Rosenhan wanted to investigate how doctors in hospitals are able to diagnose people as "sane" or "insane" - terms that today are not used in modern psychology.  In order to do this, he would use eight pseudo-patients - that is, confederates - who would act the part of someone seeking help in a hospital. There were 5 men and 3 women of various ages.  In real life, they had a range of occupations - including a graduate student, psychologist, paediatrician, psychiatrist and housewife. Rosenhan was one of the pseudo-patients himself.

The participants in this study were not the patients.  They were part of the study.  The participants were the staff of 12 hospitals in five different states across the USA. The hospitals did not give consent or know about the first step of the study.

The pseudo-patients told the doctor at the hospital that they were hearing voices. They told the doctor that the voices were often unclear, but used words like "empty" and "thud."  In 1973, based on the DSM II, this would have been considered a clear sign of schizophrenia.  Besides this symptom, they described their lives as they really were, without giving away their actual occupation if they were working in mental health care.

After being admitted to hospital, they acted completely normal. They spent their time talking with other patients and the staff, taking notes on their experience. The only way that they would get out would be if they could convince the staff that they were actually "sane."

All of the pseudo-patients was admitted. 11 were released with a diagnosis of "schizophrenia in remission," and one with a diagnosis of "manic depressive psychosis." The length of stay ranged from 7 to 52 days, with an average of 19 days. The nurse's notes showed that they observed no signs of abnormality during their stay. 

The conclusions drawn by Rosenhan were that once the initial diagnosis was made, then the label of schizophrenia remained, even though the pseudo-patients showed no signs of abnormality.  Rosenhan's study challenged the validity of diagnosis and how a diagnosis may influence the perception of staff working with a patient.

After the hospital staff received the findings of the first study, a second study was proposed. In this study, the hospitals were told that one or more pseudo-patients would try to be admitted to their hospital over a three month period. Each member of staff was asked to rate all patients who came in on a 1 - 10 scale for "likelihood of being a pseudo-patient."

Of 193 patients who were evaluated on the 10 point scale, none were actually pseudo-patients.  However, the staff judged 41 to be pseudo-patients by at least one staff member and 23 were suspected by one psychiatrist and 19 were suspected by a psychiatrist and a staff member.

This shows that validity can be affected when a doctor's credibility is at risk. They were less likely to make a diagnosis, fearing that they would make a mistake.

Rosenhan wrote: It is clear that we cannot distinguish the sane from the insane in psychiatric hospitals. The hospital itself imposes a special environment in which the meaning of behaviour can easily be misunderstood. The consequences to patients hospitalized in such an environment – the powerlessness, depersonalization, segregation, mortification, and self-labelling – seem undoubtedly counter-therapeutic.

However, can he really draw such a sweeping conclusion to this study?  There are several concerns about the study that may limit his ability to draw such a conclusion.

First, there is a limited sample size.  The study took place at only 12 hospitals. There were also only eight pseudo patients, many who were trained as psychiatrists/psychologists themselves.  Also, there was only one disorder that was tested - one that at the time was characterized by hearing voices. It is not possible to generalize this to all disorders, patients or hospitals.

There are also serious ethical concerns about this study. Rosenhan carried out a covert, participant observation.  He did not have the consent of the hospitals. In addition, staff were spending time working with pseudo-patients that could have been used with other patients. Rosenhan, in setting up the study, also created a situation that could have been stressful for the pseudo-patients involved.  Remember, one was not released for 52 days.

The behaviours of the staff were recorded by the “patients” themselves.  Considering that some of the patients were actually psychiatrists, there is a strong possibility of researcher bias in describing their experience. As they were the only pseudo-patient in the hospital, it is impossible for anyone to corroborate or refute this account of how the pseudo-patients were treated by the staff. Rosenhan wrote that the nurses noted that the pseudo-patients took extensive notes. He says that this was “seen as an aspect of their pathological behaviour.” However, the nurses’ notes simply said: engages in writing behaviour. The fact that Rosenhan reads so much into the nurse's notes is an example of researcher bias.

Rosenhan observed what is called the sick role bias. That is, when someone comes into a doctor and seeks help, the doctor assumes that there must be a problem. This then activates schema of past experience in diagnosis, and the doctor then looks for symptoms and behaviours that will allow him/her to make a diagnosis. This is an important ethical consideration in diagnosis, but it does not refute the need for diagnosis and care for those with mental illness. Rosenhan's study did lead to a lot of reflection on the behalf of mental hospitals, but the study itself should be seen as a product of its time with regard to the lack of ethical consideration and the rather dramatic presentation of the findings.