Prisons Are Built With Bricks Of Law And Brothels With Bricks Of Religion, But That Doesn’t Prove A Causal Relationship

A correlation-vs-causation debunk (Blake-riff title). An Alternet piece + a journal editorial spin a Danish study (Hjorthoj 2013) -- which compared psychiatric patients to HEALTHY controls -- into 'psychiatric care might cause suicide' (hospital admission '44x' the risk). Scott shows this is obvious selection, not causation, and dismantles the editorial's misuse of the epidemiology 'strong + dose-dependent + plausible-mechanism -> causal' criteria with parallel absurdities (police 'cause' murder at 10,000x relative risk; ice cream 'causes' heatstroke; detention 'causes' dropout). Grants a buried valid point -- the justice and psychiatric systems CAN contribute to harm (forced commitment), just not provably via this methodology -- but concludes the paper 'contributes nothing.'

Quality 68 · Strong. Strong (68). A sharp, clear, memorably-structured stats-literacy debunk (the murder/police parallel is genuinely illuminating) with characteristic fairness (the steelman ending). The core lesson (correlation != causation) isn't novel, but the application and the takedown of the misused epidemiology criteria are crisp. Scored on merit without early-cohort compression (2013).

Claude’s paradigm shift 45 · Moderate. Moderate (45). Correlation != causation is old; the fresh bit is the specific debunk of the 'strong/dose-dependent/plausible -> causal' criteria misuse and the parallel-examples device. Moderate novelty.

Real-world impact 2 · Minor. 2 -- minor/within-blog. A stats-literacy / anti-anti-psychiatry debunk of a specific study; within-discourse reach, no material footprint.