Rosenhan (1973) Flashcards
Aims of the Experiment
They tried to test whether psychiatrists could reliably tell the difference between people who are sane or insane.
Procedure of the Experiments
Study 1-
8 pseudopatients were given pseudonyms and placed into 12 psychiatric institutions. They gained an appointment and stated they heard voices (saying- ‘Thud’, ‘Empty’ and ‘Hollow’). Each participant was admitted, and then stopped showing signs.
Study 2-
The research was complete but scientists failed to believe the findings, so Rosenhan announced more pseudopatients would be released again.
Results
Study 1-
All but one of the patients received a diagnosis of schizophrenia. This shows an unreliability in diagnosis.
Study 2-
Judgements were given to 193 patients, but none were actually present.
Conclusion
He claims that they cannot distinguish between illnesses, suggesting low reliability. The study also showed the negative impact of such a label and how powerless the patients feel.
Evaluative strengths
He used a large range of hospitals (from different states and economic backgrounds), which improves the generalisability.
The applications were huge and created a change in diagnosis. This study is now a compulsory part of training in nursing and medicine.
11 out of 12 diagnoses were consistent, which supports the reliability of the study.
He practiced confidentiality, and all staff members and hospitals remained nameless.
Evaluative weaknesses
12 is quite a small sample size, so generalisability was low. Also, only USA was used, so the generalisability spanned just to America.
The participants did not follow the standardized instructions (a ninth participant was removed after breaking procedure, a student did his homework revealing his psychology, one even became involved with a nurse).
The fact they faked illness lowered validity. People wouldn’t usually lie to gain entry, so hospitals may believe all people. This lacks ecological validity (they weren’t actually insane).
The hospital staff or other patients couldn’t give consent.