Summary
This methodological paper addresses a critical but underappreciated source of bias in epidemiological research: the role of selection bias in inducing collider bias. The authors argue that whilst selection bias is known to affect representativeness and prevalence estimates, its substantial impact on association estimates is often overlooked. Through simulation evidence, they demonstrate that selection mechanisms linked to phenotypes can substantially distort genetic and phenotypic associations, and propose that understanding and adjusting for factors influencing study selection and attrition—such as through polygenic score sensitivity analyses—is essential for valid causal inference.
UK applicability
These findings are directly applicable to UK birth cohorts and epidemiological studies such as the Avon Longitudinal Study of Parents and Children (ALSPAC) and UK Biobank, where selection and attrition are known challenges. UK researchers conducting association studies should consider implementing the suggested sensitivity analyses and adjustment approaches to quantify potential collider bias.
Key measures
Bias in phenotypic and genotypic association estimates under varying selection and attrition scenarios; effects of conditioning on colliders; polygenic score predictive validity for study participation
Outcomes reported
The study examined how selection bias and attrition in large-scale cohort and cross-sectional studies can induce collider bias, leading to substantially biased estimates of both phenotypic and genotypic associations. Through simulation, the authors demonstrated that even modest influences on study selection and attrition can generate misleading association estimates.
Topic tags
Dig deeper with Pulse AI.
Pulse AI has read the whole catalogue. Ask about this record, its theme, or how the findings apply to UK farming and policy — every answer cites the underlying studies.