effects of test format and locus of control on test anxiety, The

Journal of College Student Development, Nov/Dec 1998 by Choi, Namok

Numerous empirical studies have provided evidence of the inverse relationship between test anxiety and academic performance of college students (e.g. Herman, 1990; Hunsley, 1985). Considering the detrimental effects of test anxiety on performance, identifying factors that contribute to test anxiety becomes utmost important. One factor that has been examined in relation to test anxiety is locus of control (deMan, Hall, & Stout,1991; Feather & Volkmer,1988). Internals, people with internal locus of control, perceive events to be the consequences of their actions (Weiner, 1985). Externals, people with external locus of control, perceive events to be independent of their actions. Since internals attribute failure or success to factors within self, internally controlled students tend to have control over their academic behavior, strive for achievement (Volkmer & Feather,1991), be less anxious than externals in testing situations (Shelton & Mallinckrodt, 1991), and attain higher levels of academic achievement (Gilmor & Reid,1978).

Another factor that has been separately examined from locus of control, but appears to be related to test anxiety is test format. Several studies on college students' attitudes toward different test formats indicated that college students may experience less test anxiety when a preferred test format is provided (Zoller & BenChaim, 1988), and that they tend to prefer objective tests to subjective ones (Anderson, 1987). In comparisons of the two most widely used formats, multiple-choice seems to be preferred to essay by college students (Anderson, 1987; McDaniel, Challis, & Sadowski, 1991). These investigations were, however, non-experimental in nature and the relationship between test format and test anxiety was speculated based on students' test format preference. Given the relationships among test anxiety, test format, and test format preference, based on the results of studies mentioned above, it seems logical to speculate that essay format tends to evoke higher level of test anxiety than multiple-choice format. This speculation seems to be supported by the findings of Benjamin, McKeachie, Lin, and Holinger (1981) in which highly test anxious students did more poorly on essay tests in their study.

No studies to date have simultaneously examined the effects of both locus of control and test format on test anxiety. This study was conducted to investigate the joint effects of locus of control and test format on test anxiety in a quasi experimental context. Considering that students with internal locus of control tend to have better control over their academic behavior, it seems reasonable to expect that internally oriented students would be less affected by test format. To aid the interpretation, the relationship between students' attitudes toward test format and test anxiety was also examined.

METHOD

Participants

Participants in this study (15 males and 71 females) were drawn from four undergraduate measurement classes at a large midwestern university. The course is required for a bachelor's degree in teacher education at the university. The ethnic composition of the sample was Caucasian (95%), Native American (3%), and Hispanic (2%). No extra credit was awarded for their participation.

Instruments

The Internal-External Locus of Control Scale (I-E; Rotter, 1966) was used to measure locus of control. The I-E scale consists of 23 items with the total number of external choices yielding a maximum score of 23, higher scores indicating more externality. Rotter reported a coefficients ranging from .65 to .79. The a coefficient for the I-E scale observed from the sample of this study was .73.

The Worry-Emotionality Questionnaire (WEQ; Morris, Davis, & Hutchings, 1981) was used to measure test anxiety. The WEQ is a 10 item self-report measure with two subscales; worry and emotionality. The WEQ yields a total scale score that may range from 10 to 50, higher scores indicating high levels of test anxiety. Morris et al. (1981) reported a coefficient of .81 for worry, and .86 for emotionality. The overall internal consistency reliability of the WEQ was not reported by the authors. The observed a coefficients in the current study were .77, .91, and .89 for worry, emotionality, and overall, respectively. Total scale scores of the WEQ were used to measure overall test anxiety in this study.

Procedures

Prior to the beginning of the semester, four sections of the class, each taught by a different instructor (3 Caucasian males and 1 Caucasian female), were randomly assigned to either an essay test format or multiple-choice test format. This was the format under which students would take their first course exam. Students in all sections were informed by their instructors during the first week of class what format to expect for their first exam. During the second class period, locus of control and background variables (including participants' ratings of their preference for various test formats on a 5-point scale) were assessed.