ration to the counseling of test anxious high school students. Thirty 10th grade test anxious students were randomly assigned in equal numbers to ei-.
Copyright 1981 by the American Psychological Association, Inc. 0022-0167/81/2806-0525$00.75
Journal of Counseling Psychology 1981, Vol. 28, No. 6, 525-528
Cognitive Modification and Systematic Desensitization With Test Anxious High School Students Lois L. Leal, Edward G. Baxter, Jack Martin, and Ronald W. Marx Simon Fraser University, Burnaby, British Columbia, Canada
Recent research, primarily with university undergraduates, has explored the relative effectiveness of cognitive modification versus systematic desensitization for the alleviation of test anxiety. The present study extends this exploration to the counseling of test anxious high school students. Thirty 10th grade test anxious students were randomly assigned in equal numbers to either a cognitive modification, systematic desensitization, or waiting-list control group. Raven's Standard Progressive Matrices, The State-Trait Anxiety Inventory—State form (STAI-S), and the Anxiety Differential were administered as pre- and posttests. Statistically significant Treatment X Time interactions emerged on the Raven's and the STAI-S, favoring an increase for the systematic densensitization treatment on the Raven's and a decrease for the cognitive modification treatment on the STAI-S. These findings are discussed in relation to the results of closely related studies in this area. Numerous studies have assessed the efficacy of various procedures for reducing anxiety. In Barrios and Shigetomi's (1979) review of coping-skills training and anxiety management, 25 of the 62 studies reviewed (40%) dealt with test anxiety. Systematic desensitization and cognitive modification are two common approaches. Three published studies are available that contrast the relative effectiveness of systematic desensitization and cognitive modification for the reduction of test anxiety (Holroyd, 1976; Kaplan, McCordick, & Twitchell, 1979; Meichenbaum, 1972). All three studies employed university undergraduates as participants. The first of the studies, by Meichenbaum (1972), established that students in a cognitivebehavior modification group performed better than students receiving standard desensitization or a waiting-list control group on self-report measures of test anxiety (Adjective Checklist; Gough & Heilbrun, 1965; and Anxiety Differential; Husek & Alexander, 1963). However, although both treatment groups produced superior improvement compared with the waiting-list control group on grade point average and a digit This article is based on masters theses submitted by the first two authors to the Faculty of Education, Simon Fraser University. Requests for reprints or a copy of the full-length manuscript from which this report was taken should be sent to Ronald W. Marx, Instructional Psychology Research Group, Faculty of Education, Simon Fraser University, Burnaby, British Columbia V5A 1S6, Canada.
symbol test, there were no significant differences between treatment groups. On a second analog performance measure (Raven's Standard Progressive Matrices), the control group improved as much as the treatment groups, apparently as a result of practice effects. In a more recent study, Holroyd (1976) used a placebo condition, cognitive modification, systematic desensitization, and a combined systematic desensitization and cognitive modification procedure in the group treatment of test anxiety. He found that the cognitive therapy was superior to the other methods used in the study on three measures (grade point average, StateTrait Anxiety Inventory—State form, and Anxiety Differential). On the fourth measure (a digit symbol test), cognitive modification and combined systematic desensitization and cognitive modification were more effective than the other conditions. Generally, the other treatments resulted in similar effects. The third study (Kaplan et al., 1979) varied, in a factorial design, the presence or absence of a cognitive treatment with the presence or absence of systematic desensitization. This resulted in four groups: waiting-list control, cognitive modification, systematic desensitization, and cognitive modification and systematic desensitization combined. The results favored the cognitive modification group over the other groups on all self-report measures (LiebertMorris Test Anxiety Scale and self-ratings of emotionality and worry). The general trend on self-report and performance measures (a digit symbol performance task) was that the cogni-
525
526
BRIEF REPORTS
30 students who were randomly assigned to the three experimental conditions. The cutoff criteria on these measures were employed to ensure that subjects were primarily test anxious rather than generally anxious and that their anxiety was not due to study problems. Prior to the treatment phase of the study, the Raven's Standard Progressive Matrices were administered to each subject in a simulated stress situation. The State-Trait Anxiety Inventory—State form (STAI-S) and the Anxiety Differential were then administered. The evaluative situation was expected to be anxiety eliciting and similar to regular testing conditions. The validity of this assumption was apparent in expressed verbal concerns by some subjects about finding out what marks they obtained. Subjects were desensitized to the stress instructions at the Method end of the experiment. Two graduate students in counseling conSubjects and Measures ducted the treatment group sessions. One stuThe 3010th grade students participating in this dent administered the cognitive modification study were drawn from a larger group of 122 vol- program, and the other student conducted the systematic