Effects of Time Limits on Test-Taking Behavior.
Abstract
Three related experimental studies analyzed rate and accuracy of test response under time-limit and no-time-limit conditions. Test instructions and multiple-choice vocabulary items were administered by computer. Student volunteers received monetary rewards under both testing conditions. In the first study college students were blocked into high- and low-ability groups on the basis of pretest scores. Results for both ability groups showed: (1) Higher response rates under time-limit conditions than under no-time-limit conditions; and (2) No significant differences between time-limit and no-time-limit accuracy scores. In the other tests, conditions were varied further. All three studies showed essentially zero correlations between response rate and response accuracy. Response latency data were also analyzed. These data suggested the existence of different test-taking styles and strategies under time-limit and no-time-limit testing conditions. The results of these studies suggest that number-correct scores from time-limit tests are a complex function of response rate, response accuracy, test-taking style and test-taking strategy, and therefore are not likely to be as valid or useful as number-correct scores from no-time-limit-tests.
Document Details
- Document Type
- Technical Report
- Publication Date
- Apr 01, 1976
- Accession Number
- ADA024422
Entities
People
- David J. Weiss
- T. W. Miller
Organizations
- University of Minnesota