Ability Parameter Recovery in Virtual Learning Environment Item Response Data With the Sequential Two-Parameter Logistic Item Response Theory (2PL-IRT) Model

• In Event: Psychometrics and Educational Measurement Poster Session 1

Abstract

Virtual learning environments (VLE) often allow students to practice test items at their own discretion, resulting in data that contains a mix of scores from various numbers of item attempts and omitted items across a large sample of students. Estimating student ability from data such as this can be challenging. This study evaluates the recovery of ability parameters in VLE data with multiple item attempts and missing data with the sequential 2PL-IRT model. We conduct a simulation study based on operational data, evaluating ability estimate bias and RMSE under various conditions. Results can guide researchers toward appropriately using the sequential 2PL-IRT model to estimate ability in VLE data.

Authors

• Ziying Li, University of Florida

• Corinne Huggins-Manley, University of Florida

• M. David Miller, University of Florida

• Walter L. Leite, University of Florida