What Affects the Quality of Score Transformations? Potential Issues in True-Score Equating Using the Partial Credit Model

This simulation study investigated to what extent departures from construct similarity as well differences in the difficulty and targeting of scales impact score transformation when are equated by means concurrent calibration using partial credit model with a common person design. Practical implications results discussed focus on scale equating health-related research settings. The simulated data for two scales, varying number items sample sizes. factor correlation between was used operationalize similarity. Targeting operationalized through increasing departure equal dispersion item parameters each scale. show that low goes along lower precision. In cases levels similarity, precision improves settings where range is encompassing range. With decreasing benefits more good targeting. Difficulty shifts up logits somewhat increased estimation bias but without affecting observed robustness against supports advantage applying true-score methods over identity equating, which naive baseline method comparison. Finally, larger size did not improve this study, longer improved only marginally quality equating. insights real-data example.