The SRM-G: Capturing Cognitive Dynamics in Interactive Problem-Solving Tasks

• In Event: Innovations in Validity Evaluation and Measurement Development

Abstract

Interactive problem-solving assessments allow examinees to learn from trial-and-error, potentially changing their response patterns when revisiting problem states due to learning or fatigue effects. This study introduces the Sequential Response Model with Growth parameters (SRM-G), which captures how examinees' cognitive processes evolve through accumulated experience during task interaction via growth parameters. Monte Carlo simulations across varying sample sizes and sequence lengths demonstrate superior model fit and parameter recovery when experience effects are present. An empirical application to collaborative problem-solving assessment reveals significant positive effects in intermediate states (mastering task mechanisms) and negative effects at the initial state (accumulated frustration), demonstrating the model's ability to provide nuanced ability estimates while detecting meaningful cognitive dynamics.

Authors

• Yuting Han, Beijing Language and Culture University

• Hongyun Liu, Beijing Normal University

• PUJUE WANG

• Feng Ji, University of Toronto