A Poisson Race Model Analysis of the Implicit Association Test

The Implicit Association Test (IAT) is a computerized two-choice discrimination task in which stimuli have to be categorized as belonging to target concepts (e.g., “flowers” and “insects”) or attribute concepts (e.g. “positive” and “negative”) by pressing, as quickly and accurately as possible, one of two response keys. In test blocks, target and attribute concepts are mapped onto the same response keys in different combinations. A Poisson race model analysis of the IAT is presented. Four parallel and independent Poisson processes, each of which concerns a specific concept, have been considered. Each response is associated with two processes. Information related to specific characteristics of the stimulus that has been displayed accumulates on the counter of each process. Model parameters are the rates of incoming information to the counters, and the thresholds of the counters. The process that reaches its threshold first wins and determines the response. Both latency and accuracy of the responses are taken into account. Latencies are determined by the time at which a process wins. The accuracy depends on the winning process, the category of the displayed stimulus, and the test block. The model parameters separate automatic and controlled processes involved in the IAT effect. Results of an empirical application are presented and discussed. Comparisons with other models for the analysis of the IAT are presented as well.