DOCUMENT RESUME ED 344 940 TM 018 335 AUTHOR Chang, Yu-Wen; Davison, Nark L. TITLE A Comparison of Unidimensional and Multidimensional IRT Approaches to Test Information in a Test Battery. PUB DATE Apr 92 NOTE 21p.; Paper presented at the Annual Meeting of the American Educational Research Association (San Francisco, CA, April 20-24, 1992). PUB TYPE Reports - Evaluative/Feasibility (141) -- Speeches/Conference Papers (150) EDRS PRICE MF01/PC01 Plus Postage. DESCRIPTORS *Comparative Testing; Computer Simulation; Correlation; *Error of Measurement; *Estimation (Mathematics); *Item Response Theory; Mathematical Models; Research Design; Sample Size; Scores; *Test Bias IDENTIFIERS Ability Estimates; *Multidimensional Models; Test Batteries; *Unidimensional Models ABSTRACT Standard errors and bias of unidimensional and multidimensional ability estimates were compared in a factorial, simulation design with two item response theory (IRT) approaches, two levels of test correlation (0.42 and 0.53), two sample sizes (500 and 1,000), and a hierarchical test content structure. Bias and standard errors of subtest scores (subtests within longer tests) were smaller for the multidimensional approach, but the bias of longer test scores did not improve. Sample size and test intercorrelation had a negligible effect on the results. As the number of items in a test becomes large, in an absolute sense, and relati:', to the total number in the battery, the advantage of the multidimensional approach seems to disappear. Two tables and 13 figures present results of the simulation. There is an 11-item list of references. (SLD) itgititor******************************************************************* * Reproductions supplied by EDRS are the best that can te made * * from the original document. * ***********************************************************************