the G-A and G-D activation factors; that is, the facilitators had the highest mean G-A, while the nonaffecteds had the highest G-D mean score of the three reaction types. Table 2 presents a summary of the correlational results for hypothesis 3. The predicted directional relationships between the AAT scales, test performance, and the H-A and G-A measures were found. Also, all correlations were statistically significant except the one between the AAT+ scale and the G-A measure.

D. DISCUSSION The results of this study substantiated previous research findings that reported a relationship between achievement anxiety and academic test performance. For the total sample, the AAT+ scores correlated positively with performance scores, while the AAT- scores correlated negatively. Furthermore, the predicted relationship between anxiety reaction type and academic performance was found. The mean performance scores of the TABLE 2 PEARSON PRODUCT-MOMENT CORRELATIONS BETWEENACHIEVEMENT ANXIETYTEST (AAT) SCORES,PERFORMANCE, A N D ACTIVATION-DEACTIVATION ADJECTIVECHECKLIST (AD-ACL) MEASURES AD-ACL General Activation

Measure

Performance

High Activation

AAT -b AAT-

.33** -.40**

-.35**

.17

.38**

-.23*

* p < .05, 4f = 73, one-tailed test. * * p < .01, 4f = 73, one-tailed test.

Downloaded by [University of Toronto Libraries] at 09:22 02 February 2015

46

JOURNAL OF PSYCHOLOGY

three anxiety types ideally fitted the hypothesized inverted-U curve shown in Figure 1. The facilitators scored significantly higher on the examination than the debilitators and nonaffecteds, and there was no difference between the debilitators and nonaffecteds in test performance. Results of the activation data were not as convincing. It appears that activation level may partially account for the relationship between achievement anxiety and academic test performance, but possibly in a more complex manner than previously envisioned. Correlations for the total sample between the two AAT scales and the G-A and H-A activation subscales support the position that the AAT- scale assesses an individual’s generalized perception of the degree to which his state of arousal in the testing situation has been debilitating to his test performance. And there is some support for the position that the AAT+ scale measures an individual’s generalized perception of the degree to which his state of arousal has been facilitating to his test performance. Activation data for the three anxiety reaction types supported the former finding, since debilitators had more H-A arousal before the exam than any other anxiety type. However, the remaining AD-ACL data only reflected a trend in the predicted direction. Consequently, the inverted-U hypothesis generated to account for the relationship between achievement anxiety reaction type and academic performance was only partially supported. In conclusion, the results of this study suggest that an individual may experience two general types of arousal in the typical testing situation. One state of arousal appears to enhance academic performance, while the other tends to impede performance. Why some individuals experience more facilitating than debilitating arousal in the testing situation and, therefore, have their test performance enhanced, as compared with those who experience more debilitating arousal which tends to lower their performance, is a question for future research. Research should also address itself to finding ways of providing a testing atmosphere that induces a state of arousal conducive to assessing an individual‘s maximum performance level. REFERENCES 1.

2. 3.

ALPERT,R., & HABER,R. N . Anxiety in academic achievement situations. J . Abn. 15 SOC.Psychol., 1960, 61, 207-215. CARRIER,N. A , , & JEWELL,D. 0. Efficiency in measuring the effect of anxiety upon academic performance. 1. Educ. Psychol., 1966, 57, 23-26. DATTA,L. On the unidimensionality of Alpert-Haber achievement anxiety test. Psychol.

Rep., 1967, 20, 606. 4.

DEMBER,W. N . , NAIRNE,F., & MILLER,F. J. Further validation of the Alpert-Haber achievement anxiety test. J . Abn. G SOC. Psychol., 1963, 65, 427-428.

DAVID C . MUNZ 5. 6.

7.

et al.

47

DUFFY,E. Activation and Behavior. New York: Wiley, 1962. MCKEACHIE,W. J. Interaction of achievement cues and facilitating anxiety in the achievement of women. 1.Appl. Psychol., 1969, 53, 147-148. MCKEACHIE,W. J., LIN, Y., MILHOLLAND, J., & ISAACSON, R. Student affiliation motives, teacher warmth, and academic achievement. J . Personal. 6. SOL.Psychol., 1966, 4, 457-461.

8. 9.

MUNZ,D. C., & JACOBS,P. D. An evaluation of perceived item-difficulty sequencing in academic testing. Brit. J . Educ. Psychol., 1971, 41, 195-205. MUNZ,D. C., & SMOUSE,A. D. The interaction effects of item difficulty sequence and achievement anxiety reaction on academic performance. J . Educ. Psychol., 1968, 59, 370-374.

Downloaded by [University of Toronto Libraries] at 09:22 02 February 2015

10. 11.

12.

13. 14.

F'ERVIN,L. A. Aptitude, anxiety and academic performance: A moderator variable analysis. Psychol. Rep., 1967, 20, 215-221. SMOUSE,A. D., & MUNZ,D. C. Item difficulty sequencing and response style: A follow-up analysis. Educ. Psychol. Meas., 1969, 29, 469-472. SWEENEY, C. J., SMOUSE, A. D., RUPIPER,O., & MUNZ,D. C. A test of the inverted-U hypothesis relating achievement anxiety and academic test performance. J. of Psychol. , 1970, 74, 267-273. TERRY,R. L., & ISAACSON, R. M. Item failure and performance on subsequent items of an achievement test. J . of Psychol., 1971, 77, 29-32. THAYER,R. E. Measurement of activation through self-report. Psychol. Rep., 1967, 20, 663-678.

. Activation states as assessed by verbal report and four psychological vanables. Psychophysiology, 1970, 7, 68-94. . Studies of controlled self-reports of activation. Terminal Progress Report, 16. MH-14248-01, National Institute of Mental Health, Public Health Service, Long Beach, California, 1971. 17. THAYER, R. E., & COX, S. J. Activation, manifest anxiety, and verbal learning. J . Exper. Psychol., 1968, 78, 524-526. 18. WALSH,R. P. Some correlates of test-taking anxiety. Psychol. Rep., 1968, 22, 449-450. . Test-taking anxiety and psychological needs. Psychol. Rep., 1969, 25, 83-86. 19 R. O., & O'BRIEN, B. A. Anxiety and test-taking 20 WALSH! R. P., ENGBRETSON, behavior. J. Counsel. Psychol., 1968, 15, 572-575. 15.

Department of Psychology Saint Louis University Saint Louis, Missouri 63103

A further test of the inverted-U hypothesis relating achievement anxiety and academic test performance.

The assumption that the inverted-U hypothesis, which shows performance as a function of activation level, mediates the relationship between achievemen...
478KB Sizes 0 Downloads 0 Views