Pet peeve: cancer trial reports that say they have a descriptive primary endpoint (e.g. "the primary objective was to determine the objective response rate") but then make inferential conclusions (e.g. "this drug is active in this population") without specifying their decisions threshold