psychiatry ward attribution bias

If Sanity and Insanity Exist, How Shall We Know Them?

So begins a fascinating and frightening study from 1973 that demonstrates how much circumstance dictates our perception of others, reminding us that recognizing what’s normal and what’s abnormal isn’t always obvious.

In 1973 a young man went to see a doctor complaining that he had been hearing voices. In particular he reported that he had been hearing the words ’empty’, ‘hollow’ and ‘thud’, spoken by an unfamiliar voice. After some questioning by the clinician the man was admitted into the psychiatric ward at the local hospital with a diagnosis of schizophrenia.

This was his plan and it went better than he had anticipated.

This man was actually a Stanford University psychologist named David Rosenhan who was a healthy individual with no history of psychiatric disturbance and the symptoms he reported were fake. He, one of 8 individuals (5 men, 3 women) who were observers in a fascinating experiment to see if they could fake their way into psychiatric admission and then see how long it would take for them to be discharged from the hospital. The challenge was to now convince the hospital staff that they were normal.


On being sane in insane places

The aim of their experiment, which was conducted across 12 different hospitals scattered around the east and west coasts of the US, was to see how the context of being in a psychiatric institution, and having a diagnostic label, would influence how their behavior was interpreted. Or to put it another way, whether the system of the time could effectively tell the “sane” from the “insane” in a psychiatric setting.

Given that once they arrived on the ward they acted completed normally and told staff that the voices had now disappeared you might imagine that it wouldn’t take too long for them to be found out as being pseudopatients.

Not so.

The average stay in the hospital across the group of individuals was 19 days, the longest was 52.

Misinterpretating normal

So how did this happen? Well, once each individual had passed the first hurdle of getting themselves admitted, they were then left to their own devices to see how quickly they could get themselves discharged by acting as normal as possible. But this actually was more difficult than it sounds given the fact that they were in a highly restrictive and rather dull environment.

The main problem was that they found their “normal” behaviors were time and time again misinterpreted, simply because of where they were. When they paced the corridors in boredom, they were asked whether they were doing so because they were feeling nervous. When they sat in a group outside the cafeteria waiting for lunch, they were ascribed as having “oral acquisitive” habits by a passing psychiatrist.

See related post Lunatic Asylums in the 1800s.

In other words, the context of the hospital and the label they had been given were so powerful that it caused their normal behaviors to either be overlooked, or distorted. And instead of attributing their behavior to the restrictive and dehumanizing hospital environment itself, the staff attributed it to the person’s mental state of mind. This is a nice example of fundamental attribution error where people underemphasize the role of external factors when trying to explain other people’s behavior.

What’s more, when the observers were finally discharged, it wasn’t because the staff and psychiatrists suddenly recognized them as being “normal”. It was with a diagnosis of schizophrenia in remission – they were presumed to still have the condition in some form, and it was expected to return again at some later date.

The conclusion of the study, reported in the journal Science [1], was that “it is clear we cannot distinguish the sane from the insane in a psychiatric hospital”. What’s more, in a follow-up experiment, Rosenhan told staff at one hospital that in the next 3 months there would be some pseudopatients coming through the institution and asked them to rate how likely it was that each patient they had on the ward was a pseudopatient, using a 10-point scale. What he found was that out of 193 patients admitted, 19 were rated as being possible pseudopatients by both a psychiatrist and another member of staff, when in fact none had been (at least to the knowledge of the experimenter). Whether these 19 were actually “sane” will never be known, but a frightening prospect all the same.

The persistent challenge of attribution bias

There are of course a number of caveats to this study. We aren’t in 1973 anymore. The study was being conducted in the era of DSM-II when schizophrenia didn’t have as clearly defined criteria as it does today under DSM-5. And psychiatric institutions have changed considerably since then, in part due to the observations recorded in this influential study which caused psychiatric hospitals worldwide to review their admission and training procedures.

See related post: Schizophrenia: A Metaanalysis of Resting-State EEG Studies.

But what hasn’t changed is the way that people with mental health diagnoses are given a subjectively-defined label which is often associated with personal, social and legal stigma. Once they are given this diagnosis, then it is hard to ever leave it behind. What’s more, many disorders, and in particular ADHD and autism, are still diagnosed based on evaluations and observations of context-dependent behaviors which are naturally prone to this kind of attribution biases. For example, is a child who’s distracted and not listening to an adult showing ADHD tendencies, or are they just bored and would rather be playing outside?

This isn’t to diminish or deny the extreme anguish and distress that individuals with mental health disorders experience, or the fact that mental health symptoms often do re-occur over an individual’s lifetime. But it is a reminder that subjective diagnostic systems like this are open to bias, whatever disorder you are looking at, and should always be approached with caution, especially when clinical judgements are being made based on an individual’s behavior. And when so many symptoms of mental health lie close to the line dividing normal or typical behaviors, from those considered to be atypical or abnormal, clinicians face a momentous challenge of avoiding both false positive and false negatives, and of not falling foul to the many biases which unwittingly inhabit the human mind.

See related posts: The Challenges of Mental Health Diagnosis and  The Murky Borders of Brain Dysfunction



[1] Rosenhan, D. (1973). On Being Sane in Insane Places. Science, 179(4070), 250-258. doi: 10.1126/science.179.4070.250

Leave a Reply