Compared with studies focusing on the general population, the majority of studies focusing on the low-income population are in the early stages of methodological development (with some exceptions, as described within this report). As indicated in Table 8-1, all of the studies included in this review conducted multivariate regression analysis across each outcome area.
In addition, Table 8-1 highlights that research across all of the topical areas has included some initial test of mediators to distinguish between pathways of the direct effect of religiosity and possible indirect pathways through which religiosity can affect outcomes. Few studies distinguish between organizational and nonorganizational religiosity when testing for mediating effects. One thing that is problematic in the religiosity literature is that basic tests of mediators are not consistently implemented. Further, there are fewer than five studies that attempt to comprehensively model the direction of effects by estimating simultaneous equations. Studies using this approach typically are in the health and substance abuse fields.
As with any developing research literature, the bulk of the research begins by establishing an association; researchers then use more rigorous methods to establish causal paths. Recently, a small but increasing number of studies started to use more rigorous estimation methods to advance religiosity research focused on the low-income population beyond the associational phase. Table 8-1 highlights some of the methods being used.
In the area of marriage, one study uses within-couple fixed effects analyses drawing on religiosity measures that range from general indicators of church attendance to more directly relevant relationship-specific religiosity measures (e.g., homogamy of couple religious denomination) to study marital quality. In the area of youth development, there is a set of new studies that estimates the effects of religiosity on youth development using multiple longitudinal data sets that draw comparisons between higher- and lower-income groups. These studies compare and contrast the findings using propensity score matching, instrumental variables, and fixed effects models. In the areas of health services utilization and substance abuse treatment, a number of randomized trials and quasi-experimental evaluations are testing the effect of religiosity and program interventions delivered at faith-based institutions on outcomes over time. Going forward, it will be important to build on the approach of using multiple statistical techniques that help to establish the causal paths between religiosity and outcomes, especially when randomization at the client level to test program effects is not feasible.