NotesFAQContact Us
Search Tips
ERIC Number: ED420695
Record Type: Non-Journal
Publication Date: 1998-Apr-13
Pages: 31
Abstractor: N/A
Reference Count: N/A
The Influence of Nonmodel-Fitting Examinees in Estimating Person Parameters.
Nering, Michael L.
The issue of person fit has received an increasing amount of attention by researchers in the past few years. Several studies have focused on the issue of how nonmodel-fitting responses affect the accuracy of ability estimates (e.g. R. Meijer and S. Nering, in press; Reise, 1995). The purpose of this study was to examine the effects that nonmodel-fitting response vectors (NRVs) have on the estimation of person parameters for model-fitting response vectors (MRVs). Under the assumption of local dependence in item response theory, one examinee should not influence the test results of other examinees. However, if NRVs are present in a calibration sample, and they affect the quality of item parameter estimates, this could cause error in person parameter estimates for MRVs. In this study, ability estimates and person-fit statistics were estimated for MRVs in calibration samples that contained different amounts and types of NRVs. It was found that an expected a posteriori estimation procedure tended to result in reduced bias in ability estimates, while a biweight estimation procedure resulted in person-fit statistics that were more normally distributed. The ZU3 nonparametric person-fit index was much less sensitive to NRVs compared to the parametric person-fit statistic. The results of this study demonstrated that researchers should consider the calibration sample seriously when estimating person parameters. (Contains 5 figures, 4 tables, and 29 references.) (Author/SLD)
Publication Type: Reports - Evaluative; Speeches/Meeting Papers
Education Level: N/A
Audience: N/A
Language: English
Sponsor: N/A
Authoring Institution: N/A