Causal Inference
'Causal inference' is the process of determining whether a specific factor (exposure, variable, or intervention) directly causes a change in an outcome, rather than merely being associated with it.
Key distinction
- Association ≠ Causation
A variable may correlate with an outcome without causing it.
Methods for causal inference
- Randomized controlled trials (RCTs) – gold standard for establishing causality.
- Quasi-experimental designs – such as difference-in-differences, instrumental variables.
- Causal modeling – e.g., directed acyclic graphs (DAGs), structural equation models.
- Counterfactual reasoning – comparing observed outcomes with hypothetical alternatives.
In observational studies
Causal inference requires:
- Control of confounders
- Attention to bias (selection, measurement, etc.)
- Proper temporal sequence (cause precedes effect)
Clinical importance
Causal inference underpins:
- Treatment effect estimation
- Policy decisions
- Guideline development
'In summary:' causal inference seeks to answer the question: “Does X cause Y?”, not just “Is X related to Y?”