A high level overview of three causal identification techniques to undertake geospatial impact evaluations:  regression discontinuity, "matching", and "natural experiments".