That might be depressing, but it’s not unexpected. In a new study, David Gal and Derek Rucker from Northwestern University have found that when people’s confidence in their beliefs is shaken, they become stronger advocates for those beliefs. The duo carried out three experiments involving issues such as animal testing, dietary preferences, and loyalty towards Macs over PCs. In each one, they subtly manipulated their subjects’ confidence and found the same thing: when faced with doubt, people shout even louder.

Gal and Rucker were inspired by a classic psychological book called When Prophecy Fails. In it, Leon Festinger and colleagues infiltrated an American cult whose leader, Dorothy Martin, convinced her followers that flying saucers would rescue them from an apocalyptic flood. Many believed her, giving up their livelihoods, possessions and loved ones in anticipation of their alien saviours. When the fated moment came and nothing happened, the group decided that their dedication had spared the Earth from destruction. In a reversal of their earlier distaste for publicity, they started to actively proselytise for their beliefs. Far from shattering their faith, the absent UFOs had turned them into zealous evangelists.

The case study inspired Festinger’s theory of “cognitive dissonance”, which describes the discomfort that people feel when they try to cope with conflicting ideas. Festinger reasoned that people will go to great lengths to reduce this conflict. Altering one’s beliefs in the face of new evidence is one solution but for Martin’s followers, this was too difficult. Their alternative was to try and muster social support for their ideas. If other people also believed, their internal conflicts would lessen.

Festinger predicted that when someone’s beliefs are challenged, they would try to raise support for those beliefs with paradoxical enthusiasm. Amazingly enough, during the intervening half-century, this prediction has never been tested in an experiment – that is, until now.

In their first experiment, Gal and Rucker asked 88 students to write about their views on animal testing for consumer goods, but only half of them were allowed to use their preferred hand. This may seem random, but previous studies have shown that people have less confidence in what they write with the hand they’re less comfortable with. Indeed, that’s what Gal and Rucker found in their study. When asked later, the volunteers who didn’t use their dominant hand were less confident in their views.