Literaturnachweis - Detailanzeige
Autor/inn/en | Kirisci, Levent; Hsu, Tse-Chi |
---|---|
Titel | The Robustness of BILOG to Violations of the Assumptions of Unidimensionality of Test Items and Normality of Ability Distribution. |
Quelle | (1995), (29 Seiten)
PDF als Volltext |
Sprache | englisch |
Dokumenttyp | gedruckt; online; Monographie |
Schlagwörter | Ability; Comparative Analysis; Correlation; Difficulty Level; Estimation (Mathematics); Guessing (Tests); Robustness (Statistics); Simulation; Statistical Distributions; Test Items |
Abstract | The main goal of this study was to assess how sensitive unidimensional parameter estimates derived from BILOG were when the unidimensionality assumption was violated and the underlying ability distribution was not multivariate normal. A multidimensional three-parameter logistic distribution that was a straightforward generalization of the unidimensional three-parameter logistic distribution was used as a data generation model for simulations. The study found that test dimensionality had a significant effect on the estimation of item parameters and ability. The BILOG program was not robust against the violation of the underlying ability distribution in estimating the item guessing parameter and ability for a unidimensional or multidimensional test. If dimensions of ability were correlated, and a test was multidimensional, BILOG yielded more robust estimates of item difficulty, guessing, and ability. The shape of the ability distribution affected only the estimation of ability. Comparison with the ASCAL and MULTILOG programs suggested that no program performed better than the others in estimating item and ability parameters across all test dimensions, degrees of intercorrelations of ability dimensions, and all ability distributions. Seven tables and seven figures illustrate the discussion. (Contains 10 reference.) (SLD) |
Erfasst von | ERIC (Education Resources Information Center), Washington, DC |