THE RELATIONSHIP BETWEEN TEST ANXIETY AND SILENT READING GAIN DONALD NEVILLE, PHILIP PFOST, and VIRGINIA DOBBS George Peabody College for Teachers Failure to learn to read has continued to be a concern in education. Anxiety, as a personality variable, has been shown to be related to learning. The present study investigated the relationship between test anxiety and silent reading gains as measured by standardized tests of vocabulary and comprehension skills. It was hypothesized that high test anxiety: (a) has an inverse relationship to vocabulary gain and (b) has an inverse relationship to comprehension gain. Anxiety, in this study, is assumed to be an unpleasant emotional state with physiological concomitants which are consciously experienced by the child and to be related to test and test-like situations. Within this framework, anxiety becomes a hypothetical construct having predictable consequences in terms of its effect on gain in silent reading as measured by standardized tests. Other researchers, drawing from the work of Hull and Spence, have viewed anxiety as an operational measure of drive-level. Ruebush (1963, p. 474) feels that although these definitions may differ . . . "they demonstrate how the construct of anxiety may be useful in the systematic analysis and prediction of children's behavior in learning and problem-solving situations." Many recent studies, using the Sarason Test Anxiety Scale for Children (Sarason, and others, 1960), have examined the relationship of anxiety to achievement and intelligence. Sarason (1963) and Cox (1962) found that anxiety relates negatively to school achievement. Others (Feldhusen and Klausmeier, 1962; Waite, and others, 1958; Sarason, 1961) have found a negative correlation between anxiety and IQ. The most obvious conclusion from these data would be that highly anxious children are less intelligent and/or lower achievers. One might argue that children who are less intelligent and lower achievers deal less effectively with their environment and, therefore, acquire higher levels of anxiety. However, an alternative hypothesis would suggest that depressed test scores result from the interfering effects of anxiety. If this were the case, it might be well to question the validity of both the intelligence and achievement test scores of highly anxious children 45 Downloaded from http://aerj.aera.net at PENNSYLVANIA STATE UNIV on May 11, 2016 (Sarason, 46 AERJ • VOLUME 4 • NUMBER 1 • JANUARY 1967 METHOD The subjects were 54 boys enrolled in a six weeks summer reading program offered by the Child Study Center of George Peabody College for Teachers. The chronological age range of the subjects was seven through fourteen, with a mean of 10.5. The grade level range was three through nine, with a mean of 5.5. Before instruction began, the subjects were given the Peabody Picture Vocabulary Test, Form B; the appropriate level of the Lorge-Thorndike Group Intelligence Test, Form A; the Test Anxiety Scale for Children (TASC) by Sarason; and either the Gates Basic or Advanced Primary Reading tests. Adequate validity and reliability data have been reported for all of the above instruments. The TASC was selected because: (a) it was designed for children, (b) it measures anxiety as it is aroused by test situations, and (c) it has been used previously to study the relationship between anxiety and learning. The orientation of the instructional program might be characterized as one of diagnostic teaching. The teacher emphasized helping the child to gain skills in those areas where he had initially exhibited deficiencies. At the end of the summer reading program the subjects were given a parallel form of the Gates Reading Tests. To determine the silent vocabulary and comprehension gain, each subject's score on the pretest was subtracted from his score on the posttest. In order to analyze the data, the subjects were divided into three groups according to their scores on the TASC: High Anxiety (HA), Middle Anxiety (MA), and Low Anxiety (LA). (See Table 1.) Since a TABLE 1 Composition of Anxiety Groups High Anxiety Middle Anxiety Low Anxiety Total Group N Range of Scores Mean Score 17 18 19 54 15-26 9-14 0-8 0-26 20.0 9.44 4.68 11.59 major purpose of this study was to examine how the level of anxiety was related to achievement gain, it was necessary to establish the fact that the groups were not different on several factors