synthetic control design
A quasi-experimental statistical method that constructs a counterfactual 'synthetic' version of a treated unit from a weighted combination of untreated units to estimate a policy's causal effect.
As defined in 152 | Charis Kubrin on Criminology, Incarceration, and Hip-Hop
Where it comes up · 1
Most important claims · 3 of 3Search all claims for “synthetic control design” →
- A common analytical error is that when a policy is enacted, observers track crime trends afterward and immediately attribute any change to that policy, ignoring that crime and other outcomes are affected by many factors (guns, drugs, gangs, economic conditions, joblessness, poverty, demographics, policing); isolating one policy's impact requires more sophisticated methods like synthetic control design.152 | Charis Kubrin on Criminology, Incarceration, and Hip-Hop
- The synthetic control design method is a quasi-experimental approach that addresses major challenges to determining causality when a pure experiment is impossible (you cannot randomly assign a policy to some states and not others), offering the benefits of an experimental approach minus random assignment to treatment or control.152 | Charis Kubrin on Criminology, Incarceration, and Hip-Hop
- Using a synthetic control design to model a counterfactual world where marijuana was not legalized, research found that legalizing marijuana lowered suicide rates in California, with possible mechanisms including people substituting marijuana for alcohol, though the mechanisms remain unexplored.152 | Charis Kubrin on Criminology, Incarceration, and Hip-Hop