Guns And States

Independent data journalism on guns and homicide ('reverse voxsplaining'). Shows Vox conflates 'gun deaths' (homicide + suicide) - the gun/homicide link is weak, the gun/suicide link strong; re-runs the Harvard Miller-Azrael-Hemenway studies and finds their 2002 gun proxy (FS/S) actually measures Southern-ness, while the 2007 study (real CDC ownership data + a Southern dummy + confounders) holds up as a likely-causal gun-homicide correlation; argues most of the US-vs-Europe murder gap is a 'culture of violence' (Southern/honor culture, robbery rate), not guns; and Fermi-estimates that Australia-style control might avert ~2,000 homicides/year, probably worth it. Calibrated confidence levels throughout, data shared.

Quality 81 · Excellent. A rigorous, honest, original re-analysis - he ran the regressions himself and caught the FS/S-measures-Southernness artifact - that both debunks and steelmans and reaches calibrated conclusions; a model of reasoning about a politicized empirical question. Capped just below the top by being explicitly tentative and correlational. Excellent.

Claude’s paradigm shift 42 · Moderate. Sharpens the gun debate with a fresh decomposition (suicide vs homicide) and a culture-of-violence quantification, but applies standard regression reasoning rather than introducing a new frame. Moderate.

Real-world impact 2 · Minor. A rigorous, original re-analysis of the guns/homicide data (he ran the regressions himself and caught the FS/S-measures-Southernness artifact) — a model of reasoning on a politicized empirical question, but explicitly tentative, correlational, and discourse-level with no material or policy change — low RWI.