Across industrialized societies, religious individuals exhibit higher fertility than their secular peers. The higher fertility of religious people has been associated with various indicators of religiosity, including attendance at religious services (Adsera, 2006; Berghammer, 2012; Frejka & Westoff, 2008; Peri-Rotem, 2016; Philipov & Berghammer, 2007; Shaver et al., 2020; Shaver, Sibley, Sosis, Galbraith, & Bulbulia, 2019), self-reported importance of religion (Baudin, 2015; Frejka & Westoff, 2008), or even simply declaring a religious affiliation (Frejka & Westoff, 2008; Peri-Rotem, 2016; Philipov & Berghammer, 2007; Shaver et al., 2019). While the magnitude of fertility differences between religious and less religious individuals varies spatially, temporally, and across denominations (Peri-Rotem, 2016; Schoonheim & Hülsken, 2011), the effect has been consistently noted in industrialized nations, particularly in the US and the UK (Goldscheider, 2006; Rowthorn, 2011).

Evolutionary researchers hypothesize that larger family sizes may lead to lower socioeconomic and educational outcomes for individual children raised in these families. This is because parents have finite energetic resources, theoretically resulting in lower parental investment per child in larger families (Lawson and Mace, 2009, Lawson and Mace, 2011), or in other words in a tradeoff between the quality and the quantity of offspring (Walker, Gurven, Burger, & Hamilton, 2007). This “resource dilution” effect has been used to explain findings that children from larger families may be slightly physically smaller (Hagen, Barrett, & Price, 2006; Öberg, 2015), have shorter lifespans (Lynch, 2016), and lower education attainment (Downey, 1995) for example, though such trade-offs are not ubiquitous across empirical analyses (Lawson & Borgerhoff Mulder, 2016; Lawson & Mace, 2011). The “quantity-quality” trade-off may be particularly strong in high-income, low-fertility settings where parents perceive high parental investments are necessary for child success (Hill & Kaplan, 1999; Walker et al., 2007; Lawson, Alvergne, & Gibson, 2012; Lynch, 2016).

Despite these expected quantity-quality tradeoffs, children who grow up in religious families do not seem to suffer the negative consequences of a larger family size. Rather, the opposite pattern emerges: with some exceptions (e.g. Bartkowski, Xu, & Bartkowski, 2019; Darnell & Sherkat, 1997; Sherkat, 2010; Sherkat, 2011), children raised in religious families experience benefits associated with their higher religiosity. Measures of religiosity of both parents and children are positively related to child and adolescent wellbeing across a range of indicators including better academic performance (Abar, Carter, & Winsler, 2009; Muller & Ellison, 2001); better social skills, self-control, and cognitive performance (Bartkowski, Xu, & Levin, 2008); fewer risk-taking behaviors (Abar et al., 2009; Yonker, Schnabelrauch, & DeHaan, 2012); lower likelihood of use and abuse of licit and illicit substances (Cotton, Zebracki, Rosenthal, Tsevat, & Drotar, 2006; Fletcher & Kumar, 2014); lower risk of mental illnesses such as depression (Cotton et al., 2006; Miller & Gur, 2002; Wright, Frost, & Wisecarver, 1993; Yonker et al., 2012); and higher self-esteem (Yonker et al., 2012). Although these results suggest important benefits of religiosity for child outcomes, studies rarely control for the independent effect of family size which is related to both religiosity and child outcomes. Nonetheless, the many studies which indicate that religiosity has a positive effect on a range of outcomes for children and young adults is puzzling given the typically higher fertility of such families.

The finding that children from more religious families in high-income countries do not appear to experience negative effects of large family size has been referred to as the paradox of religious fertility. This paradox may be explained by the religious alloparenting hypothesis, which proposes that higher levels of co-operation within religious communities, may offset the costs associated with larger family sizes. The type of co-operation proposed to be at work is allomaternal investment, meaning investments (time, energy, or other resources) provided directly to a child who is not the allomother's, or indirectly to that child through other support to the mother (Shaver, 2017). Participation in religious ritual is associated with higher trust and co-operation between co-religionists (Power, 2017; Purzycki & Arakchaa, 2013; Sosis & Ruffle, 2003) and in some settings, engaging in costly religious practices predicts a higher likelihood that someone will be trusted with childcare (Purzycki & Arakchaa, 2013). Separately, allomaternal support has been linked to improved health of children and higher fertility for mothers (e.g. Kramer & Veile, 2018; Sear & Coall, 2011; Sear & Mace, 2008; Sear & Mace, 2009; Sear, Steele, McGregor, & Mace, 2002), though there is variation in this effect (e.g. Sadruddin et al., 2019). It is possible, then, that higher levels of allomaternal support among religious people might mitigate the detrimental impacts of large family sizes on child outcomes. Indeed, secondary analyses by Shaver and colleagues demonstrated support for this hypothesis. Shaver et al. (2019) found in a New Zealand sample that greater religiosity predicts a higher likelihood that an adult without young children will engage in allomothering. In the UK, Shaver et al. (2020) found church attendance to be positively associated with increased social network support and practical support from co-religionists, and support from co-religionists to be associated with higher maternal fertility and improved cognitive outcomes for children. Thus, in high-income settings, these trends suggest that religious women may be able to mobilize more support from their social networks, which may mitigate quantity-quality tradeoffs relative to non-religious mothers who may have fewer social resources.

