Factors Affecting Accuracy of Comparable Scores for Augmented Tests under Common Core Standards

This paper examined various factors affecting linking accuracy of the augmented tests containing both common assessments and state-specific item sets in the context of the Common Core State Standards (CCSS). Linking was conducted using the common-item nonequivalent groups design with pseudo and simulated data. The factors included group ability differences, differential effect sizes, latent-trait correlations, differential latent-trait correlations between the common assessment and state-specific item set, and different equating methods (classical and IRT equating). Higher latent-trait correlation between the common assessment and state-specific item set was associated with smaller bias, and if the latent-trait correlation exceeded 0.8, IRT equating methods provided adequate linking even though the group ability difference was large and the differential effect size was moderately large. The same latent-trait correlations for the old and new tests resulted in smaller bias than differential latent-trait correlations for the old and new tests.