The question
In the early 1970s psychiatry faced criticism from many sides, and one sharp question concerned diagnosis itself. David Rosenhan, a psychologist who taught psychology and law at Stanford, asked it in the opening line of his paper: if sanity and insanity are real categories, how would anyone tell them apart? His test was simple. Send people with no symptoms into psychiatric hospitals and see whether the staff noticed.
How it was done
According to the paper, published in Science in January 1973, eight people took part: a psychology graduate student, three psychologists, a paediatrician, a psychiatrist, a painter and a homemaker. Three were women and five were men. Rosenhan himself was the first.
Between them they sought admission to 12 hospitals in five states, on the East and West coasts. They ranged from old to new and from research centres to ordinary hospitals; all but one were public or university-run. Each volunteer told the admitting staff one thing that was false: they had been hearing a voice saying words like “empty”, “hollow” and “thud”. They also gave a false name and, where needed, a false job. Everything else about their lives they described truthfully.
Once admitted, the paper says, they stopped reporting any symptom and behaved as they normally would. They had been told they would have to get out on their own, by convincing staff they were well. They took notes, at first secretly and then openly, and staff rarely remarked on it.
What came out
The results Rosenhan reported were striking:
- No pseudopatient was identified by staff. Eleven admissions were diagnosed as schizophrenia; the one at the private hospital as manic-depressive psychosis.
- Stays lasted 7 to 52 days, 19 days on average.
- Discharge came with the diagnosis “in remission”: according to the paper, each pseudopatient left with schizophrenia “in remission”, meaning present but currently without symptoms. Rosenhan read this as proof that the label stayed even when the behaviour did not support it.
- Fellow patients were more suspicious than the staff. On the first three admissions, 35 of 118 patients on the ward voiced doubts that the newcomer was really ill.
- Ordinary behaviour was read through the diagnosis. One nurse’s notes on a pseudopatient who was writing in his notebook said: “Patient engages in writing behavior.”
- Contact was scarce. Time spent each day with psychiatrists, psychologists and other doctors averaged under seven minutes. The pseudopatients were handed nearly 2,100 pills, of which they swallowed two.
Rosenhan added a second study. At one research and teaching hospital whose staff had heard the findings and doubted such errors could happen there, staff were told that one or more pseudopatients would try to get in during the next three months, and they rated each new patient. Of 193 admissions, 41 were judged, with high confidence, to be pseudopatients by at least one staff member, and 23 by at least one psychiatrist. Rosenhan wrote that no pseudopatient, at least from his group, had come.
His conclusion was that psychiatric hospitals could not reliably tell people who were unwell from people who were not, and that the setting and the label, more than the person, drove what staff saw. He was careful to say that the staff were not cruel; most of them, he wrote, clearly cared.
What really happened (replications and critiques)
The first challenge: Robert Spitzer. In 1975 the Columbia psychiatrist Robert Spitzer published a detailed critique, which he expanded in 1976. His main points:
- Faking a symptom says little about whether diagnosis works. In an analogy he borrowed from the psychiatrist Kety, someone who secretly drinks blood and then vomits it in an emergency room would be diagnosed with a bleeding ulcer, and no one would conclude that medicine cannot diagnose ulcers.
- Given a patient who reports hearing voices and nothing else explains it, considering schizophrenia was reasonable, though “diagnosis deferred” might have been better.
- “In remission” was not routine. At his own hospital, none of more than 300 discharges with schizophrenia in the preceding year carried that code. Of 12 other hospitals his assistant phoned, 11 said they never or almost never used it. So, Spitzer argued, the discharge diagnoses showed that staff had noticed the pseudopatients were well.
- The second study, in which the hospital was primed to expect impostors, made false alarms almost inevitable.
Spitzer also noticed that Rosenhan’s own reply in 1975 said eight discharges were “in remission”, not all twelve. He called the paper “pseudoscience presented as science”. Spitzer went on to chair the task force behind the 1980 DSM-III, the manual that introduced explicit criteria for each diagnosis.
