Table 1 Analytical techniques included in Medical Research Council guidance on natural experimental studies 1
Analytical technique | Brief description |
|---|---|
Controlling for observable characteristics | |
Matching | Involves finding unexposed individuals (or clusters of individuals) which are similar to those receiving the intervention, and comparing outcomes in the two groups |
Regression adjustment2 | Measured characteristics that differ between those receiving the intervention and others can be taken into account in multiple regression analyses |
Propensity scores | An estimate of the likelihood of being exposed given a set of covariates, propensity scores are usually estimated by logistic regression, and can be used to match exposed with unexposed units (which may be individuals or clusters of some kind) using values of the propensity score rather than the covariates themselves |
Controlling for unobservable characteristics | |
Difference in differences | Involves comparison of change over time in exposed and unexposed groups, which enables control of unobserved individual differences and common trends |
Instrumental variables | An instrumental variable is a factor associated with exposure to an intervention, but independent of other factors associated with exposure, and associated with outcomes only via its association with exposure |
Regression discontinuity | This approach exploits a step change or `cutoff’ in a continuous variable used to assign treatment, or otherwise determine exposure to an intervention. The assumption is that units (individuals, areas, etc.) just below and just above this threshold will otherwise be similar in terms of characteristics that may influence outcomes |