Stanford professor David Rosenhan wanted to prove that psychiatric hospitals of his day would instantly label and mistreat even the mildest of symptoms, so he convinced a group of students to go undercover at various hospitals and then wrote an influential study about the results. Rosenhan’s paper “On Being Sane in Insane Places” was published in the journal Science in 1973 and had a big impact at the time.

Rosenhan’s eight healthy pseudopatients allegedly each followed the same script to gain admittance to psychiatric hospitals around the country. They each told doctors that they heard voices that said, “Thud, empty, hollow.” Based on this one symptom alone, the study claimed, all of the pseudopatients were diagnosed with a mental illness — mostly schizophrenia.

And once they were labeled with a mental illness, it became impossible to prove otherwise. All eight were kept hospitalized for an average of 19 days — with the longest staying an unimaginable 52. They each left “against medical advice,” meaning the doctors believed that they were too sick to leave. A total of 2,100 pills — serious psychiatric drugs — were reportedly prescribed to these otherwise healthy individuals.

At the time, the collective American imagination was deeply suspicious of psychiatry and its institutions. It was the era of Ken Kesey’s “One Flew Over the Cuckoo’s Nest” and movies like “Shock Corridor” and “The Snake Pit.” Rosenhan — who was both an insider who studied abnormal psychology, and an outsider who was a psychologist rather than a psychiatrist — was the perfect person to pull back the curtain on psychiatry’s secrets.

Author Sussana Cahalan heard about the study by accident and began looking into it. She wanted to find out why Rosenhan had never finished the book he was paid to write about it or why he never again revisited the topic in any subsequent paper. What she discovered was a shock to her. Rosenhan counted his own experience among the eight people mentioned in the study, but he misrepresented that experience in order to make it appear the hospital had overreacted to his presentation:

Rosenhan’s medical record during his undercover stay at Haverford found that he had not, as he had written in his published paper, only exhibited one symptom of “thud, empty, hollow.” Instead, he had told doctors that he put a “copper pot” up to his ears to drown out the noises and that he had been suicidal. This was a far more severe — and legitimately concerning — description of his illness than he had portrayed in his paper.

Obviously there’s a big difference between hearing non-distinct voices and telling people you are suicidal because you can’t stop the voices unless you put copper pots over your ears. One is worrisome, the other is life-threatening.

But that wasn’t the most surprising part of what Cahalan found. After months investigating, she was unable to locate most of the other “pseudopatients” mentioned in Rosenhan’s paper. She did identify one of them. His name is Bill Underwood and his experience did match the claims in Rosenhan’s study. But the other six pseudopatients don’t seem to exist. She speculates that may be why Rosenhan never finished the book he was contracted to write and in fact never wrote about the topic again.

Cahalan did identify another student who participated in the study but Harry Lando was dropped because his experience didn’t match what Rosenhan was looking for:

Lando had summed up his 19-day hospitalization at the US Public Health Service Hospital in San Francisco in one word: “positive.”

Even though he too was misdiagnosed with schizophrenia, Lando felt it was a healing environment that helped people get better.

“The hospital seemed to have a calming effect. Someone might come in agitated and then fairly quickly they would tend to calm down. It was a benign environment,” Lando, now a psychology professor at the University of Minnesota, recalled in an interview.

But instead of incorporating Lando into the study, Rosenhan dropped him from it.

Cahalan argues that Rosenhan was probably right about the conditions at many psychiatric hospitals in the early 70s, but that’s obviously not an excuse for inventing data. She has written a book about her findings titled “The Great Pretender” which will be published tomorrow.

It has been a tough couple of years for influential studies. Last year an expose of the famous Stanford Prison Experiment revealed that the participants were coached. One of the most famous moments from the experiment, in which a student playing an inmate appears to have a mental breakdown, was revealed to be bad acting by the participant: “If you listen to the tape, it’s not subtle. I’m not that good at acting.” And one of the guards who began getting tough with the “prisoners” admitted his own performance was based on a character from the film Cool Hand Luke.

Just last month another influential study that claimed to demonstrate gender bias through “blind auditions” for orchestras was debunked and shown to not really support the conclusions the authors, and especially the media, made for it. That study was cited more than 1,500 times and mentioned by Justice Ruth Bader Ginsburg in a dissent.

Here’s Rosenhan describing his experiment: