Modeling Item Responses with Response Times as Collateral Information in Large-Scale Educational Assessments

In large-scale educational assessments, it is becoming common practice to collect the item response times next to the actual item responses. These response times have many uses in large scale assessments, including data quality checks, checks on technical problems in data recording, and checks on aberrant response processes like guessing, careless responding, and responding with item preknowledge. In the present chapter, we focus on response times as collateral information concerning individual differences in the ability underlying the educational domains of an assessment. The main aim is to improve the reliability of large-scale educational assessments. We discuss various psychometric models that are available to simultaneously analyze responses and response times, we illustrate the added value of the response times to the reliability of the 2015 PISA cycle, and we discuss the current challenges and future directions of response time research in large-scale educational assessments.