CHIDIM - A FORTRAN PROGRAM FOR ASSESSING THE DIMENSIONALITY OF BINARYITEM RESPONSES BASED ON MCDONALDS NONLINEAR FACTOR-ANALYTIC MODEL

The assumption of unidimensionality is central to most item response t heory (IRT) models. procedures based on McDonald's nonlinear factor an alytic model and accompanying principle of weak local independence hav e proven to be useful with respect to identifying the number of dimens ions underlying a set of binary item responses. In particular, an appr oximate chi-square statistic, which tests the null hypothesis that the off-diagonal elements of a residual correlation matrix are equal to z ero after fitting an m-factor model to a set of item responses, was sh own to be accurate in identifying the true number of dimensions underl ying various simulated test structures. This article briefly describes CHIDIM, a FORTRAN computer program that calculates the approximate ch i-square statistic. The use of the program and the interpretation of t he results are illustrated through a simulated data set.