The amount of difference in the outcome variable that the investigators wish to detect between intervention and control groups in a study.