Proficiency classification and violated local independence
An examination of pass/fail decision accuracy under competing rasch models
The purpose of this study was to examine the use of a misspecified calibration model and its impact on proficiency classification. Monte Carlo simulation methods were employed to compare competing models when the true structure of the data is known (i.e., testlet conditions). The conditions used in the design (e.g., number of items, testlet to item ratio, testlet variance, proportion of items that are testlet-based and sample size) reflect those found in the applied educational literature. Decision Consistency (DC) was high between the models, ranging from 91.5% to 100%. Testlet variance had the greatest effect on DC. An empirical example using PISA data with nine testlets is also provided for the consistency of pass/ fail decisions between the competing models.