Bayesian Analysis of Joint Modeling Response Times with Dynamic Latent Ability in Educational Testing

• Activity Number: 589
• Type: Topic Contributed
• Date/Time: Wednesday, August 12, 2015 : 2:00 PM to 3:50 PM
• Sponsor: Section on Bayesian Statistical Science
• Author(s): Xiaojing Wang* and Abhisek Saha and Dipak K. Dey
• Companies: University of Connecticut and University of Connecticut and University of Connecticut
• Keywords: Item Response Theory ; Longitudinal Data ; Joint Modeling ; State Space Models ; Markov Chain Monte Carlo ; Latent Traits

Abstract

In educational measurement testing, inferences about latent traits of test takers have been mainly based on their responses to the items while the time taken to complete an item has been often ignored. Luckily, in computerized testing, the response time of items can be collected at no additional cost. In this paper, a new class of state space models in joint modeling response time with time series dichotomous responses will be put forward, where item response time is used as auxiliary information for the response accuracy to improve the precision for the estimates of latent traits. The models can be applied either retrospectively to the full data or on-line, in cases where real-time prediction is needed. The models are studied through simulated examples and applied to a large collection of reading test data obtained from MetaMetrics, Inc.