ERIC Number: ED350350
Record Type: RIE
Publication Date: 1992-Apr
Reference Count: N/A
Analysis of Spring 1990 Field Tryouts of MEAP Essential Skills Mathematics Test Items Using a Partial Credit Model.
Brown, William L.
The partial credit model of G. N. Masters (1982), a one-parameter unidimensional polychotomous Rasch model, was used to reduce the error of measurement, particularly for students near the cut score, and to permit measurement to reflect the actual ability of a student more accurately by reducing the degree of misfit for students near the cut scores. Two implementations of the extended Rasch model, the FACETS and BIGSCALE programs, were used in the application of polychotomous scoring to the Michigan Educational Assessment Program (MEAP) spring 1990 item tryouts of the Essential Skills Mathematics Tests. FACETS produces a Rasch analysis of many-faceted data through iterative maximum likelihood estimation. The dichotomous analysis was made with BIGSCALE, which fits the one-parameter Rasch measurement model to the data. Item responses were examined for approximately 264 10th-grade students. The analysis did not show a major difference in the students exceeding the cut score under the partial credit model as compared with dichotomous scoring, but it did show that error of measurement was reduced near the cut score. The degree of misfit to the model was slightly increased for students near the cut score. Appendix A contains portions of the examination. Appendixes B, C, D, and E give analysis results. Thirty-eight graphs and 47 references are included. (SLD)
Publication Type: Reports - Research; Speeches/Meeting Papers
Education Level: N/A
Authoring Institution: N/A
Identifiers: Dichotomous Variables; Essential Skills Mathematics Tests; Michigan Educational Assessment Program; One Parameter Model; Partial Credit Model; Rasch Model
Note: Paper presented at the Annual Meeting of the American Educational Research Association (San Francisco, CA, April 20-24, 1992).