Blinding

Blinding is a technique used to minimize bias in a study by preventing the participants from knowing which treatment option they are allocated to receive. Blinding can take three main forms: single, double, and triple blinding.

Single blinding involves the subjects being unaware of their treatment regimen. In a study comparing a new drug to a drug already in use, a subject may have a preconceived notion that one drug is more effective than the other. This can cause a potentially exaggerated effect (see placebo effect) for either the standard or the investigational treatment.

Double blinding involves keeping both the investigator and the subject unaware of the treatment assignment. This eliminates potential investigator bias for measuring efficacy or detecting potential adverse events associated with a treatment.

Triple blinding involves keeping the treatment assignments unknown to the subject, to the investigator, and to the person performing the data analysis.

Various study designs can be blinded to varying degrees. To be considered scientifically sound, most study designs require double blinding. The use of quantification and universal statistical criteria (p value < .05, and so forth) often eliminates any potential bias that could be introduced at the data analysis level. If a study involves a subject or investigator being “blinded,” this information should appear in the consent form. For studies involving a placebo, this placebo blinding, as well as under what conditions the “blind” may be broken, must also be explained in simple language in the consent form. Washington, DC., or http://ohsr.od.nih.gov/mpa/