A method of avoiding bias in the context of treating a disease. In a medical experiment the comparison of treatments could be biased if either the patient, the doctor administering the treatment, or the data analyst knew which treatment was allocated to which patient. If the patient (or the doctor, or the data analyst) is unaware of which treatment is being given, then they are said to be ‘blind to’ the treatment allocation process. If neither the patient nor the doctor is aware of the treatment allocation then the process is said to be a double-blind trial.

Typically the patient’s permission will be required for participation in such a trial. Agreement by a patient to being treated in an unknown fashion may be difficult to obtain. In Zelen’s design, the allocation of treatments to patients is at random, but the allocation is revealed to the patient so that the patient’s decision concerning participation is fully informed.

In such an experiment, one treatment is often a dummy—for example, a pill of the same shape and flavour as the genuine pill, but with no active ingredients. This is called a placebo. In practice there is often a placebo effect, in which the treated person shows a benefit despite the treatment theoretically having no effect. The effect is a reflection of the psychological benefit of believing that one is being given an effective treatment.

The related Hawthorne effect recognizes that individuals behave differently when aware that they are being studied, whilst the Pygmalion effect recognizes that individuals are influenced by the expectations that others have of them.