both attributes and was examined separately. “No Religion” endorsements formed the reference group (n = 625 cases). [Note that 460 cases (11.5%) were unclassified due to minimal representation.] As well, across these types, religious proscription of alcohol was assessed by endorsement that a family’s religious affiliation had rules against all alcohol use (a “Religious Rules” indicator). Substantial within-group heterogeneity was evident within these categories, but between-groups main effects have been sufficiently robust as to validate these constructs. Finally, religious affiliation during childhood was obtained from mother’s report of family affiliation during child’s ages 6–13, and was considered a family-level characteristic.
