ERIC Number: ED189174
Record Type: RIE
Publication Date: 1980-Apr
Reference Count: 0
An Analysis of Rasch Person Fit Statistics in College Placement Tests.
Smith, Richard M.; Mitchell, Virginia P.
To analyze test item and person interactions according to the Rasch Model, 822 incoming community college freshman were administered the New Jersey Basic Skills Placement Test. The Logical Relationships and Elementary Algebra subtests were chosen to examine the effect of differing average item difficulty on descriptions of fit statistics and on frequency of person misfit. Three user-defined item groups: sequence, subscores, and difficulty, were chosen to highlight three misfit statistics: working too slowly to finish test, misunderstanding content, and guessing. A fourth misfit statistic, total t test, was also calculated. On the four misfit statistics, the percentage of misfitting students was 25% (Logical Relationships) and 33% (Elementary Algebra). After editing techniques were applied to the item response strings, the algebra subtest became valid for 90% of the misfitting students. The appendix contains samples of the PANAL computer program output for typical cases of misfit, as well as editing suggestions for producing meaningful ability estimates. (CP)
Publication Type: Speeches/Meeting Papers; Reports - Research
Education Level: N/A
Authoring Institution: N/A
Identifiers: New Jersey College Basic Skills Placement Test; Rasch Model
Note: Paper presented at the Annual Meeting of the National Council on Measurement in Education. (Boston, MA, April 8-10, 1980).