Answers are grounded in this paper and cite their sources.
• Primary clinical research can be categorised as either observational or experimental.• Bias is present in all observational studies.•The relationship between an exposure and outcome can be distorted by a third extraneous variable, a confounder.•Statistical strategies to address confounding include matching, stratification, regression, propensity scoring and instrumental variable analysis.•The main weakness of propensity scoring is that it can only control for known and measured confounding.
No takes yet. Share an insight, caveat, or question.
Hewson et al. (2025) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: