The measure and mismeasure of fairness: a critical review of fair machine learning