A confusion bias that occurs when the subjects assigned to the experimental group do not receive the intervention or receive another intervention, or when some subjects in the control group receive the intervention.