ABSTRACT
This chapter summarizes the research on using response time data to detect various forms of test cheating. It explores statistical procedures for identifying aberrant or irregular response behavior by the test takers, which requires response times for each item, examinee responses to each item, and item parameters. The chapter describes Bayesian procedures for identifying aberrant responses and response time patterns in testing for examinees that test online. This approach uses a combination of a regular response model and a response time (RT) model in a hierarchical framework that allows us to use the collateral information in the responses to make the detection of aberrant response times more powerful, and the other way around. The chapter presents two approaches in RT modeling: the lognormal RT model as defined by W. J. van der Linden and a hierarchical RT model as presented by van der Linden and F. Guo.
