The Pew Research Center for People and the Press has a new report out with a fairly perplexing finding. The percentage of people who think President Obama is a Christian has declined sharply over the past year. In three studies from March 2008, October 2008, and March 2009, the percentage thinking he was a Christian stayed between 47-51%. Then, in the most recent study, it dropped to only 34%. The study was in the field from July 21-August 5th, so I think much of this predates the recent “Ground Zero Mosque” flap. Also, everything at follows is with the caveat that any one survey can always produce a result from the tails of a distribution, so we’ll want to see other surveys replicating this result before we can confidently conclude that public opinion in this regard has indeed shifted. But this is a blog, so let’s just posit for the moment that it indeed has shifted and continue with the discussion.
My question for those of you who study public opinion is the following: don’t we normally expect that factual knowledge about individuals increases over time as that person spends more time in the public eye? And it’s not like this issue wasn’t politicized from the start – opponents of Obama tried to make an issue of his religion during the campaign. So why would we expect fewer people now to know that he is a Christian?
Here’s the best answer I can come up with, and it’s a bit troubling. If we take a Bayesian perspective on the whole thing – people start with a prior, and then update as they get more information – here is the one stylized fact we actually know: Obama’s approval rating has dropped over the past year. So if we imagine that people start with a prior that Obama is a Christian, but it’s not particularly strong. Then time passes and they like him less – that’s the new piece of information. So then they have to update their belief as to whether Obama is a Christian, and lo and behold, fewer people think he’s a Christian. From this perspective, it suggests that – given enough initial uncertainty about a person’s religion – liking that person less makes one less likely to think he is a Christian. Interestingly, the drop in “Obama is a Christian” seems to be evenly split between thinking he is a Muslim and not knowing (other holds steady at 2%).
I’m open to other suggestions, explanations, etc. Also interested in whether readers have any other good examples of knowledge of factual information about a person declining as the person gets better known by the public?