Effects of answer changing on tests by 133 fifth- and sixth-grade students were investigated to determine the degree to which these students changed answers, the extent to which they gained from changing, and the reasons they advanced for the changes. Students were from a predominantly suburban district (n=88) and a predominantly rural district (n=45). Relationships of answer changing behavior with test score, gender, and test-taking strategy were also examined. The majority of changes were wrong-to-right for all subgroups. High-scoring students changed fewer answers but had a higher proportion of changes from wrong-to-right than did other students. Reasons for changing, determined through interviews, were similar to those given by adults in previous studies, but with more clerical changes. For each test-taking strategy, students who changed tended to gain. Students should not be discouraged from changing answers on tests, especially given the preponderance of gain from changing, even for subgroups. Three tables present study data and there is a 21-item list of references. (Author/SLD)