To investigate, Daws and Hampshire surveyed more than 63,000 people online, and had them complete a 30-minute set of 12 cognitive tasks that measured planning, reasoning, attention and working memory. The participants also indicated whether they were religious, agnostic or atheist.
As predicted, the atheists performed better overall than the religious participants, even after controlling for demographic factors like age and education. Agnostics tended to place between atheists and believers on all tasks. In fact, strength of religious conviction correlated with poorer cognitive performance. However, while the religious respondents performed worse overall on tasks that required reasoning, there were only very small differences in working memory.