Gerald Monroe comments on Many methods of causal inference try to identify a “safe” subset of variation