A Multimodal Item Response Modeling for Personality Assessment in Organizational Research

Recent advances in process data collection have made it possible to efficiently collect multimodal behavioral indicators, such as response times and eye-tracking measures. These multimodal data have been widely applied in cognitive and achievement assessments, where they have improved the accuracy of latent construct estimation. However, the use of informative multimodal process data in noncognitive assessments, such as personality measures widely used in organizational research, has received considerably less attention. To address this gap, we integrate response time and eye-tracking data into a conventional item response model to capture respondents’ response processes, thereby improving differentiation across trait levels and enhancing noncognitive assessment. Simulation studies were conducted to evaluate the performance of the proposed model and compare it with a conventional IRT model. Results indicate that model parameters can be accurately recovered and that incorporating multimodal data significantly improves the accuracy of person latent trait estimates. Finally, an empirical analysis was conducted to demonstrate the applicability and advantages of the proposed model in personality assessment.