l5-skin-in-the-game_psychology-replication-crisis

The replication crisis refers to a large-scale effort where only 39 of 100 published psychology papers successfully replicated. This 61% failure rate indicates that many peer-reviewed findings were artifacts of small samples, flexible analysis, and publication bias rather than real effects.

Academic incentives reward publishing novel results, not true ones, and researchers face no consequences for being wrong. Since most novel findings are false by base rate, a system selecting for novelty over reliability systematically produces attractive falsehoods.

Effects like priming, anchoring, and framing either failed to replicate or replicated at much smaller magnitudes than claimed. The field's incentive structure favored surprising publishable results over rigorous verification, producing overconfidence without actionable knowledge.

Many heuristics pathologized as irrational are rational once ergodicity and tail risk are understood. What looks irrational from an ensemble perspective is often the only sane strategy for an individual who cannot afford ruin over time.

Distrust findings when researchers bore no downside for being wrong, when results are surprising and novel, and when no independent replication exists. The base rate of false positives in novelty-driven publication systems is structurally high.