Natural Experiments from Observational Data
In social and public health sciences, there is a need for more confidence in determining whether associations based on observational data are causal.
Confidence in the causality of issues leading to better health not only improves our understanding of how we might intervene to achieve this, but also translates into confidence that future interventions and policies will succeed. Discussion of this issue frequently focuses on technical aspects of statistical analysis, but it is equally important to focus on the research questions asked and the study design used.
There is a need to understand the impact of policy and other issues on population health and inequalities using natural experiments. Work in this theme builds on previous work and expertise in our Unit.