The archive: Susannah Cahalan. The paper’s data could not be checked: hospitals and volunteers were anonymous, and according to Scull, Rosenhan did not give Spitzer the records he asked for. After his death, the journalist Susannah Cahalan began working through his papers. Her book The Great Pretender (2019), as described by reviewers and by Scull, reported:
- Rosenhan’s own admission was different. His records from Haverford State Hospital, where he was admitted as “David Lurie”, show he told the psychiatrist about suicidal thoughts, sensitivity to radio signals and being able to hear others’ thoughts, which goes well beyond the single voice described in the paper.
- Most pseudopatients were never found. Besides Rosenhan, Cahalan identified only Bill Underwood and Harry Lando, the volunteer whom Rosenhan left out. She found no trace of the others despite a long search.
- Underwood’s account did not match. He recalled little preparation beyond a brief tip on hiding pills, and Cahalan concluded that the writs of habeas corpus the paper says were prepared for each volunteer did not exist. Reviewers note that figures in the archive for his stay, and for the size of his hospital, differ from Rosenhan’s.
- The left-out volunteer had a good experience. Lando had published his own account in 1976. He spent 19 days in an unlocked 18-bed ward of a large public hospital, with almost one staff member per patient and more than an hour a day with doctoral staff. He found the place largely therapeutic and saw none of the powerlessness Rosenhan described, although he too felt that ordinary behaviour was read as a symptom. The paper says the ninth volunteer was dropped for falsifying parts of his history; Cahalan, citing Rosenhan’s notes, suggests the real reason was that Lando liked the hospital.
Reviewers differ on how far Cahalan goes. In Nature, Alison Abbott wrote that Cahalan could not be completely certain Rosenhan cheated. Yet Kaplan, in Psychiatric Times, quotes Cahalan concluding that the facts were distorted intentionally, while Scull thought her closing chapter held back.
The charge of fraud: Andrew Scull. In a 2023 article in History of Psychiatry, the historian Andrew Scull, using Cahalan’s material with her permission, called the study “a successful scientific fraud”. He points out, for example, that a draft listed nine pseudopatients and that the published figures stayed the same after one was removed, which he calls a statistical impossibility. He also argues the paper helped push American psychiatry toward DSM-III. The historian Roderick Buchanan, reviewing Cahalan’s book, likewise judged the study’s evidence no longer credible, while noting that Rosenhan still used parts of Lando’s experience in the published paper.
What is established, and what is not. Established, from documents: Rosenhan’s own symptoms were more severe than the paper says, a volunteer whose experience cut against the conclusion was excluded, and claims about preparation and data do not match what the archive and the surviving volunteers show. Not established: whether the unidentified pseudopatients existed, and what Rosenhan intended. Rosenhan died in 2012, so he cannot respond. No one has properly repeated the study; a 2004 book by Lauren Slater seemed to describe a repeat, but in her reply to critics she indicated that she had not run one.
Ethics
The design depended on deceiving clinical staff, who did not know they were being observed and, apart from the administrator and chief psychologist at Rosenhan’s own hospital, who knew in advance, were not told; the paper does not say whether anyone was informed afterwards. Rosenhan defended the concealment as a regrettable but needed first step, and kept the names of hospitals and staff out of the paper. The volunteers took real risks too: admission is not always easy to end. Lando, looking back, said he was not ready to recommend faking psychosis to get into a hospital, because it raises serious ethical issues.
What it means today
The study became one of the most cited papers in psychology. By the end of 2012 it had 867 citations in Web of Science, while no critique of it had more than 42. That gap is part of the story: a vivid result spread much faster than the slow work of checking it.
What remains is not nothing. Concerns about how diagnoses can colour the way people are seen, and about poor conditions in some institutions, existed before Rosenhan and do not depend on his paper. But his study cannot be used as evidence for them. And diagnoses describe real conditions that many people live with; questioning how reliable they are is different from dismissing them.
A few habits follow, though they go beyond anything the study tested:
- A famous result is not the same as a checked one. Ask whether anyone has looked at the data.
- When a study cannot be checked at all, for example because every participant is anonymous, treat it as a claim, not a finding.
- Notice when a label is doing the explaining. “Writing behavior” is just writing.
Would you have passed?
