The relation of mathematics anxiety to situationally-assessed test anxiety, math performance, physiological arousal, and math avoidance behavior was investigated. Sixty-three undergraduates (23 men, 40 women) completed Suinn's Mathematics Anxiety Rating Scale, Fennema and Sherman's Mathematics Anxiety Scale, Sandman's Anxiety Toward Mathematics Scale, and Spielberger's Test Anxiety Inventory prior to completing three math tasks. During the math tasks, heart rate, skin conductance level, skin fluctuations, and avoidance behavior were monitored. Subjects also completed Deffenbacher's Post-Task Questionnaire, a situational measure of test anxiety, worry, and emotionality. Results indicated that math anxiety measures were more highly rated to each other than to test anxiety. Math anxiety accounted for 14-23% of the variance in two math tasks, whereas, math ability (SAT-Q) accounted for 30-42%. Rarely, did math anxiety add to the variance accounted for by ability. The physiological and avoidance measures showed little relation to math anxiety.As Dew, Galassi, and Galassi (1983) noted, math anxiety has sparked a great deal of interest, but a number of fundamental unanswered questions still surround the construct and its measurement. The purpose of this study was to investigate some of these issues more comprehensively and in the context of performance situations.First, the relation of math anxiety to test anxiety, and particularly situation-specific (state) test anxiety and its worry and emotionality components, was explored. If math and test anxiety are related, this relation should be especially evident in a math test-like situation. Second, the extent to which math anxiety interferes with performance was investigated for three math tasks to determine whether math anxiety in-This study is based on the first author's doctoral dissertation, supervised by the second author.

The study investigated four basic questions about mathematics anxiety: (a) the degree of math anxiety experienced by men and women; (b) the internal consistency and test-retest reliabilities of three math anxiety measures; (c) the relationship of math anxiety instruments to each other; and (d) the relationship of math anxiety to test anxiety and its worry and emotionality components. The subjects were 769 students at the University of North Carolina at Greensboro. Math anxiety was measured by Fennema and Sherman's Mathematics Anxiety Scale (MAS), Sandman's Anxiety Toward Mathematics Scale (ATMS), and Suinn's Mathematics Anxiety Rating, Scale (MARS). Test anxiety was measured by the Spielberger Test Anxiety Inventory. Results indicated nonequivalent internal consistency and test-retest reliability coefficients for the three math anxiety measures with the ATMS having the lowest coefficients. Small but significant gender differences were found on two (MARS and MAS) of the measures. The math anxiety measures were moderately related to each other and, almost invariably, they were more closely related to each other than to test anxiety and its components.Mathematics anxiety has sparked a great deal of interest in both the popular (e.g., Tobias, 1976) and professional literature (e.g., Richardson & Woolfolk, 1980). It has been proposed as an important factor in educational and career development (Betz, 1978); assumed to be more prevalent in women than men (Fox, Fennema, & Sherman, 1977); used to explain why women tend to enroll in fewer math courses (Betz, 1978); and invoked to account for the finding that women often perform more poorly in math (Richardson & Suinn, 1972). Math anxiety has also been viewed as related to or as a form of test anxiety (Richardson & Woolfolk, 1980). The worry and emotionality components proposed for test anxiety (Liebert & Morris, 1967) have even beenThis study is based on the first author's doctoral dissertation, supervised by the second author and submitted to the University of North Carolina at Chapel Hill.

No abstract