desensitization program. Each unteers. Three self-report measures were used to assure counselor followed a manual that detailed the that the students selected for the study were in- treatment procedures step-by-step. These deed test anxious, were not suffering from general manuals were constructed to ensure that both anxiety in addition to test anxiety, and did not treatment groups engaged in exactly the same have deficiencies in study skills. The Achieve- combinations of counselor instruction, group ment Anxiety Test (Alpert & Haber, 1960) was discussion, and active individual and group used as a measure of anxiety specific to academic counseling. All groups met for six 1 -hour weekly situations, and thus was a proxy for test anxiety. sessions. Following the completion of the test anxiety The Fear Survey Schedule (Wolpe, 1969) was used to assess fear of things or events in addition programs, subjects again performed on the Rato tests. The Study Habits Checklist (Preston ven's Standard Progressive Matrices in the sim& Botel, 1967) was used to obtain students' self- ulated stress situation and completed the Anxiety reports of their study habits. Differential and STAI-S. The dependent variable measures were Raven's Standard Progressive Matrices (Raven, 1965), Treatments used to measure the effects of test anxiety in a The cognitive modification procedure was test-like situation, and two self-report measures of anxiety, the Anxiety Differential (Husek & based on that used by Holroyd (1976) and Alexander, 1963) and the State-Trait Anxiety Meichenbaum (1972). Anxiety was explained to Inventory—State form (Spielberger, Gorsuch, & the subjects as resulting from their thoughts and Lushene, 1970). self-statements occurring before and during exams. Subjects were told that an awareness of these thoughts and self-statements was necessary Procedure so that they could create incompatible responses Individuals scoring under 34 on the Achieve- to anxiety engendering ones. Subjects learned ment Anxiety Test were excluded from the study to label emotional arousal, recognize inapprobecause a pilot study indicated that this proce- priate responses, and replace self-defeating dure would ensure that only the upper 10% of test thoughts and self-statements with more positive anxious students in the school would be sampled. alternatives. From an initial 42 students selected on the basis Procedures for systematic desensitization of their scores on the Achievement Anxiety Test, generally followed those of Holroyd (1976). Deep 29% did not meet selection criteria set for the muscle relaxation was paired with presentation Study Habits Checklist (raw score not less than of imaginary anxiety provoking scenes. The 62) and Fear Survey Schedule (raw score not more pairing was meant to inhibit autonomic arousal than 187), or did not care to participate, leaving and to elicit an alternate response pattern to the
tive-only treatment was most successful, followed by the combined, desensitization, and control groups, respectively. The present study extends the experimental examination of the relative effectiveness of systematic desensitization and cognitive modification in the alleviation of test anxiety to a new population—high school students. On the basis of the Holroyd (1976), Kaplan et al. (1979), and Meichenbaum (1972) results, the following conservative predictions were made: (a) the cognitive modification group will be superior to the other two groups on posttreatment self-report measures of test anxiety and (b) there will be no significant differences among treatment groups on posttreatment performance measures.
BRIEF REPORTS stimulus of the exam situation. While the subject is in a relaxed state, the counselor has the individual imagine mild anxiety-provoking test situations. Eventually, the subject imagines situations closely approximating an immediate test situation while still trying to remain relaxed. Students assigned to the waiting-list control group participated in the pre- and posttesting under the simulated stress situations. Results Cronbach alphas were calculated as an index of reliability for the three screening instruments using scores from the entire student sample (N = 122). The lowest alpha was .84. The alpha coefficients (n = 30) for the three outcome measures at pretest were: STAI-S, .84; Anxiety Differential, .74; and Raven's, .87. Alphas at posttests were: STAI-S, .83; Anxiety Differential, .69; and Raven's, .86. To test for the adequacy of random assignment, one-way analyses of variance (ANOVAs) were computed among the treatment groups on all three dependent variables. All three of these analyses resulted in nonsignificant F ratios. Table 1 reports the means and standard deviations for the three treatment groups on the three dependent variables. Separate 3 X 2 ANOVAs with repeated measures on the second factor (pretest-posttest) were calculated for each dependent variable. There were no statistically significant effects on the Anxiety Differential. Table 1 Means and Standard Deviations for Treatment Groups on All Dependent Variables Pretest Measure and treatment group
Postteat
M
SD
M
SD
42.00 39.40 35.10
8.87 4.81 9.87
35.20 40.10 35.40
7.90 8.25 9.66
40.40 40.80 41.70
4.58 11.74 4.83
43.90 45.00 41.30
3.64 9.01 6.95
52.20 46.20 54.00
12.63 10.96 6.62
46.10 45.50 53.70
10.02 9.95 7.86
STAI-S
CM SD CO Raven's
CM SD CO Anxiety Differential
CM SD CO
Note, n = 10 for all groups. STAI-S = Stait-Trait Anxiety Inventory—State; Raven's = Raven's Standard Progressive Matrices; CM = cognitive modification; SD = systematic desensitization; CO = control.
527
There was a significant main effect for time on the Raven's, F (I, 27) = 10.55, p < .01, as well as a significant interaction effect, F (2, 27) = 3.65, p