A Practical Guide to Expressing Psychological Theories in Evidence Accumulation Models

Evidence accumulation models (EAMs) explain and predict human choices and response times (RTs). Their parameters can be estimated based on observed choice-RT distributions. Contrasting accumulation-model parameters across experimental design cells can be a powerful way to test theories of the latent cognitive processes underlying choice and RT. For example, the prospective memory (PM) decision control model instantiated a novel cognitive-control theory of PM decisions and quantitatively compared it to existing theories. This tutorial demonstrates how to embed such psychological theories in EAM parameters using the R package EMC2, which simplifies model programming and increases computational efficiency beyond previous approaches. We first show how a broad class of cognitive models can be expressed by mapping EAM parameters to experimental conditions using an augmented linear model language. We then walk through two examples of how to specify, fit and evaluate such models. The first is a standard application of the PM decision control model. The second showcases how the model can be extended to address a more complex paradigm where an automated decision aid provides advice about choices. We conclude by discussing how this approach can be applied more broadly.