To quote a poster from earlier on this thread:
The original Rosenhan Experiment was an amazing, and brave, study designed to test the accuracy of psychiatric diagnoses. The study had two parts. First, Professor David Rosenhan and seven associates had themselves admitted to a variety of mental hospitals. They described auditory hallucinations and were diagnosed as paranoid schizophrenics. After they were admitted, they acted normally and waited for the hospital staff to notice. While many of the real patients noticed that the imposters lacked symptoms of mental illness, the hospital staff did not. In some cases, Rosenhan’s associates were not released for months.
In a second part of Rosenhan’s study, a hospital was told that one or more imposters would attempt to be admitted as psychiatric patients. After hearing about Rosenhan’s initial test, this institution had claimed that similar errors would not occur there. Staff at the hospital subsequently determined that 41 out of 193 patients were highly likely to be imposters. But, in reality, Rosenhan had sent no imposters to the hospital. Rosenhan ultimately concluded that “one thing is certain: any diagnostic process that lends itself too readily to massive errors of this sort cannot be a very reliable one.”