The proposed research aims to analyze patterns of several dimensions of religiosity, to create multidimensional measurement models of these constructs, and to structurally model their effects on other endogenous constructs, including health status and psychological well-being. Exogenous constructs include education, income, marital status, and other correlates or predictors of religiosity or the other latent constructs. Through replicated secondary data analysis using seven national or regional datasets on aging, differences in these patterns and the factorial invariance of these measurement models and structural effects of religiosity will be examined comparatively by gender, race and ethnicity, and age. Both cross-sectional and multiwave panel models will be analyzed, and several multivariable and multivariate statistical techniques will be employed. These include analysis of covariance with multiple-comparison tests, confirmatory factor analysis, and structural-equation modeling. The latter two analyses will be conducted using LISREL-7.16 and EQS/EM-3.0, statistical software packages for covariance-structure modeling. The seven datasets to be used include the NORC General Social Survey, the NCOA Myth and Reality of Aging survey, the Quality of American Life survey, the National Survey of Black Americans, the Americans' Changing Lives study, the Three-Generations Study of Mexican Americans, and the San Antonio Longitudinal Study of Older Mexican Americans. In sum, the proposed research seeks to advance empirical health-related research on religiosity within gerontology through (a) the use of covariance-structure modeling techniques, (b) confirmation of multidimensional latent constructs for measuring religiosity, (c) multisample and split-sample replication, (d) the positing of alternative, testable, operational models based on mid-range theory, and (e) a comparative approach focused on gender, race and ethnicity, and age.
Showing the most recent 10 out of 13 publications
