I am working on a new causal inference method for estimating treatment effects and want to test it against current methods. I would like to find a dataset where the counterfactual is either observed in an almost identical unit (i.e. the same person at a later point in time, not a different person; I realize this is impossible for the exact same unit) or the set up allows for good estimation of the counterfactual. I've spent a while trying to brainstorm ideas, and would like some outside input. I think biological studies that involve genetically identical mice given and the Card-Krueger minimum wage data near the PA/NJ state line are potential examples.