Classic study: Rosenhan (1973) On being sane in insane places

Aims and study 1 procedure

Aims: to see whether sane people could be detected in psychiatric hospitals, and how patients were treated.

Sample: 8 sane pseudo-patients (3 women, 5 men: a psychology graduate student in his 20s, three psychologists, a paediatrician, a psychiatrist, a painter and a housewife) at 12 hospitals in 5 US states, varying in age, funding and staffing.

Procedure: a field study using participant observation. Each pseudo-patient phoned for an appointment and said they heard voices saying 'empty', 'hollow' and 'thud'. Everything else they said about their lives was true. Once admitted they stopped simulating symptoms, behaved normally, joined ward activities and recorded observations, at first secretly, then openly. They were given medication, which they did not swallow.

Study 1 results

• None of the pseudo-patients was detected by staff. All but one were admitted with a diagnosis of schizophrenia, and all were discharged with 'schizophrenia in remission'.
• Stays lasted 7 to 52 days, average 19 days.
• 35 of 118 real patients voiced suspicions that the pseudo-patients were sane (29.7%).
• About 2,100 tablets were given; only 2 were swallowed.
• Nurses spent about 90% of their time in the ward office; contact with psychiatrists and psychologists averaged under 7 minutes a day.
• Note-taking was recorded in nursing notes as 'writing behaviour', interpreting normal behaviour through the label.

Study 2

Staff at a teaching hospital were told that over 3 months one or more pseudo-patients would try to be admitted. They rated each new patient. Of 193 patients, 41 were judged with high confidence to be pseudo-patients by at least one staff member, 23 by at least one psychiatrist, and 19 by a psychiatrist and another staff member. Rosenhan sent none.

Conclusion: the diagnostic process cannot reliably tell sane from insane; once a person is labelled, the label colours how all their behaviour is seen. In study 1 staff erred towards calling the healthy sick; in study 2, when their reputation was at stake, they called the sick healthy.

Evaluation

• Ecological validity: real hospitals and real staff who did not know they were being studied.
• Ethics: staff were deceived and could not consent; pseudo-patients may have taken resources from real patients.
• Small sample of 8 pseudo-patients; Rosenhan was one of them, which may introduce bias.
• Study 2 had no 'true' answer to check judgements against.
• Fairness to psychiatrists: doctors are expected to take reported hallucinations seriously, so admission may have been reasonable; the label 'in remission' was unusual and arguably shows staff did notice normal behaviour.
• Impact: it raised concerns about labelling and diagnosis that influenced later revisions of DSM (9.3.8).

Worked example

How does Rosenhan's study 2 provide evidence about the reliability of diagnosis?

1. Staff were expecting pseudo-patients and judged 41 of 193 genuine patients to be likely fakes with high confidence.
2. No pseudo-patients were sent, so every one of these judgements was wrong.
3. Different staff also disagreed (41 by any staff member, 23 by a psychiatrist), showing low agreement.
4. So diagnosis depended on expectations and context rather than consistent criteria: poor reliability.

Answer: Staff judgements were inconsistent and driven by expectation, suggesting poor reliability.

Check yourself

1. Rosenhan's pseudo-patients stayed between 7 and 52 days. What was the range of stay?
   Show answer

   45 days. 52 - 7 = 45 days.

2. 35 of 118 real patients suspected the pseudo-patients were sane. What percentage is that? Give your answer to 1 decimal place.
   Show answer

   29.7 %. 35 / 118 x 100 = 29.7%.

3. In study 2, 41 of 193 patients were judged pseudo-patients by at least one staff member. What percentage is that? Give your answer to 1 decimal place.
   Show answer

   21.2 %. 41 / 193 x 100 = 21.2%.

4. Which words did the pseudo-patients say their voices said?
   • Empty, hollow, thud
   • Kill, die, hurt
   • Help, stop, run
   • Mother, father, home
   Show answer

   Empty, hollow, thud. Chosen to suggest existential emptiness, not typical of psychosis.

5. With what diagnosis were the pseudo-patients discharged?
   • Sane
   • Manic depression
   • Schizophrenia in remission
   • No diagnosis
   Show answer

   Schizophrenia in remission. The label stayed, marked 'in remission'.

6. Which is the main ethical criticism of Rosenhan's study 1?
   • Hospital staff were deceived and did not consent
   • The pseudo-patients were children
   • Participants were given electric shocks
   • It used animals
   Show answer

   Hospital staff were deceived and did not consent. Staff did not know they were being studied.

7. Explain what Rosenhan's 'writing behaviour' finding shows about labelling.
   Show answer

   Once the pseudo-patients were labelled with schizophrenia, staff interpreted normal note-taking as a symptom, showing the label changed how all their behaviour was seen.. Labels make normal behaviour look like illness.

Practise with instant feedback Print this lesson

Sources

• Pearson Edexcel International Advanced Subsidiary/Advanced Level in Psychology specification (issue January 2026)
• Pearson IAL Psychology component guide: WPS04 (Issue 2, November 2024)
• Pearson IAL Psychology summary of studies: WPS04

Last updated 8 November 2026. Spotted a mistake? Tell us.