Item-response changing as a function of test anxiety was investigated. Data supported the hypothesis that high test-anxious students make more item-response changes than low test-anxious students. Also, both high- and low-anxious students profit to a similar extent proportionally from answer changing. (Author/GK)