The search for causal inferences: using propensity scores post hoc to reduce estimation error with nonexperimental research.

The search for causal inferences: using propensity scores post hoc to reduce estimation error with nonexperimental research.