In 1997 a Taiwanese group called Chen Tao — God's Salvation Church — moved to Garland, Texas, a suburb of Dallas. Its leader, Hon-Ming Chen, a former sociology professor, had determined that Garland was significant because the name sounded like God Land.
The group bought houses on the same street, dressed in white, and made a prediction of unusual precision.
On 25 March 1998, at 12:01 a.m., God would appear on television — on channel 18, on every set in North America, regardless of whether it was tuned there or even switched on. Then on 31 March, God would descend physically at Chen's house, in a form resembling Chen himself, and would speak to those assembled and shake their hands.
Chen invited the press. Dozens of journalists came, and a national audience watched what was, in effect, a live replication of the most famous field study in social psychology.
Festinger, Riecken and Schachter had infiltrated a Chicago group in 1954 whose leader predicted a flood and a rescue by flying saucer, and documented what happened when the date passed. Their finding was that under specific conditions — deep conviction, public commitment, an unequivocal disconfirmation, and social support — belief does not collapse. It is reinterpreted, and proselytizing frequently increases, because recruiting others restores the consensus that disconfirmation damaged.
The specified conditions in Garland were met exactly. Members had moved countries, sold property, uprooted families. The commitment was maximally public. And the disconfirmation was unambiguous.
Nothing appeared on channel 18.
What Chen did next is unusual and is the reason this chapter is worth reading. On 31 March, when God did not descend, he stood before the assembled press and said that his prediction had not been fulfilled, that his statements could be considered nonsense, and that he was willing to be stoned or crucified as a false prophet.
Nobody took him up on it. And most of his followers stayed.
Some accepted a reinterpretation — that God had in fact appeared but in a form not everyone could perceive, or that the prediction had been a test. Others simply stayed with the community. The group later relocated, revised its doctrine, and continued in reduced form.
The finding worth extracting is about the relative cost of two revisions. Abandoning a belief that your identity, relationships, home and finances are built on requires dismantling all of it. Revising the interpretation of a single event requires almost nothing. People take the cheaper option, and they are not being stupid when they do — they are minimizing total disruption, which is what the mind does with contradictions generally.
Which is why disconfirming evidence, on its own, is among the weakest instruments available for changing anyone's mind about anything.