Coming soon. This is not a test of anything clinical. You’ll read short notes about the same ordinary afternoon on a hospital ward: someone waiting by a door, writing in a notebook, talking with another patient. Sometimes the note starts with a diagnosis, sometimes it doesn’t. You describe what you think is going on. At the end you’ll see whether the heading changed your words, and why this one exercise, like Rosenhan’s study, is too small to prove anything on its own.
Questions people ask
What was the Rosenhan experiment?
In a 1973 paper in Science, David Rosenhan reported that eight people with no psychiatric symptoms got themselves admitted to 12 US psychiatric hospitals by saying they heard a voice. Once inside they acted normally, yet none was identified by staff. Eleven were admitted with a diagnosis of schizophrenia and one with manic-depressive psychosis; the paper says each was discharged with schizophrenia 'in remission', though Rosenhan later put that figure at eight.
Was the Rosenhan experiment fraudulent?
It is contested. Journalist Susannah Cahalan found that Rosenhan's own hospital records describe more severe symptoms than the paper reported, and that a volunteer with a positive experience was left out. She could identify only two pseudopatients besides Rosenhan. Historian Andrew Scull calls the study a fraud. Rosenhan died in 2012 and cannot answer.
Has the Rosenhan experiment been replicated?
No. A 2004 book by Lauren Slater appeared to describe a repeat, but in her reply to critics she indicated that she had not actually conducted a study. The design has never been properly repeated.
Was the Rosenhan experiment ethical?
It relied on deceiving hospital staff, who did not know they were being studied. Rosenhan acknowledged the concealment was unpleasant but argued it was unavoidable. One of the volunteers, Harry Lando, later wrote that he could not recommend faking psychosis to enter a hospital, citing serious ethical issues.
Sources
- Rosenhan, D. L. (1973). On being sane in insane places. Science, 179(4070), 250–258. doi:10.1126/science.179.4070.250
- Spitzer, R. L. (1975). On pseudoscience in science, logic in remission, and psychiatric diagnosis: A critique of Rosenhan's 'On being sane in insane places'. Journal of Abnormal Psychology, 84, 442–452.
- Spitzer, R. L. (1976). More on pseudoscience in science and the case for psychiatric diagnosis: A critique of D. L. Rosenhan's 'On being sane in insane places' and 'The contextual nature of psychiatric diagnosis'. Archives of General Psychiatry, 33, 459–470.
- Lando, H. A. (1976). On being sane in insane places: A supplemental report. Professional Psychology: Research and Practice, 7(1), 47–52. doi:10.1037/0735-7028.7.1.47
- Cahalan, S. (2019). The great pretender: The undercover mission that changed our understanding of madness. Grand Central Publishing.
- Scull, A. (2023). Rosenhan revisited: Successful scientific fraud. History of Psychiatry, 34(2). doi:10.1177/0957154X221150878
- Scull, A. (2020, 25 January). Review of The great pretender by Susannah Cahalan. The Spectator.
- Abbott, A. (2019). Double deception in the asylum? Nature, 574, 622–623.
- Kaplan, R. M. (2020, 9 March). The great pretender: The undercover mission that changed our understanding of madness [Book review]. Psychiatric Times. www.psychiatrictimes.com/view/great-pretender-undercover-mission-changed-our-understanding-madness
- Buchanan, R. D. Book review: The great pretender. Journal of the History of the Behavioral Sciences. doi:10.1002/jhbs.22021
- Ruscio, J. (2015). Rosenhan pseudopatient study. In R. L. Cautin & S. O. Lilienfeld (Eds.), The encyclopedia of clinical psychology. Wiley. doi:10.1002/9781118625392.wbecp425
- Psychiatric News (2016). Robert Spitzer, M.D., dies at age 83, hailed as pioneering diagnostician. doi:10.1176/appi.pn.2016.2a5
- Stanford Law School (2012, 16 February). Stanford Law School mourns the loss of David L. Rosenhan, professor of law & psychology, emeritus. law.stanford.edu/press/stanford-law-school-mourns-the-loss-of-david-l-rosenhan-professor-of-law-psychology-emeritus