related to achievement; namely, IQ, age, and grade. Three simple analyses of variance comparing the three anxiety groups on each of these factors resulted in F ratios well below that required for significance. Thus, it was assumed that any effect these factors may have had on reading gains could be attributed to chance, Downloaded from http://aerj.aera.net at PENNSYLVANIA STATE UNIV on May 11, 2016 47 TEST ANXIETY AND SILENT READING RESULTS To determine if these groups performed differently on vocabulary tasks than they did on comprehension tasks, an analysis in which the gain in these skills could be viewed both in terms of their relatedness and in terms of their separateness was needed. For this purpose the Type I analysis of variance, as suggested by Lindquist (1953), was chosen to compare the three anxiety levels on gains in comprehension and vocabulary. It resulted in a significant F ratio for the interaction term (Table 2). Therefore, the first analysis was followed by two simple analyses of variTABLE 2 Type I Analysis of Variance Comparison Source df Sum of Squares Between S's Anxiety Levels (B) error (b) Within S's Vocab. & Comp. (A) Interaction (A X B) error (w) 53 2 51 54 1 2 51 5375 586 4789 3271 114 524 2633 F 3.12 (N.S.) 2.21 (N.S.) 5.08 (p < .01) ance computations; one each for vocabulary and comprehension. The resulting F ratios suggested that there were no differences among the three anxiety levels in vocabulary gains but that there were differences in comprehension gains (Table 3). Subsequent t comparisons disclosed that the TABLE 3 Simple Analyses of Variance Comparisons of Vocabulary and Comprehension Gains Source df Sum of Squares Comprehension Gains Between groups Error (within) 2 51 1097.8 4810.3 Vocabulary ocabulary Gains Between groups Within groups (error) 2 51 13.3 2610.8 F 5.82 (p < .01) .13 (N.S.) MA group made gains in comprehension significantly greater than either of the other groups (HA-MA, p < .001; MA-LA, p < .05). Also, it was found that the HA group gained significantly more on vocabulary than they did on comprehension (p < .05). In fact, as can be seen from Fig. 1, the HA group had a lower mean comprehension score on the posttest than they did on the pretest. Downloaded from http://aerj.aera.net at PENNSYLVANIA STATE UNIV on May 11, 2016 48 AERJ • VOLUME 4 • NUMBER 1 • JANUARY 1967 Fig. 1—Vocabulary and Comprehension Gains. DISCUSSION On the basis of these results, the first hypothesis is rejected and the second hypothesis is accepted. High test anxiety does have an inverse relationship to comprehension gain but not to vocabulary gain. However, both very high and very low test anxiety are associated with little or no gain in silent reading comprehension, while a medium level of anxiety seems to be associated with greater gain. Hebb (1955) has theorized about the curvilinear (inverted U-shaped) nature of the effect of anxiety upon learning. Only a few studies have supported his thesis at this time, but most research has examined this relationship using only high and low anxiety groups. Cox (1960), using an adaptation of the TASC for Australian children, divided his subjects into high, medium, and low anxiety groups, and found this inverted U relationship between anxiety and a single composite academic score composed of arithmetic, composition, grammar, spelling, and reading. Reed (1960) reviewed the status of research in anxiety and concluded that: (a) severe anxiety seems to depress learning, (b) a mild degree of anxiety may function in a positive manner for some forms of learning, and (c) a very low level of anxiety may depress learning. I t must be added Downloaded from http://aerj.aera.net at PENNSYLVANIA STATE UNIV on May 11, 2016 TEST ANXIETY AND SILENT READING 49 that Reed's conclusions were based on research done with the manifest anxiety scale which is often seen as a measure of drive level or motivation; whereas, the present study used the TASC which was developed from a psychoanalytic viewpoint and views anxiety as negatively correlated with performance. Even so, it is apparent that the effect of anxiety upon comprehension in this study supports Hebb's (1955) thesis and Reed's (1960) conclusions. Perhaps this can be explained if vocabulary and comprehension can be considered in the same way that Spence (1956) has viewed the effect of anxiety on simple and complex learning tasks. It seems reasonable