Religious Affiliation, Participation and Fertility: A Cautionary Note

Abstract

When studies of religion's effect on fertility seek to discriminate between nominal and committed followers, they often turn to measures of participation in church activities. Such an approach is conceptually defensible but empirically problematic with the typical cross-sectional data set, because the temporal ordering is wrong: such studies usually ascertain religious participation at the time of the interview, after the birth of children. I explore the problem empirically with data from the 1963 Glock-Stark Northern California Church Member Study. Finally, I discuss alternative indicators of religiosity for cross-sectional analysis, and the need for appropriate panel data.