Journal article
Reviewing and Changing Answers on Computer-adaptive and Self-adaptive Vocabulary Tests
Journal of educational measurement, Vol.35(4), pp.328-345
12/1998
DOI: 10.1111/j.1745-3984.1998.tb00542.x
Abstract
Results obtained from computer-adaptive and self-adaptive tests were compared under conditions in which item review was permitted and not permitted. Comparisons of answers before and after review within the "review" condition showed that a small percentage of answers was changed (5.23%), that more answers were changed from wrong to right than from right to wrong (by a ratio of 2.92:1), that most examinees (66.5%) changed answers to at least some questions, that most examinees who changed answers improved their ability estimates by doing so (by a ratio of 2.55 to 1), and that review was particularly beneficial to examinees at high ability levels. Comparisons between the "review" and "no-review" conditions yielded no significant differences in ability estimates or in estimated measurement error and provided no trustworthy evidence that test anxiety moderated the effects of review on those indexes. Most examinees desired review, but permitting it increased testing time by 41%.
Details
- Title: Subtitle
- Reviewing and Changing Answers on Computer-adaptive and Self-adaptive Vocabulary Tests
- Creators
- Walter P. Vispoel - University of Iowa
- Resource Type
- Journal article
- Publication Details
- Journal of educational measurement, Vol.35(4), pp.328-345
- Publisher
- Blackwell Publishing Ltd
- DOI
- 10.1111/j.1745-3984.1998.tb00542.x
- ISSN
- 0022-0655
- eISSN
- 1745-3984
- Number of pages
- 18
- Language
- English
- Date published
- 12/1998
- Academic Unit
- Psychological and Quantitative Foundations
- Record Identifier
- 9984371277502771
Metrics
4 Record Views