to assert that comprehension is a more complex task than vocabulary, and that the complexity of the task and the high anxiety resulted in inferior performances among the HA group. It is difficult to explain the LA group in this light, but it appears possible that the successful performance of a complex task requires more personal involvement than does the successful performance of a more simple task. Thus, the LA group had enough involvement to perform as well as the MA group on the vocabulary task, but not enough to perform as well on the comprehension task. Another possible explanation would assume that many LA Ss are highly defensive and therefore like the HA Ss. An attempt was made to examine the LA group for defensiveness, in terms of their scores on the Lie Scale from the General Anxiety Scale. The reading gains of the LA Ss with high lie appeared to be more like the HA Ss than they were like the other LA Ss, but no valid conclusion could be drawn. The problem concerning LA youngsters who are highly defensive certainly needs further consideration, and at the present time the authors are engaged in a study which may provide relevant information. In light of present knowledge, it seems that it would be worth while for those concerned with the education of children to be cognizant of the fact that anxiety is a relevant factor in influencing children's performance on tests, and probably is a negative influence on general academic performance. Perhaps through further study it will be possible to develop programs which will minimize the detrimental effects of anxiety on learning. REFERENCES Cox, F. N. "Correlates of General and Test Anxiety in Children." Australian Journal of Psychology 12: 169-77; December 1960. Cox, F. N. "Educational Streaming and General Test Anxiety." Child Development 33: 381-90; June 1962. FELDHUSEN, JOHN F., and KLAUSMEIER, HERBERT J. "Anxiety, Intelligence, and Achievement in Children of Low, Average, and High Intelligence." Child Development 33: 403-09; June 1962. HEBB, DONALD 0. "Drives and the Conceptual Nervous System." Downloaded from http://aerj.aera.net at PENNSYLVANIA STATE UNIV on May 11, 2016 Psychological Review 62: 243- 50 AERJ • VOLUME 4 • NUMBER 1 • JANUARY 1967 F. Design and Analysis of Experiments in Psychology and Education. Boston: Houghton Mifflin Co., 1953. 393 pp. RUEBUSH, BRITTON K. "Anxiety." Child Psychology. Sixty-Second Yearbook, Part I, National Society for the Study of Education. Chicago: University of Chicago Press, 1963. Chapter 11, pp. 460-516. REED, HORACE B. "Anxiety: The Ambivalent Variable." Harvard Educational Review 30: 141-53; Spring 1960. LINDQUIST, EVERET SARASON, IRWIN G. "Test Anxiety and Intellectual Performance." G. "Test Anxiety and the Intellectual Performance of College Students." Journal of Educational Psychology 52: 201-06; August 1961. SARASON, SEYMORE B., and others. Anxiety in Elementary School Children. New York: John Wiley and Sons, 1960. 351 pp. SARASON, SEYMORE B., and others. "A Test Anxiety Scale for Children." Child Development 29: 105-13; March 1959. SPENCE, KENNETH W. Behavior Theory and Conditioning. New Haven, Connecticut: Yale University Press, 1956. 262 pp. WAITE, RICHARD R., and others. "A Study of Anxiety and Learning in Children." Journal of Abnormal Psychology 57:267-70; November 1958. SARASON, IRWIN (Received August, 1965) AUTHORS NEVILLE, DONALD D. Address: Child Study Center, Box 158, Peabody College for Teachers, Nashville, Tenn. 37203 Title: Assoc. Prof, of Education and Psychology; Director, Child Study Center Age: 39 Degrees: B.S., M.Ed., Kent State Univ.; Ed.D., Univ. of Florida Specialization: Learning problems in elementary school children. PFOST, H. P H I L I P Address: Box 163, Peabody College for Teachers, Nashville, Tenn. 37203 Title: Research Scientist Age: 36 Degrees: B.A., M.Ed., Northwestern State College; Ed.D., George Peabody College Specialization: Remedial reading, clinical education, learning disorders. DOBBS, VIRGINIA Address: 620 Barrywood Drive, Nashville, Tenn. 37220 Title: Director, Evaluation and Research ESEA, Metropolitan Public Schools Age: 52 Degrees: B.S., Middle Tennessee State Univ.; M.A., Ed.D., George Peabody College Specialization: Educational research, psychology, special education. Downloaded from http://aerj.aera.net at PENNSYLVANIA STATE UNIV on May 11, 2016 Journal of Abnorma
© Copyright 2026 Paperzz