Evolutionary modules and Bayesian facilitation: The role of general cognitive resources

Abstract

Although decisions based on uncertain events are critical in everyday life, people perform remarkably badly when reasoning with probabilistic information. A well-documented example is performance on Bayesian reasoning problems, where people fail to take into account the base-rate. However, framing these problems as frequencies improves performance spectacularly. Popular evolutionary theories have explained this facilitation by positing a specialised module that automatically operates on natural frequencies. Here we test the key prediction from these accounts, namely that the performance of the module functions independently from general-purpose reasoning mechanisms. In three experiments we examined the relationship between cognitive capacity and performance on Bayesian reasoning tasks in various question formats, and experimentally manipulated cognitive resources in a dual task paradigm. Results consistently indicated that performance on classical Bayesian reasoning tasks depends on participants’ available general cognitive capacity. Findings challenge the postulation of an automatically operating frequency module.