Tuesday, June 26, 2018

Differences in Belief Updating Depending on Belief Content

A study by Walter and Murphy in Communication Monographs shows that political misinformation is harder to correct than health misinformation. It also shows that the more educated someone is, the harder it is to correct. This both contradicts and coheres with things we thought we already knew.

We already knew that the more educated people were, the harder it is to correct misinformation they believe (here and here) and depressingly, even that increasing people's access to information only makes them dig in further; it doesn't matter if the extra knowledge is inside or outside our skulls, but where politics is concerned, we will recruit it to confirm what we already believed. But notice that most of these kinds of studies (good roundup here) are about politics. What about other kinds of beliefs?

That's the take-home from the Walter and Murphy study, and one interpretation is that it's identity protection and/or tribal affiliation that all this motivated reasoning is protecting - and that politics is very identity-defining. It's therefore interesting that when given the chance to improve one's own personal traits, people actually chose things things less central to their identity - for instance, musical ability over moral sense - which is of course completely absurd (study is in a post at Overcoming Bias which I'm having trouble finding and will update when I do.) If your moral sense is that much more important, why would you not choose to improve that? (If what you're really doing is reinforcing your identity, it makes perfect sense.) This is why Paul Graham gives the advice of shrinking your identity, to prevent all your identity-defining beliefs from biasing your thinking.

By this argument, in this sample, people's health beliefs were not as identity defining as their political beliefs. A way to test this would be to measure the importance of such beliefs to people's overall identity, and then see if the health-politics difference is as strong, and also if there's a good relationship to identity-centrality. There are lots of food extremists and antivaxxers whose health beliefs might be as central or more central to their identity than their other beliefs, including their political ones. You might also wonder why the U.S. got into its current failure mode of politics having become so central to identity (moreso than religion by many measures) - my guess is some combination of only two parties for 3x10^8 people, along with the decreasing significance of other meaning-providing communities and overlapping status hierarchies. I'm hardly endorsing religion as the answer, although there are plenty of people for whom the vacuum of meaning left by the decreasing role of religion has made them turn to politics - namely, young white working class people, who despite being less religious, voted enthusiastically for Trump.

No comments: