A Major Advance in Our Knowledge about Religion and Health
While there have been many prior studies on spirituality and health and even many meta-analyses within the religion and health literature, most of these are relatively weak in design, employing cross-sectional data, collected only at one time point, which makes it difficult to determine what causes what. Does religious service attendance protect against smoking, or do those who smoke feel less comfortable attending religious services? To make these determinations, we need longitudinal data, gathered over time, though even with such data, we can only approach genuine knowledge about causation when the accumulated evidence, generally from multiple studies, is sufficiently strong and unambiguous to render any alternatives untenable. And as we’ve noted previously, achieving genuine knowledge about causal claims is also difficult because of the possibility that researchers are neglecting some additional “confounding” factor that explains the apparently causal relationship, or because of statistical uncertainty, or questions of generalizability, or biases that can creep into the research process.
One approach to trying to address these challenges is meta-analysis, or the synthesis of previous studies on a particular topic. This still doesn’t solve all the issues since…
