AO3 SZ Classification

• What is reliability?

  Consistency
• What is validity?

  Accuracy
• Do sz classifications have high or low reliability?

  High
• Who found support for high reliability?

  Jakobsen et al
• What did Jakobsen et al. do?

  Tested the reliability of the ICD 10 classification system
• What was the sample in Jakobsen‘s study?

  100 Danish with history of psychosis
• What was the reliability Jakobsen et al. found?

  98 %
• Who found high reliability for diagnosis of sz in 180 individuals using DSM-5?

  Osorio et al.
• Osorio et al. (2019):

  Inter-rater reliability: 97 %

  Test-retest: 92 %
• Who found evidence of low validity in the diagnosis of sz by comparing ICD-10 and DSM-5?

  Cheniaux et al.
• What was the procedure of Cheniaux et al.’s study?

  2 psychiatrists independently assess the same 100 clients using DSM-5 and ICD-10
• What were the findings of Cheniaux’s study?

  In total, across both psychiatrists, use of ICD produced 68 positive cases while the use of DSM-5 produced 39 cases
• What does the findings in Cheniaux’s tell us about sz diagnosis?

  Disagreement between diagnostic manuals means some diagnosis must have been in preset (not valid). SZ is either over or under diagnosed according to diagnostic system, suggesting low validity
• Who conducted a study where fake volunteers tried to gain admission in hospitals?

  Rosenhan
• In Rosenhan’s study, 8 sane volunteers attempted to gain admission to 12 different hospitals across 5 states in the US. They had a clinical interview and complained that they had been ‘hearing an unfamiliar voice’ saying words like ‘empty’, ‘thud’ or ‘hollow’
• What were the findings of Rosenhan’s study?

  7 out of 8 volunteers were diagnosed with schizophrenia and admitted. None of the staff recognised that they were normal and they stayed in the hospital for an average of 19 days and diagnosed with ‘sz in remission (improving)’
• What was Rosenhan’s follow up study?

  A hospital was told that during the next 3 months, 1 or more fake patient would attempt to be admitted. In reality, no fake patients would attempt to be admitted
• What were the findings in Rosenhan’s follow up study?


  Of the 193 patients, 41 were deemed to be ‘fake’ by a psychiatrist. This suggests that psychiatrists cannot reliably tell the difference between sane and insane. This suggests the validity of diagnosis was low