Fig. 3From: Testing for carryover effects after cessation of treatments: a design approachThis graph shows estimated rates of false positives in a parallel-group design for the rules mentioned, most are far from the normal 5 %Back to article page