The Evolution of the Evolutionary Demography of Religion Project

By John Shaver

When I was a graduate student at UConn, my advisor, Rich Sosis, invited Eric Kaufmann to campus to give a talk about his new book, Shall the Religious Inherit the Earth? Demography and Politics in the Twenty-First Century. In his talk, which focused on only a portion of the book, Kauffman argued that despite widespread shifts away from religion in the United States, religion was likely to grow in the foreseeable future. This is because, he suggested, religious people have many more children than secular people. His models predicted that in the United States, by the mid- to late-Twenty-First Century, Catholics and Muslims would far outnumber secular people and be two of the largest social groups.

Although I’d had some exposure to life history theory as an undergrad, at roughly the same time as Kaufmann’s talk, I was enrolled in Rich’s behavioral ecology class where life history theory was a major focus. I was also reading Hill and Hurtado’s Ache Life History for another class on anthropological theory. During Kaufmann’s talk, which was not guided by evolutionary theory, it became immediately apparent that, from a life history perspective, the high fertility of some religious groups poses an evolutionary problem: parental resources are critical to child outcomes and child success, but they are also finite. Since each additional child requires more resources, there is a tradeoff between child number and child success. If religious people have more children on average, then these children should be of “lower quality” (to use the parlance of life history theory) relative to secular children. Or, perhaps there was something unique to religious groups that helps them to mitigate the tradeoff between child number and child success? Kaufmann’s models were missing something, and indeed, the relationships between religion and fertility had not yet been rigorously subjected to evolutionary analyses.

Rich’s historical work examining the survivorship of 19th Century American communes, as well as his fieldwork on Israeli kibbutzim had shown that religious people, compared to their secular counterparts, tend to be more cooperative with one another. In discussions after Kaufmann’s talk, Rich and I discussed the evolutionary problem of religious fertility and speculated that perhaps religious cooperation extends to childcare and could help to explain the differences in fertility between secular and religious groups. I was off shortly to do fieldwork in Fiji and it was clear that in order to test these ideas, we would need large samples, much larger than the village where I was conducting my PhD research. Any further thought or work in the area would have to wait.

After a postdoc in the Czech Republic, I was lucky enough to land another postdoc to work with Joseph Bulbulia on the New Zealand Attitudes and Values Study (NZAVS) at Victoria University of Wellington in Aotearoa New Zealand. The NZAVS, started in 2009, is a longitudinal study of now over 60,000 New Zealanders. While I was in Wellington, the journal Religion, Brain & Behavior had a call for a special issue, in which scholars were invited to submit “Hilbert Problems,” and their potential solutions. These Hilbert Problems were to be the largest contemporary problems facing the evolutionary and scientific study of religion.

I took this as an opportunity to think more about those issues that were first raised for me after Kaufmann’s talk. My submission detailed the evolutionary problem of high fertility among religious groups (and called it the paradox of religious fertility). The very brief article also offered a potential solution: religious cooperation extends to alloparenting and that higher levels of cooperation among religious mothers can help to explain their higher fertility, as well as positively impact their children’s development.

But these ideas remained underdeveloped and unevaluated. Rich, Joseph, a few others and I began to analyze NZAVS data to test aspects of the model. The NZAVS asks people several demographic questions as well as questions about how often they engage in childcare each week (and many, many more other questions). We found that among 12,980 New Zealanders, ethnic groups with higher fertility, specifically Māori and Pasifika, were more likely to alloparent than ethnic groups with lower fertility. After adjusting for ethnic group differences, moreover, we found that religious adults had higher fertility and were more likely to alloparent than their secular counterparts. These results were largely consistent with the solution proposed in the Hilbert Problems special issue. But, many questions remained, and these questions required fieldwork for answers.

At an International Association for the Cognitive Science of Religion Conference in Vancouver, I presented the results of the NZAVS data. Afterwards, on the way to dinner, I asked Nick Gibson, program officer at the John Templeton Foundation (JTF), if he thought JTF would be interested in funding a project to investigate these issues in Fiji. He suggested that I think “much bigger,” and consider developing a cross-cultural research project. He also suggested that I submit a planning grant application and propose to invite advisors to collaboratively develop a full research proposal.

Luckily, I received that planning grant from JTF. I used this as an opportunity to meet with and get advice from scholars whose work I respected at a two-day workshop in 2018. Among the ten participants at the workshop were Rich, Rebecca Sear, and Mary Shenk. At the meeting I proposed some preliminary ideas for a research design, including additional secondary analyses and some rough ideas for a large cross-cultural study. After the meeting Rebecca and Mary suggested that the project would benefit from working with Health and Demographic Surveillance Systems (HDSS) for the cross-cultural portion of the project. HDSSs are longstanding research infrastructures across the world that collect data from entire populations within fairly large geographic areas. Researchers can collaborate with HDSSs to draw from their existing data, develop sampling frames, and work with their local research teams to collect novel data. Both Rebecca (in The Gambia) and Mary (in Bangladesh) had worked with HDSSs in the past.

Given Rich’s work in the area, Rebecca and Mary’s demographic research, and history with HDSSs, it made sense for the four of us to combine efforts and submit a large JTF grant together. In 2019 our project was funded. Unfortunately COVID-19 put fieldwork on hold for a while. But while primary data collection was delayed, we wrote up additional analyses of pre-existing data as well as analyses from online data we collected for this project. Thanks to our fantastic postdocs and HDSS collaborators, we were finally able to begin fieldwork in 2021 (in Bangladesh, The Gambia, India, Malawi, and the United States). Now in July 2022, fieldwork is nearly complete. And by the end of this year, we can finally evaluate some of those ideas that first emerged after a department seminar at UConn about 12 years ago.