On the day that Barack Obama was elected President, I asked readers of my Comment Central blog to let me know, in a few words, their greatest hope for him. I received thousands of replies. And most of them were touching, optimistic and only a little naive.

There were two types of reply that left me perplexed. The first were the hundreds of people who wrote to say that they hoped that the new President would be shot. You can’t read that over and over again without being perplexed about human nature. A second type expressed the hope that Obama would produce his birth certificate. This surprised me, because until then I had been unaware of the mad conspiracy theory that Obama was not eligible to be President, having been born in Kenya rather than America.

I thought it would be a good idea to seek out and post on Comment Central documentary evidence that Obama had indeed been born in the US. I needn’t have wasted my time. Judging by the replies, the evidence made not the slightest difference. I was forced to conclude that the bonkers idea that Obama was not legally allowed to be President was adopted merely to reduce the cognitive dissonance of those who believed that such a thing — an African-American becoming chief executive — could not happen.

The same sort of issue — the persistence of misperceptions in the face of evidence — has also been intriguing Brendan Nyhan, of Duke University, North Carolina, and Jason Reifler, of Georgia State University. And they have published two fascinating papers providing the results of experiments that they conducted into whether it is possible to correct such errors of fact.

Their conclusions are not a cause for optimism. Nyhan and Reifler provided their research subjects with newspaper articles on diverse subjects — the Iraq war and stem-cell research among them — which incorporated errors of fact, such as, for instance, that weapons of mass destruction had been discovered in Iraq. They surveyed the group for their views and then distributed further articles. These put right the misperception. Or should have done. But that is not what the academics found when they re-examined their subjects.

The second survey of attitudes showed two things. First, correcting a misperception doesn’t really work when the original misperception fits snugly with the subject’s ideology. Second, and worse still, attempting to correct errors often produces a backlash, with the error becoming more firmly believed.

Three weeks ago Nyhan and Reifler published results of their experiments specifically on President Obama. They were investigating the persistence among a minority — about 12 per cent, or so — of the false idea that the President is a Muslim. Again, they showed how difficult such notions are to shift among those inclined to believe them.

The statement “The President is not a Muslim” is ineffective. Indeed, it makes the belief, the association between Obama and Muslim, stronger. The statement “The President is a Christian” appears to work better. This may account for the fact that, as one US commentator has recorded, Obama talks more of his Christianity in public than George W. Bush did.

But there is a wonderful little twist to the data. The academics found, however, that “The President is a Christian” worked in only one condition — if a non-white person was present in the room and the survey respondent was trying to please him.

daniel.finkelstein@thetimes.co.uk

Times Recommends

• You created the job. You should do it
• A swindler with a ready maid plan
• Charming to pillory disabled Brown