ABSTRACT
When the panel data from the classes of 1976-1984 are compared closely with the cross-sectional data from high school seniors at matched times, the patterns suggest the following curvilinear relationship with age: religious attendance does drop somewhat during the first few years after high school, but after about the mid-twenties attendance increases slightly compared with what we suspect are overall societal trends. Confirmation of that conclusion is provided by another set of tabulations, reported in Bachman et al. (2001) ; when we examined data from all base-year and follow-up surveys collected in the years 1990-1998 (thus holding period effects constant while examining age differences), we found that 33% of the seniors (modal age 18) attended weekly or more often, compared with 25% of those aged 19-20, 2 1 % of those 21-24, 23% of those 25-28, and 28% of those 29-32. Again, the relationship with age is clearly curvilinear.
