A participant who knows they received the active treatment reports differently from one who knows they did not, particularly on outcomes that depend on how someone feels.
A participant who knows they received the active treatment reports differently from one who knows they did not, particularly on outcomes that depend on how someone feels. Blinding removes that channel. It is not about dishonesty; it is about a well-documented effect that operates whether or not anyone intends it.
The person recording an outcome can shade a borderline judgement toward the result they expect. Where an outcome requires interpretation — reading an image, grading a symptom — blinding the assessor matters more than blinding the participant, and trials should state which were blinded rather than claiming the word in general.
It usually means participants and investigators, but reporting is inconsistent enough that the phrase alone tells a reader little. The useful detail is a list: who was blinded, and who was not. Papers that provide the list are easier to trust than papers that provide the label.
Surgical and behavioural interventions often cannot be concealed from anyone involved. That is a real constraint rather than a failing, and the right response is to choose outcomes less sensitive to expectation, which good trials in those fields do explicitly.
A treatment with a distinctive side effect tells participants which arm they are in. Careful trials ask participants at the end which group they believed they were in, and report whether the guesses were better than chance. Very few do, and the ones that do are signalling something about their standards.
An unblinded trial reporting a subjective outcome is the combination most likely to overstate an effect. That does not make the result worthless; it makes the size of the effect less reliable than the direction, and a careful reader weights it accordingly.
Blinding conceals the assignment. A placebo gives the control group something to receive, so the comparison is between treatment and the experience of being treated rather than between treatment and nothing. The two are separate design choices and a trial can have one without the other.
Whether the methods name who was blinded, whether the outcome depends on judgement, and whether anything about the treatment would make the assignment obvious to a participant.
The claims on this page about who knows which group a participant is in, and what changes when they do can be checked against sources outside this site: