1/ The PHQ-2 & PHQ-9 are those ubiquitous depression screening questions in medical offices. New article tells us they're "validated" against structured interviews used in research. But... what are those interviews validated against?
It's not the snarky
jamanetwork.com/journals/jama/…
It's not the snarky
jamanetwork.com/journals/jama/…
2/ question it may appear to be. Seriously, what were they validated against? The PHQ screeners were designed so non-experts without training in psych could make psych diagnoses by following the instructions. But here's the kicker... the "gold standard" structured interviews
3/ (like SCID) are also nearly always administered by non-experts, typically research assistant or students—not by psychiatrists or psychologists. So the interviews were *also* designed so non-experts could make psych diagnoses by following paint-by-number instructions.
4/ The questions in the PHQ and the structured interviews ask about the same checklist of things. In practice, there's no actual expertise involved in the assessments on either side of "validation" equation. Responses to rote questions administered by non-experts
/5 in one office predict responses to basically same rote questions administered by non-experts in another office. That is literally how it's typically done. Then ppl speak solemnly about "established validity." Few look behind the word "validity."
6/ Researchers will ask, what was questionnaire validated against?" And answer is usually another questionnaire. No asks next question... "And what was THAT validated against?"
Because answer is usually a tautology.
End of today's psychometrics lesson.
Because answer is usually a tautology.
End of today's psychometrics lesson.
• • •
Missing some Tweet in this thread? You can try to
force a refresh
