Farms, Findings, and False Causation: A Systematic Review of Allergy Prevalence

The dominant paradigm in allergy epidemiology is the "farm effect," which posits that exposure to farming environments provides a protective effect against allergies. This has led to research aimed at identifying specific environmental or microbial factors responsible for this protection. This preprint challenges this paradigm by proposing an alternative explanation: the observed protective effect is not a genuine biological phenomenon but a statistical artifact arising from methodological flaws, specifically collider-stratification bias and selection bias (a "healthy worker effect"), where individuals predisposed to allergies are less likely to remain in farming environments.

This preprint systematically deconstructs the evidence supporting the "farm effect" paradigm. Through a meta-analysis, it demonstrates that the purported protective effect vanishes in studies using representative random sampling, while persisting only in those with biased, stratified sampling. By identifying this fundamental methodological anomaly that has been largely ignored by the field, and by showing that selection bias can fully account for the observed data, the paper exposes a deep-seated failure within the current paradigm. It argues that the puzzle researchers have been trying to solve—finding the protective "needle in the farm dust"—may not exist at all.