Instability of Person Misfit and Ability Estimates Subject

Unexpected response patterns on tests and their problematic interpretation has led to an intense research activity (i) to investigate the sources of such responses and (ii) to model statistics that claimed to detect person misfit in an examinee’s response pattern. The rationale behind this effort was the claim that the test scores of these examinees with unexpected response patterns may fail to provide a useful and valid measure of their ability. In this study we have followed-up ‘misfitting’ examinees in a Mathematics test and during interviews we have asked them to work out items again on which they have provided unexpected responses. Pupils’ response patterns were then changed based on their responses during interviews and fit analyses were re-run in order to obtain new estimates of their abilities and of the fit statistics. When old and new estimates were compared using a paired ttest, there was an overall increase in ability estimates and an overall reduction in pupils’ Infit and Outfit values. However only the change in the Infit values was statistically significant. By examining the change in ability estimates for each individual pupil we have found cases where the two measures were seriously discordant, raising questions about the validity of test-measurement for these pupils.



  • Version
  • 1 Download
  • 156.81 KB File Size
  • 1 File Count
  • July 19, 2018 Create Date
  • July 19, 2018 Last Updated