image_printPrint This Post

Across the world, religious people tend to have more children than their secular counterparts. This fact is relatively well known across the social sciences, but from an evolutionary perspective, the high fertility of religious communities is puzzling.

Humans face tradeoffs in how to best allocate the limited resources they have available to them. One critical tradeoff is between the quantity and “quality” of their offspring. The more children parents have to care for, the less time and energy parents can invest in each child. Studies have documented this tradeoff and found that family size is negatively related to a child’s cognitive and physiological development, as well as their socioeconomic success in adulthood.

We would, therefore, expect, due to differences in family sizes, children born to religious parents to fare less well in terms of developmental outcomes than children born to secular parents. Yet, while there have been few studies that compare the success of religious and secular children, the available evidence suggests that children born to religious parents fare as well as those born to secular parents. What, then, can explain the paradox of religious fertility?

In a recent article, my collaborators and I investigated this paradox using a large longitudinal database of mothers and their children from Bristol, UK, known as the Children of the 90s Health Study. In this work, we integrate two lines of research to propose and preliminarily test, a solution to the religious fertility paradox. First, studies consistently find religious people tend to be more cooperative with one another than their secular counterparts. Second, alloparenting – or the provisioning of childcare by individuals other than the parents – has been shown to both positively influence a woman’s fertility, and improve children’s health and developmental outcomes.

We hypothesized that religious cooperation extends to alloparenting and that higher levels of cooperation among religious mothers can help to explain their higher fertility. We also predicted that the support that religious mothers receive would also positively impact child developmental outcomes. In other words, we tested the idea that religious cooperation mitigates the costs of high fertility to children.

Using the Children of the 90s dataset, we first investigated whether women who invest in ritual behavior have stronger support networks. In England, and in the Avon sample, church attendance is a centrally important ritual. We assessed two forms of social support. First, we assessed general social network support based upon questions such as how many friends an individual has, and how many people they can borrow money from. Second, we assessed whether or not women received help from members of their congregation. Over the 10 years of the study, we found that church attendance was positively related to both social network support, and aid from co-religionists. Religious women have more social resources at their disposal.

To evaluate our key predictions, we next investigated whether or not these two forms of support were related to a woman’s fertility. Unexpectedly, we found a negative relationship between social network support and a woman’s fertility. Aid from co-religionists, though, was positively related to a woman’s fertility, as expected. Women who get help from members of their congregation have more children.

Over the 10 years of data we analyzed, though, a woman’s social network support decreased, while aid from co-religionists remained consistent. Religious mothers appear to remain social as their children age, while secular women’s social networks become smaller. This suggests that while most mothers’ social networks faded over time, religious support networks remained strong.

Next, we examined whether these two forms of social support were related to a child’s physiological and cognitive development. In terms of physiological development, the Children of the 90s health data have been used to show that children with more siblings are shorter. Our models recovered this effect, but we didn’t find evidence for relationships between a mother’s social network support, or aid from co-religionists, and a child’s height.

In terms of cognitive development, and again replicating previous research, we found that the more siblings a child had, the lower they scored on three cognitive tests. These tests were administered to children when they entered school (aged 4-5), one year later (aged 5-6), and when they were 8. These kinds of tests are known to be biased, and our models reveal evidence of these biases – the children of wealthier and better-educated mothers scored higher on these tests. We found, though, that social support and aid from co-religionists were both associated with higher child test scores, particularly at later stages of development. This suggests that women’s social networks positively affect child cognitive development, at least in terms of the way that cognitive development is assessed on these tests.

These results only support some of our hypotheses. Yet our findings are mostly consistent with the idea that religions in modern environments support cooperative breeding strategies: women who receive help from members of their congregation have higher fertility, and this aid, as well as more general forms of social support, were both associated with improved child cognitive development. By positively influencing social support, religion in the UK may help some women have more children, without sacrificing the success of these children.

Still, many open questions remain, and we were unable to evaluate alternative hypotheses, such as whether differences in fertility between religious and secular groups are affected by differences in pro-natal and gender norms. We also don’t know whether these findings will generalize to non-Christian and/or non-Western contexts. We are currently conducting a large cross-cultural study in Bangladesh, The Gambia, India, Malawi, and the United States to better investigate these issues. In doing so, our larger project hopes to advance our understanding of the interrelationship between religion, fertility, and child development. Researching the evolutionary dynamics surrounding religion’s influence on family size and child success is not just of interest to the scholarly community. Due to its relevance for economic and social development, health, and demographic projections, we expect our project will be of significant interest to governments, NGOs, and public policy officials. We’ll keep you updated on our findings.


Shaver, J.H., Power, E., Purzycki, B., Watts, J., Sosis, R., Sear, R., & Shenk, M. (2020). Church attendance and alloparenting: An analysis of fertility, social support, and child development among English mothers. Philosophical Transactions of the Royal Society, Biological Sciences.

Shaver, J. H., Sibley, C. G., Sosis, R., Galbraith, D., & Bulbulia, J. (2019). Alloparenting and religious fertility: A test of the religious alloparenting hypothesis. Evolution and Human Behavior40(3), 315-324.

Published On: June 29, 2020

John Shaver

John Shaver

image_printPrint This Post

John Shaver is Senior Lecturer and Head of the Religion Programme at the University of Otago. He is an evolutionary anthropologist whose work explores the complex relationships between religion, cooperation, and social inequality. He is also particularly interested in understanding cross-cultural and intra-cultural variation in fertility, and the impacts of family size on child development. To explore these issues, he has conducted research in the Czech Republic, Fiji, Mauritius, New Zealand and the United States.

One Comment

  • Andrew Atkinson says:


    Have you ever thought, that there’s an argument to be made about religious fertility being ultimately maladaptive at the species level? Like the antlers of some deer (which clearly evolved too) the size of a group can become too big to support. I’d be interested to know your thoughts.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.