Exploring causality from observational data: An example assessing whether religiosity promotes cooperation

Abstract

Causal inference from observational data is notoriously difficult, and relies upon many unverifiable assumptions, including no confounding or selection bias. Here, we demonstrate how to apply a range of sensitivity analyses to examine whether a causal interpretation from observational data may be justified. These methods include: testing different confounding structures (as the assumed confounding model may be incorrect), exploring potential residual confounding, and assessing the impact of selection bias due to missing data. We aim to answer the causal question “Does religiosity promote cooperative behaviour?” as a motivating example of how these methods can be applied. We use data from the parental generation of a large-scale (n = approx. 14,000) prospective UK birth cohort (the Avon Longitudinal Study of Parents and Children), which has detailed information on religiosity and potential confounding variables, while cooperation was measured via self-reported history of blood donation. In this study, there was no association between religious belief or affiliation and blood donation. Religious attendance was positively associated with blood donation, but could plausibly be explained by unmeasured confounding. In this population, evidence that religiosity causes blood donation is suggestive, but rather weak. These analyses illustrate how sensitivity analyses can aid causal inference from observational research.

Identifying culture as cause: Challenges and opportunities

“…Unconfoundedness is well understood in cultural evolutionary studies (Major-Smith, 2023), but it is not enough to make causal claims in observational studies. Rather, researchers need to also invoke positivity.…”

“…Moreover, cultural evolutionary scholars can rely on several tools to guide their reasoning about omitted common causes. Most notably, DAGs are a common logical–statistical tool to guide covariate selection (for an example in cultural evolutionary studies, see Major-Smith, 2023). Cultural evolutionary researchers can also rely on sensitivity analyses to assess the fragility of their results to confounding (e.g.…”

“…Thus, including them might be akin to adjusting for a mediator (something we should usually avoid doing, see e.g. Major-Smith, 2023). At worst, adjusting for these covariates might even lead to collider bias owing to conditioning on common effects.…”

“…regression where suspected confounders are adjusted for, see, e.g. Major-Smith, 2023), briefly reviewing their limits. Third, we present some empirical strategies that try to approximate an ideal experiment where culture is assigned randomly to individuals or entire groups: instrumental variable estimation, regression discontinuity design and epidemiological approach.…”