Stepped trials: Magnifying methodological muddles — the supernatant effect
AbstractStepped trials such as STAR*D suffer from a serious methodological flaw. In each stage of the trial one is dealing with a different population sample that has a different percentage…