Study design: blinding

Statistics

Also called masking, it’s a tool to reduce bias by preventing researchers and/or subjects from knowing what is going on. You can be blinded to the hypothesis, the case-control classification, the assignments or treatments. Without blinding, humans tend to inadvertently corrupt studies with asymmetric measurements and assessments. Single-blinded refers to the subject being in the dark. Double-blinded means the researcher, too, is unaware of the allocation. Triple-blinding is sometimes mentioned and usually means even more people are unaware such as the data crunchers or safety monitoring committee. Unblinding is when everyone/thing comes out of the closet, so to speak. Usually this is intentional at the end of a trial. This is in contrast to an open trial where both the researcher and subject are aware. Blinding is preferred, but not always possible (diet interventions, surgical trials, parachute allocation), and required for USPSTF level 1 evidence classification.

Sources

    Med J Aust. 2005 Aug 1;183(3):165; author reply 166

    [PubMed: 15651970]

    BMJ;2003 Dec 20;327(7429):1459-61

    [PubMed: 14684649]

71%

Answered correctly

2014

Year asked

69%

Answered correctly

2011

Year asked

78%

Answered correctly

2008

Year asked