Construct-Irrelevant Variance: Examining Differential Speededness in Clinical Science Subject Exams

• In Event: 66.019 - Multicultural and Diversity Issues in Measurement and Assessment

Abstract

This study examined speededness for clinical science examinations. The study included medical students who took five computer-administered clinical science exams. Traditional measures of speededness including item response times and difficulty for exam sections based on item presentation order and groups conditioned on total test ability were examined. Measures of speededness indicated that time limits were sufficient for each exam based on total test indices. Response time and difficulty by section and ability group indicated that lower ability groups showed larger decreases in response time and difficulty than higher ability groups. Speededness did not appear to be a problem for these exams. Speededness may be confounded with perceived time pressures and content for lower ability groups and should be examined.

Authors

• Linette P. Ross, National Board of Medical Examiners

• Carol A. Morrison, National Board of Medical Examiners

• Allie Routhenstein