| Summary: | The writer compared results of computerized vocabulary tests under conditions in which item review was allowed and not allowed. He argues that comparison between the review and no-review conditions did not yield statistically significance differences in number correct scores, ability estimates, or measurement error but points out that examinees in both conditions strongly desired review opportunities. He asserts that comparisons of answers before and after review within the review condition indicated that only a small percentage of answers was altered, that more answers were altered from wrong to right than vice versa, that a large proportion of examinees changed answers to at least some questions, and that most examinees who altered answers enhanced their performance by doing so. Furthermore, he contends that performance gains after review were greater for those of higher ability and that review was desired more by examinees with higher test anxiety.
|
|---|