The proposed research aims to investigate patterns, measurement models, and predictors of religious involvement among older black and white Americans. These aims will be accomplished through a research plan organized around four interrelated objectives. Specific objectives include (1) describing patterns of religious involvement in older blacks and whites and identifying predictors of religiosity, with a substantive focus on racial, age, gender, and other social-structural differences; (2) confirming multidimensional measurement models of religious involvement in different groups of black and white respondents and examining differences in model structure across racial, regional, and denominational classifications; (3) analyzing patterns, models, and predictors of religious involvement across three-generational families; and, (4) performing panel analyses of religious involvement. In accomplishing each of these objectives, special attention will be given to multidimensional measurement of religious involvement, higher-order analyses which control for the effects of relevant exogenous variables, and the confirmation of findings across racial, age, gender, and other social and religious categories. In conducting the proposed research, five datasets will be used. These include the National Survey of Black Americans, the Three Generation Family Study, the National Survey of Black Americans Panel Study, the Americans' Changing Lives Study, and the National Opinion Research Center's General Social Survey. Each of these surveys is based on large, national probability samples, each contains sizeable numbers of black respondents, and each dataset includes a variety of indicators of organizational, nonorganizational, and subjective religious involvement. In analyzing these data, a number of multivariate statistical procedures will be used, including multiple regression, logistic regression, analysis of variance and covariance, and several procedures based on covariance-structure-modeling. These include confirmatory factor analysis, structural-equation-modeling, and factorial-invariance analysis.