Plenty is doing.
"When the technology of moderator identification (of identifying variables that altered correlations amongst variables-thus allowing possible findings that patterns of covariation were different in, e.g., minority groups) became well-known, almost every study either identified or alluded to the presence of potential moderators of predictive relationships.... Recently [however] it has become clear that moderators are much less frequent than was originally assumed."
Disclaimer: this quote appears here only to spark discussion. It is not endorsed one way or the other. Make up your own mind. Or just refresh the page for another viewpoint. From a collection assembled by the late Chris Brand.