Double blinding has multiple definitions and is interpreted in different ways.
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.
Unless stated, it is not clear which of these people were blinded in a “double blind” study.
Because the meaning of double blind is ambiguous, we recommend not using it. It is better to consider explicitly who was blinded, and who was not blinded, and how that might have protected against or led to a risk of bias, including placebo effects, differences in the care provided to the participants in a study other than the treatments being compared (performance bias), or differences in how outcomes are measured, in treatment comparison groups (measurement bias).
If you feel that this definition hasn't helped you to understand the term, click on our monkey to let us know.