Nonparametric Item Response Theory: A Maturing and Applicable Measurement Modeling Approach

Publisher Website
Google Scholar
Abstract
The papers in this Special Issue are briefly discussed. The practical applicability of nonparametric item response theory (NIRT), NIRT's capacity to address fundamental modeling issues and to function effectively with smaller sample sizes than parametric item response theory (PIRT), along with how NIRT addresses lack-of-fit issues, are emphasized. NIRT assesses latent multidimensionality using item pair conditional covariances and other approaches. This and the relation of latent multidimensionality analyses to construct validity are discussed. The need for NIRT analyses to provide analogues for users to the parameters supplied by PIRT analyses, is stressed. The interaction between fundamental order relationships and polytomous NIRT is discussed. It is concluded that NIRT is an important area of research that should continue to be emphasized in the future.