The authors present results of their winning solution to the National Institute of Justice recidivism forecasting challenge. Their team, “MCHawks,” placed highly in both terms of accuracy (as measured via the Brier score), as well as the fairness criteria (weighted by differences in false positive rates between White and Black parolees).
The authors used a non-linear machine learning model, XGBoost, although they detail their search of different model specifications, as many different models’ predictive performance is very similar. Their solution to balancing false positive rates is trivial; they bias predictions to always be “low risk” so false positive rates for each racial group are zero. They discuss changes to the fairness metric to promote non-trivial solutions. By providing open-source replication materials, it is within the capabilities of others to build just as accurate models without extensive statistical expertise or computational resources. (Publisher abstract provided)