Replication and Meta-Analysis in Parapsychology

This is a big deal: a statistician from UC Davis got a paper about psychic research published in one of the most prestigious statistics journals around, complete with seven expert commentaries. Jessica Utts walked through a surprisingly common scientific blunder: when a replication study finds nothing significant, researchers often wrongly declare the original effect was fake, when really the new study just lacked enough statistical power (the ability to detect small effects). She then laid out meta-analyses (studies that pool many experiments together) across four types of psychic phenomena, and the numbers are genuinely striking. In telepathy experiments using the ganzfeld method (a sensory-deprivation setup), participants hit the right target 34.4% of the time versus the expected 25% by chance alone, with odds against luck of about 1 in 20,000. That effect size is actually triple the effect of aspirin in preventing heart attacks. The expert commentators were split but intrigued: two statisticians calculated the evidence at 100-to-200-to-1 in favor of something real happening. Utts concluded that a genuine anomaly exists that demands explanation.