On God and grades.

Not a family matter: The effects of religiosity on academic outcomes based on evidence from siblings 

Ilana M.HorwitzaBenjamin W.DomingueaKathleen MullanHarrisb 

Abstract

Religiosity has been positively linked with multiple measures of academic success, but it is unclear whether the “effect” of religiosity on academic outcomes is causal or spurious. One source of heterogeneity that may contribute to a child's level of religiosity and his/her academic success is family background. This paper is the first to use sibling differences to estimate the associations between religiosity on short and long-term academic success. Our analysis yields two main results. First, more religious adolescents earned higher GPAs in high school, even after including family fixed effects. Second, because they earned higher GPAs in high school, more religious adolescents completed more years of education 14 years after their religiosity was measured. Our findings suggest that adolescents' religious commitments influence their schooling in both the short and long term and should be more actively included and theorized as important drivers of educational and economic stratification. 

Introduction 

Religion permeates every aspect of American society. Religious commitments shape where Americans live, how they vote, who their friends are, and even how happy they are. Religion is a particularly salient feature for millions of American teenagers: one in two see faith as central to their daily life, and one in three say they pray daily (Smith and Denton, 2005).1 When it comes to academic performance, religiously engaged adolescents appear to have better academic outcomes than those who are not religiously active. They earn higher GPAs (Glanville et al., 2008; Good and Willoughby, 2011; McKune and Hoffmann, 2009; Milot and Ludden, 2009; Regnerus and Elder, 2003; Tirre, 2017; Toldson and Anderson, 2010), aspire to go farther in school2 (Al-fadhli and Kersen, 2010; Muller and Ellison, 2001; Regnerus, 2000), and actually stay in school longer (Brown and Gary, 1991; Kim, 2015; Lee et al., 2007; Lee and Pearce, 2019; Lehrer, 2010, 2004; Loury, 2004; Mohanty, 2016). The theoretical reason for this positive association is that increased religiosity tends to deter young people from risky behaviors, promotes social capital and network closure, and motivates youth to act in ways that adhere to the moral grounding of their religious teachings (Smith, 2003).

While the evidence suggests that more religiously engaged students have better academic outcomes, questions remain as to how to interpret this evidence. The existing evidence has been derived from observational approaches that inherently limit the scope of inference; as a consequence, there is uncertainty about whether the “effect” of religiosity on academic outcomes is causal or spurious (Bagiella et al., 2005; Cochran et al., 1994; Freeman, 1986; Regnerus and Smith, 2005). We are particularly concerned about the role of family background—a key source of heterogeneity that influences children's level of religiosity as well as their academic success (Eirich, 2012; Ludwig and Mayer, 2006). While previous studies have attempted to eliminate family-level confounders by including a set of observed family-level controls (e.g., parental education, family income, and family structure), these controls do not effectively address family-level heterogeneity, especially when these factors are unobserved (Kim, 2018). Thus, previous studies may be overstating the actual effect of religiosity, which could be null or even negative. Approaches that allow for more stringent analyses of such observed associations are useful in such settings. Here, conduct within-family analysis by analyzing sibling pairs to better understand the association of religiosity, high school GPA, college aspirations, and educational attainment. Data on sibling pairs allow us to separate the contribution of religiosity from families by examining whether sibling differences in religiosity translate into sibling differences in academic achievement  (since family differences are muted between siblings).