A Dual-Process Diffusion Model

This paper presents a simple formal analytical model delivering qualitative predictions for response times in binary-choice experiments. It combines a dual-process/multi-strategy approach with the standard diffusion model, modeling a utility decision process and a heuristic decision process as diffusion processes of evidence accumulation. For experiments with objective alternatives (including many tasks in judgment and decision making), the model predicts that errors will be quicker than correct responses in case of process conflict and slower in case of alignment, capturing a well-documented asymmetry regarding slow or fast errors. Further, the model also predicts that correct responses are slower in case of conflict than in case of alignment, capturing the well-known Stroop effect. The model is also extended to cover experiments with subjective alternative evaluations, that is, preferential choice. In this case, results depend on whether trials are hard or easy, that is, on whether the heuristic can be interpreted as relatively automatic or not.