The stats-first approach to issues of race and incarceration isn’t effective—in fact, it’s potentially counterproductive.