A Hacker News links post revisits the shaky foundations of classic psychology studies. The piece highlights how a new analysis of Leon Festinger's 1956 cognitive dissonance study, 'WHEN PROPHECY FAILS,' suggests the original findings might be flawed. Up to half the cult members Festinger observed were likely undercover researchers, and the cult's reaction post-failure wasn't a simple doubling-down.

This joins other high-profile psychology findings under scrutiny, like the car crash memory experiment. That iconic study, cited nearly 4,000 times, showed verb choice dramatically altered speed estimates. A 2025 replication with 446 participants found no effect, challenging its robustness.

The author also discusses their own research on public opinion shifts, noting a reanalysis found people are actually quite good at estimating change, contradicting their initial findings.