Firestar: Computerized Adaptive Testing Simulation Program for Polytomous Item Response Theory Models

Publisher Website
Google Scholar
Abstract
Polytomous items are used extensively in the assessment of patient-reported outcomes (PROs). PRO measures are increasingly administered using computer adaptive testing (CAT) platforms (Cella, Gershon, Lai, & Choi, 2007). Post hoc CAT simulations are con- ducted occasionally as a means to evaluate item pools, select items for static short forms, determine optimal CAT settings, compare measurement efficiency of CAT to that of static measures, and so on. Firestar is a computer program designed for simulating CAT with poly- tomous items (two-category items are supported as special cases). The item response theory models supported by the program include Samejima's (1969) graded response model and Muraki's (1992) generalized partial credit model (GPCM). Both Masters's (1982) partial credit model (PCM) and Andrich's (1978) rating scale model are also supported as special cases of the GPCM.