On testing conditional qualitative treatment effects (Q2313289)

This paper deals with hypothesis testing regarding the optimal treatment regime. Therefore, the authors introduce the notion of a conditional qualitative treatment effect (CQTE) and derive several helpful equivalent formulations for the null hypothesis that there is no CQTE. Based on kernel density estimations, they derive novel testing procedures for this problem and prove that the theoretical level is asymptotically kept. In case of specific alternatives, also asymptotic expansions of the power are given. The proposed method is illustrated in a simulation study and on real world data.