I feel like this is the case for most economics studies I've read (note: not nearly an expert). I enjoy the types of studies done and think they have value, but am concerned they can be taken as direct, strong supporting evidence when by their nature they make controlling for confounding variables impossible.