A Dynamic Dual Process Model for Binary Choices: Serial Versus Parallel Architecture

Abstract

AbstractDual process theories have become increasingly popular in psychology, behavioral economics, and neuroscience, assuming that two processes, here generically labeled as System 1 and System 2, have antagonistic characteristics such as automatic versus deliberate, impulsive versus rational, fast versus slow, and more. In decision-making a choice results from an interplay of these two systems. However, most existent dual-process approaches are merely verbal descriptions without providing the means of rigorous testing. The prescribed dynamic dual process model framework is based on stochastic processes and produces testable qualitative and quantitative predictions. In particular, it makes precise predictions regarding choice probability, response time distributions, and the interrelation between these quantities. The focus of the present paper is on the architecture of the two postulated systems: serial versus parallel processing. Using simulation studies, I illustrate how different factors (timing of System 1, time constraint, and architecture) influence model predictions for binary choice situations. The serial and 6 parallel processing versions of the framework are fitted to published data.