Validity and Reliability

Question 1

Rosenhan (1973) - used DSM II

Answer 1

Aim - to investigate whether psychiatrists in a naturalistic setting could tell the difference between sane and insane people

Procedure:
1) 8 mentally healthy subjects checked themselves into mental hospitals
2) during interview with psychiatrist they claimed to hear voices in their head that said words like “empty” and “thud”
3) Were only told to lie about this nothing else
4) Upon admission, they acted normally and reported that voices had stopped, sought to leave, wrote down observations secretly

Results:
- 7/8 patients admitted for diagnosis of schizophrenia
- when discharged, all diagnosed with “schizophrenia in remission”
- took patients on average 19 days to get out of hospital by their own means (7-52)
- none of the staff suspected they were healthy
- normal behavior often misinterpreted as symptom of disorder
- staff of avg spent 7min a day with patient, largely ignored

Conclusion - psychiatrists lack ability to distinguish mental disorder from sanity, questions validity of diagnoses
explained by labelling (anchoring effect + confirmation bias)

Question 2

Validity

Answer 2

• refers to correctness of a diagnosis
• construct validity: whether a tool or system measures the concept it claims to measure
• content validity: whether diagnosis is influenced by situational or contextual factors

Question 3

Reliability

Answer 3

• stability of diagnosis over time, place, circumstance, and argument between clinicians
• inter-rater reliability: can be asked by asking more than one practitioner to attempt to make a diagnosis
• test-retest reliability: whether the same person will receive the same diagnosis if assessed more than once

Question 4

Nicholls et al (2000)

Answer 4

• two practitioners asked to use DSM-IV, ICD-10 or the Great Ormond Street Hospital’s own diagnostic system (GOS)
• used to diagnose 81 children with eating problems who had come to a specialist clinic
• correlational study in the sense since reliability was measured based on rates of agreement between the two practitioners

Results:
- When the practitioners used ICD-10, inter-rater reliability was only 0.357
- 0.636 when they used DSM-IV
- 0.879 for the GOS system.
- shows that the GOS system is more reliable than either of the other two.

Conclusion
- researchers suggest that the success of the GOS system is because it is specifically designed for use with young children.