Thou shall not kill: An analysis of religion, violence, and IQ

Like religion, higher IQ is associated with lower rates of antisocial behavior and crime.

Illustration by Lesley Becker/GlobeStaff; Adobe
Illustration by Lesley Becker/GlobeStaff; AdobeIllustration by Lesley Becker/Globe Staff; Adobe

Editor’s note: The research this article was based on has been retracted by the authors, and thus the findings reported in this article may no longer be valid. You may find the retraction notice here: https://www.psychologicalscience.org/publications/psychological_science/clark-2020-retraction-editorial

Throughout history artists, philosophers, and poets have contrasted human depravity with the moral perfection of God, the only force that could rein in the tempestuous impulses of humans. Many scholars have agreed, arguing that human morality depends on the moral codes, judgmental deities, and communal rituals provided by religions — leading some to warn that increasing irreligiousness will erode community and cooperation, breeding violence and instability.


Skeptics, however, rightly point out that today’s low ebb of religious participation has coincided with the most peaceful moment in history. Perhaps the hand-wringing about secularization was wrong. Or is the story more complicated?

Modern data and more careful analyses suggest the latter.

Whereas much social scientific research demonstrates that religion can indeed make people more helpful and less selfish, other data show that many of the most peaceful societies are also the least religious. These patterns suggest that certain cultural characteristics may influence how advantageous religion is for regulating antisocial behavior. One possible factor is cognitive ability.

Religious belief has declined especially among advanced industrialized societies comprising highly educated and intelligent citizens. Like religion, higher IQ is associated with lower rates of antisocial behavior and crime, and so groups with relatively higher average IQs might have other means of maintaining and enhancing cooperation without the vivid moral lessons and strict rules of traditional religions. In a paper forthcoming in Psychological Science, we tested whether lower levels of religious participation within populations would be associated with higher levels of homicidal violence, but not in countries with relatively high average IQs. Two studies supported this hypothesis.

First, a longitudinal analysis with 176 countries found that declines in religious participation from 1945 to 2010 were associated with concurrent increases in homicide rates — but not in countries with relatively high average IQ. We then confirmed this pattern using modern data from 195 countries, finding that lower religious participation was strongly associated with higher homicide rates in nations with relatively low average IQ, but the relationship was weak to nonexistent in countries with relatively higher average IQ. These patterns held when using three different measures of country-level IQ, three different measures of country-level religious participation, and controlling for other related factors such as wealth, inequality, and educational differences between countries.


These findings suggest that religions may reduce violent antisocial behavior more among populations with lower average cognitive ability than among societies with more cognitively gifted citizens. All large societies have had to construct cultural solutions — such as laws and justice systems — to manage the more violent and selfish tendencies of their people. Religions likely have also served this function and continue to do so, at least in many countries across the globe. These new findings suggest that today, as countries have become more cognitively sophisticated, they may have unlocked additional routes to managing humankind’s more violent tendencies — without religion.

Religion can be a controversial topic. Some nonbelievers are reluctant to admit that anything good can come of religions — seeing them instead as great delusions that poison everything. Some adherents see their religion as a divinely inspired tradition that produces only good, with any negative consequences dismissed as unrelated or the fault of stubborn nonbelievers. Yet, social science has demonstrated that religions produce both costs and benefits. Our findings suggest that the balance of these costs and benefits might depend on the cultural context and the people who observe (or do not observe) these religious traditions.


Like all scientific studies, ours is not without caveats. First, though our theorizing and longitudinal analysis imply a causal relationship, our data are correlational. Other scholars may discover more plausible or parsimonious causal stories for the patterns we observed, though we did aim to statistically rule out some obvious alternative explanations (e.g., the effect does not appear to be driven by wealth differences). Second, none of these country-level differences — in religious participation, violence, or IQ — is static. Our analysis applies to the past 70-75 years, but these patterns may change in the future as societies continue to advance and develop, create new technologies and cultural solutions, and globalize. Finally, these data do not suggest that declines in religious participation are inevitably harmful in particular societies. Declines in religious participation are associated with a variety of costs and benefits, and sometimes the benefits (e.g., promoting individual rights) will outweigh the costs. And other secular norms and institutions may be able to replace religion and supply similar violence-deterring benefits.

Humans, though flawed and prone to selfishness, have created many effective cultural norms and narratives, religion among them, that improve their societies and enhance their quality of life. Religion is by no means the only — or even the best — force to tame people’s worst impulses, but it may yet serve some violence-deterring value in many societies across the globe.


Cory Clark is an assistant professor of psychology at Durham University in the United Kingdom. Bo Winegard is an assistant professor of psychology at Marietta College in Ohio. Azim Shariff is an associate professor of psychology at the University of British Columbia.