Consider an intervention or event that has occurred in a single site, such as one state in the US. The standard method for evaluation of the effects would be to compare aggregate outcomes across the intervention site and control sites. The synthetic control method is a refinement of this standard approach. Here pre-intervention data from controls is used to determine a weighted average of the controls that reproduces the pre-intervention trend in the intervention site. In this way the outcome observed in the intervention sites is compared with the intervention that would have been expected in the intervention sites if it behaved in the same way as the synthetic control. The synthetic control can be considered the counter-factual outcome at the intervention site. The choice of controls is not left to the investigator and the difference in expected outcomes between the intervention and the control units is made explicit. The method reminded me of propensity scoring in clinical epidemiology, in the sense that both methods seek to provide an unbiased estimate of the treatment effect.
— Richard Lilford, CLAHRC WM Director
- Abadie A, Diamond A, Hainmueller J. Synthetic control methods for comparative case studies: Estimating the effect of California’s Tobacco Control Program. J Am Stat Assoc. 2010; 105(490): 493-505.