Milk and cooking oil look far more dangerous on a phone screen than in a lab notebook. That contrast is not an accident. Large cohort studies and meta-analyses, tracking hundreds of thousands of people, repeatedly show no consistent signal that moderate intake of dairy or common vegetable oils drives spikes in cancer incidence, once confounders such as smoking, obesity and alcohol use are controlled with multivariable regression.
What social feeds sell instead is a story. One dramatic tumor, one grieving family, one before‑and‑after photo. Fear compresses statistics. A single anecdote, drenched in detail, hijacks availability bias and makes rare outcomes feel routine, while relative risk changes under five or ten percent vanish from memory. In nutritional epidemiology, dose–response curves, confidence intervals and randomized controlled trials carry the real weight of evidence, yet they are visually dull and cognitively slow.
The harsher truth is that the villains are mostly boring. Ultra‑processed diets, chronic inflammation, tobacco, excess body fat and ultraviolet radiation account for a large share of preventable cancers, according to major oncology registries and pooled hazard ratio estimates. That pattern exposes an uncomfortable quirk of human cognition: the brain is wired to chase vivid threats, not probabilistic ones, which is why a viral warning about a splash of milk travels faster than a sober table of relative risks.
