... cookie cutter correlation to causation type argument, so to strengthen it, ... Specifically, we need to deny an endogenous variable ... to both depression levels and reported recreation time), deny that this correlation ... to do with "reporting;" we need to assume to ...