Beyond religion
When Prophecy Fails
The most famous study of failed prophecy, what it claimed, and what checking the archives just found.
The famous story
In 1954 a Chicago housewife, Dorothy Martin, channeled messages predicting a great flood for December 21 and rescue by flying saucer. Psychologist Leon Festinger and colleagues infiltrated her group to watch what happens when a specific prophecy meets a specific deadline. Their 1956 book When Prophecy Fails reported the result that made it a classic: when the flood did not come, the committed members did not abandon the belief, they doubled down, welcoming the press they had previously shunned and proclaiming their faithfulness had saved the world. The book
From this came cognitive dissonance theory: when evidence collides with a belief you have publicly committed to at cost, the cheapest repair is often not dropping the belief but reinforcing it.
The part almost nobody knows, and it matters
In 2025, researcher Diana Kelly published a peer-reviewed archival re-examination in the Journal of the History of the Behavioral Sciences, drawing on newly available materials, and found the textbook story badly oversold: the group appears to have largely recanted and dissolved rather than doubled down, and the authors' account misrepresents what happened. The paper and accessible coverage
Later research on other failed-prophecy movements had already found a mixed picture: some groups reinterpret and persist, some quietly dissolve, some fracture. "Disconfirmation always increases fervor" was never the reliable law the retellings made it.
Why tell the story this way
Because the corrected version teaches two lessons instead of one. The doubling-down pattern is real, you can watch it in the Millerite Great Disappointment of 1844, in endlessly recalculated end-times dates, and in the reinterpretation of the Missouri temple prophecy covered elsewhere on this site. But the second lesson is the sharper one: a vivid, satisfying study went essentially unchecked for seventy years because it confirmed what everyone wanted to believe about believers. Skeptics repeated it the way believers repeat faith-promoting stories, and it took one person actually reading the archives to notice. The failure mode this site describes does not belong to religious people. It belongs to people.
Why this should make you question
Two questions, and both matter. When your community's specific expectation fails, watch what happens next: is the miss counted, or reinterpreted? And when a story flatters your side, do you check it before repeating it? This page exists because the author's side failed that test for seventy years.