The Efficacy Of Everything In Psychiatry In One Graph Plus Several Pages Of Dense But Necessary Explanation

Discussion of Huhn et al (2014)'s graph of effect sizes for drugs vs therapy across psychiatric disorders. Conveys useful empirical content (effects mostly ~0.5, comparable to general medicine -- debunking 'psychiatry is less effective,' per Leucht) with characteristic honesty: he admits he doesn't fully grok standardized mean differences on non-normal distributions, flags three surprises (drugs > therapy for borderline and insomnia -- but tiny samples; drugs better at preventing relapse than acute episodes -- with a sharp selection-bias self-correction). The killer methodological point: low-quality psychotherapy trials showed effect size 0.74 vs 0.22 for high-quality ones ('proof it's way too easy to inflate positive findings'), and psychotherapy research 'isn't even good enough to have publication bias.'

Quality 71 · Strong. A solid, informative journal-club-style post with genuinely sharp methodological observations (trial-quality inflation; the publication-bias quip) and honest statistical humility -- but fundamentally a discussion of someone else's graph. Strong.

Claude’s paradigm shift 40 · Moderate. Moderate. Analysis plus caveats on others' meta-analytic data; the trial-quality point is sharp but not novel.

Real-world impact 2 · Minor. A solid journal-club post on Huhn et al's effect-size graph conveying useful empirical content (psychiatric effects mostly ~0.5, comparable to general medicine) with sharp methodological observations and honest statistical humility — but a discussion of someone else's graph. Conceptual influence within psychiatry discourse, no material change — low RWI.