Abstract
This study used cluster analysis to identify sets of individuals similar to one another across multiple measures of religiousness and then ascertained if these groups differed by sociodemographic characteristics, health risk behaviors, physical health and functional status, and mental health. The authors identified six clusters using data from 1,000 community-dwelling older adults in a k-means cluster analysis of a modified version of the Duke University Religion Index. Participants in the strongly religious, moderately religious, and minimally religious clusters had the highest scores on the health, functional status, and mental health variables. The privately practicing moderate attender and the privately practicing nonattender groups were similar to each other and generally had poorer health, functional status, and mental health. The typology illustrates the importance of simultaneously considering multiple measures of religiousness and the complexity of the relationships among religiousness, sociodemographic characteristics, and health and well-being.