This award will provide funding to support her project, "Multiple Group Joint Modeling of Item Responses, Response Times, and Action Counts in Collaborative Problem-Solving Assessments."
Abstract
Measurement invariance is an important fairness issue. Multiple group modeling is one of the methods to address the measurement variance issue. Traditional psychometric models focus on item responses to estimate response accuracy. However, including response times in the joint modeling with item responses helps estimate students’ working speed as well as improve the measurement precision of response accuracy. In computer-based assessments, the number of actions (i.e., action count) is a good indicator of student engagement. In addition, action counts, as additional auxiliary information, can further enhance the measurement precision of response accuracy. This study proposes multiple group joint modeling of item responses, response times, and action counts to address student engagement and measurement invariance issues in collaborative problem-solving assessments. Specifically, the proposed model is used to address the fairness issue between socioeconomic status groups. Simulation studies are conducted to evaluate the model parameter recovery. The impact of ignoring the multiple group structure is also examined by comparing the proposed model with the single group model. Empirical data analysis is carried out to illustrate an application of the proposed model using the Programme for International Student Assessment (PISA) 2015 collaborative problem-solving items. The model- data fit of the multiple group model and the single group model is evaluated. In the meantime, the item parameter invariance between different socioeconomic status groups is investigated.