In this analysis, we examine a portion of this pathway and evaluate the relationship between religiosity, orientation towards and allomaternal support from kin, and fertility. We focus specifically on kin of the focal child in this analysis, and choose to do so for three reasons. First, cross-culturally, allomaternal support tends to come from kin, and particularly from close kin such as the child's father, siblings, grandparents, and/or aunts (Ivey, 2000; Crittenden & Marlowe, 2008; Meehan, Quinlan, & Malcolm, 2013; Kramer, 2019). All else being equal, kin have a vested interest (i.e. indirect fitness gains) in ensuring that genetically related individuals survive to reproduce (Hamilton, 1964), thereby increasing the likelihood that kin will provide allomaternal support. In industrialized and post-industrialized settings, increased mobility generally results in kin networks that are more geographically dispersed (Colleran, 2020; Newson, Postmes, Lea, & Webley, 2005; Zelinksy, 1971), and mothers may therefore rely more on non-kin for support than in pre-industrial settings (Newson et al., 2005; Sear & Coall, 2011). However, analyses of allomaternal networks in the US and UK in the dataset used here demonstrate that the bulk of allomaternal support received by young children comes from women's partners (often her children's father) and maternal kin, and that non-kin contribute relatively small amounts of non-professional care, i.e. care not provided from paid or institutional sources (Spake et al., 2021).

Second, the presence of close kin has often been associated with women's fertility and child outcomes. Particularly in low-income settings, the presence of certain types of kin (though who this person is varies considerably across contexts) has been associated with higher fertility, earlier ages of first birth, and shorter interbirth intervals (Mattison, Scelza, & Bloomfeld, 2014; Allal, Sear, Prentice, & Mace, 2004; Lahdenperä, Lummaa, Helle, Tremblay, & Russell, 2004; Scelza, 2010; Snopkowski & Sear, 2013). There are several mechanisms which could explain a positive relationship between the presence of kin and fertility: kin may provide essential support which reduces the costs (real or perceived) of beginning or continuing reproduction (McAllister, Pepper, Virgo, & Coall, 2016; Sear & Coall, 2011; Snopkowski & Sear, 2015; Turke, 1989); people from large families, and thus those with many kin around, may desire larger families due to inherited (socially or genetically) family orientation (Beaujouan & Solaz, 2019); or family members may provide pressure to have a large family through pro-natal information or advice (Newson et al., 2005; Mathews & Sear, 2013).

Third, religious individuals may be more family oriented than their non-religious peers across the life course. Parents who are more religious have higher quality relationships with their children than less religious parents (Pearce & Axinn, 1998; Smith & Kim, 2003; Regnerus & Burdette, 2006). Religious fathers and grandparents tend to spend more time with their children, and are more likely to provide assistance to children and grandchildren than are less religious fathers (Bartkowski & Xu, 2000; King, 2003; King & Elder, 1999; Lynn, Grych, & Fosco, 2016). Around the world, countries that are more religious tend to have higher rates of contact between kin (Murphy, 2008), particularly face-to-face contact but also other forms of contact such as phone calls, e-mails, or postal mail (Kalmijn & De Vries, 2009). Therefore, evidence suggests a pathway between kin orientation, support to mothers, and fertility that religious individuals may be especially exposed to due to their higher level of kin orientation.

Given the possible pathways between religiosity and kin orientation and support, and kin contact and support and fertility, we hypothesize that religious women may be more kin oriented, which may lead to greater receipt of support, and through increased support, to higher fertility. To evaluate this hypothesis, we conducted an online survey of women in the UK and the US, in which we asked about religiosity, fertility, and allomaternal support. The project procedures, predictions, and statistical approach were pre-registered prior to data collection on the Open Science Framework (OSF), available at https://osf.io/jhvn9, save the last prediction of this paper (see Section 2.2 for methodological deviations from the pre-registration). Here, we examine a subset of these predictions. Specifically we predicted that:

Religiosity will be positively associated with (a) kin-density of social networks and (b) residential proximity to kin

Religiosity will be positively associated with allomaternal investments in the form of (a) household help to women and (b) care to children

Religiosity will be positively correlated with fertility

Women who receive more allocare from kin will have higher fertility

The amount of allomaternal investment received from kin will mediate the relationship between religiosity and fertility