Blinding is intended to prevent people who can potentially introduce bias into a treatment comparison through knowledge of the treatment allocation from doing so.
People who can potentially be blinded include the people receiving the treatments being compared, the people delivering the treatments, data collectors, people who assess the outcomes, data analysts, the data safety and monitoring committee, and manuscript writers.
Biases that blinding can help to prevent - depending upon who is blinded - include:
- placebo effects;
- differences in the care provided to the participants in a study, other than the treatments being compared (performance bias);
- differences in how outcomes are measured in treatment comparison groups (measurement bias);
- interpretation bias, reporting bias, and sponsor bias.
If you feel that this definition hasn't helped you to understand the term, click on our monkey to let us know.