Implicit Association Tests and Suicidality

Reviews Nock & Banaji (2007), which used an Implicit-Association-Test variant (self/other paired with self-harm vs not-self-harm images) to detect suicidal intent in psychiatric patients. Scott is genuinely interested in a non-self-report suicide screen but sharply skeptical of THIS study: the self-harm-image confound (it may just detect past cutters), the unconvincing 'controlled for self-harm' claim (presence but not amount), and especially the weak prospective-validity analysis (comparing the 2 later attempters' mean to all 71 non-attempters, who weren't matched on prior history). The methodological critiques are sound. But the post's framing explicitly endorses the IAT itself as 'really powerful' and wasted on 'nebulous social justice projects' (the CIA-double-agent-IAT aside). REPLICATION CAVEAT: the post leans on the IAT being a valid, powerful instrument; the IAT (implicit-bias paradigm) has since shown weak test-retest reliability and poor predictive validity, undercutting that premise. The study-specific critiques still hold; quality-as-it-reads-now is docked and confidence set to Med.

Quality 58 · Solid. Competent, genuinely skeptical study-review with a sharp prospective-validity teardown, but a minor 2013 journal-club-style post whose enthusiasm for the IAT-as-powerful-tool has aged poorly. Docked for leaning on the now-shaky IAT paradigm -> high-Solid, Med confidence.

Claude’s paradigm shift 35 · Slight. Slight-to-Moderate. Reviews someone else's application of an existing instrument; little that was conceptually new.

Real-world impact 2 · Minor. A skeptical journal-club review of an IAT-based suicide-screen study; a minor methodological critique within the discourse (now itself dated by the shaky IAT paradigm), with no practice change → RWI 2.