| Abstract: | Serial retesting was investigated as a means of improving the test performance of high-anxious students relative to that of their low-anxious peers. The Test Anxiety Scale for Children was given to about 300 seventh grade students in a suburban school; the middle 50% were eliminated to maximize experimental variance. Six groups were established for high and low anxiety levels and for three ability levels. The Academic Promise Tests (except for the Abstract Reasoning Test) were then given four times at 1-week intervals in the fall of the school year. All groups continued to improve significantly in a monotonic, linear fashion, with the greatest improvement occurring on the Numerical Test. Contrary to expectation, there were no significant interactions among previous achievement levels, predispositional test anxiety levels, and test trials. The greatest change on the Numerical Test occurred between the first test trial and the second test trial. |
