Choosing exogeneity assumptions in potential outcome models

Summary There are many kinds of exogeneity assumptions. How should researchers choose among them? When is imposed on an unobservable like a potential outcome, we argue that the form be chosen based kind selection unobservables it allows. Consequently, can assess plausibility any assumption by studying distributions treatment given consistent with assumption. We use this approach to study two common assumptions: quantile and mean independence. show both assumptions require nonmonotonic relationship between outcomes. discuss how selection. also define new weaker version independence allows for monotonic unobservables. then implications choice identification. apply these results in empirical illustration effect child soldiering wages.