>Even the doctor wouldn't know if this were a double-blind test. Actually there's plenty of evidence that doctors do tend to know which groups patients are in.

>Even the doctor wouldn't know if this were a double-blind test. Actually there's plenty of evidence that doctors do tend to know which groups patients are in. In SSRI trials, they can be about 80% accurate (can find the ref at work tomorrow if needed). Patients also tend to know — probably because of the different side effect profiles in drug vs placebo arms. This is one of the major arguments for running trials with active placebo arms — that is, a placebo which does have side effects but which is assumed or known non-efficacious for the problem under investigation.