There is a reliably replicated negative correlation between intelligence and religiosity: on average, more intelligent people are less religious. The figure usually cited comes from a meta-analysis of 63 studies (Zuckerman, Silberman & Hall, 2013), and the pattern has held across larger meta-analyses since, though the effect size shrinks under stricter methods.
Three explanations get offered, and they aren’t equally strong. The first is non-conformity: intelligent people are less inclined to hold a belief simply because it’s the socially endorsed default, and inherited religion is exactly that kind of default. The second is analytical over intuitive thinking, which correlates with disbelief but is the shakiest leg, because the causal version, where prompting analytic thought reduces belief, largely failed to replicate. Treat it as association, not a lever.
The third is the interesting one. What religion provides (a way to cope with death and uncertainty, a template for self-discipline, a source of self-worth) is also supplied by higher intelligence through other routes, so the religious version is less needed. This is the explanation with real weight, and it carries an uncomfortable implication: if intelligence substitutes for religion’s functions, then stripping religion from someone who lacks those substitutes is subtraction, not liberation.
References
Zuckerman, Silberman & Hall (2013). “The Relation Between Intelligence and Religiosity: A Meta-Analysis and Some Proposed Explanations.” Personality and Social Psychology Review. DOI: 10.1177/1088868313497266. The 63-study meta-analysis; originates the functional substitution hypothesis.
Dürlinger & Pietschnig (2022). “Meta-analyzing Intelligence and Religiosity Associations: Evidence from the Multiverse.” PLOS ONE. DOI: 10.1371/journal.pone.0262699. 89 studies, N = 201,457; r = −.14, a smaller but robust negative relation.
Sanchez et al. (2017). Large preregistered replication finding little causal effect of analytic priming on belief (d ≈ 0.07); the reason the analytic-thinking explanation should be stated as correlation only.