Those studies also don't demonstrate any causality, only correlation. It seems not unlikely that districts willing to try an experimental voting method also are more interested in running diverse candidates.