Modeling for Directly Setting Theory-Based Performance Levels

Abstract

This paper presents the Latent Class Level-PCM as a method for identifying and interpretinglatent classes of respondents according to empirically estimated performance levels. The model,which combines elements from latent class models and reparameterized partial credit modelsfor polytomous data, can simultaneously (a) identify empirical boundaries between performancelevels and (b) estimate an empirical location of the centroid of each level. This provides moredetailed information for establishing performance levels and interpreting student performancein the context of these levels. The paper demonstrates the use of the Latent Class L-PCM on anassessment of student reading proficiency for which there are strong ties between the hypothesizedtheoretical levels and the polytomously scored assessment data. Graphical methods for evaluatingthe estimated levels are illustrated