When researchers mis‑specify the functional form of control variables (for example, modelling age as a simple quadratic), they can manufacture associations between exposure and outcome that disappear under appropriate adjustment. Those artefactual findings travel from papers into press and policy debates, creating misplaced alarm or misdirected regulation.
— Recognizing and fixing control‑form errors is essential to prevent bad science from driving public fear, regulatory action, and media narratives about health risks.
Cremieux
2026.04.28
100% relevant
The article reanalyses a public dataset where authors claimed higher urinary glyphosate linked to lower bone mineral density; replacing the paper’s quadratic age control with a more flexible age adjustment eliminated the association.
← Back to All Ideas