Economic epistemology is hard. Your data can never be controlled for all confounding variables. So you try to do the best you can. With the difficulty of isola

Economic epistemology is hard. Your data can never be controlled for all confounding variables. So you try to do the best you can. With the difficulty of isolating variables in social data, I doubt the objectivity of large-scale empirical economics ever improves. Other social sciences are probably worse. That's why they are stuck with the modifier "social". So your criticism is right. The best large-scale empirical economics can offer is plausible-sounding models and examples that back them up.