Stop Confounding Yourself! Stop Confounding Yourself!

Critiques a headline-grabbing study claiming childhood bullying causes bad adult outcomes even after "controlling for confounders." Scott's point: you only ever control for an imperfect measurement of the confounders you happened to think of - and a bully's instinct for who can safely be targeted captures more about a child's real vulnerabilities (height, actual social skills) than a teacher's checklist (the Bristol Social Adjustment Guide) ever could. So residual correlation may be unmeasured confounding, not causation; an RCT-style intervention would settle it.

Quality 64 · Strong. A sharp, broadly-applicable methodological point made vividly through a real study - the limits of "adjusting for confounders" - but short and narrow (one study). Strong.

Claude’s paradigm shift 30 · Slight. Residual confounding / imperfect-proxy bias is well known in epidemiology; the contribution is a memorable framing, not a new idea. Slight.

Real-world impact 2 · Minor. A vivid, broadly-applicable methodological point on the limits of 'controlling for confounders'; a clarifying idea within the discourse, with no material-world reach → RWI 2.