In observational nutritional epidemiology, self-reported food frequency questionnaires generate tiny relative-risk ratios indistinguishable from socioeconomic and healthy-user confounding, rendering non-randomized dietary correlations scientifically uninformative.