NPR interview on partisanship versus facts

I was interviewed by NPR's Shankar Vedantam for a segment that just aired on Morning Edition:

The story features my research with Jason Reifler on self-affirmation and misperceptions. Here's the title and abstract of our manuscript (PDF):

Opening the Political Mind?
The effects of self-affirmation and graphical information on factual misperceptions

People often resist information that contradicts their preexisting beliefs. This disconfirmation bias is a particular problem in the context of political misperceptions, which are widespread and frequently difficult to correct. In this paper, we examine two possible explanations of the prevalence of misinformation. First, people tend to resist unwelcome information because it is threatening to their worldview or self-concept. Drawing from social psychology research, we therefore test whether affirming individuals’ self-worth and thereby buttressing them against this threat can make them more willing to acknowledge uncomfortable facts. Second, corrective information is often presented in an ineffective manner. We thus also examine whether graphical corrections may be more effective than text at reducing counter-arguing by individuals inclined to resist counter-attitudinal information. Results from three experiments show that self-affirmation substantially reduces reported misperceptions among those most likely to hold them, suggesting that people cling to false beliefs in part because giving them up would threaten their sense of self. Graphical corrections are also found to successfully reduce incorrect beliefs among potentially resistant subjects and to perform better than an equivalent textual correction. However, contrary to previous research, affirmed subjects rarely differ from unaffirmed subjects in their willingness to accept new counter-attitudinal information

For more information on our research, please see our original article on the difficulty of correcting factual misperceptions, our New America report summarizing the literature on misinformation and fact-checking, and our study of race of researcher effects and the Obama Muslim misperception (all PDFs).

Comments

[LeahRosin]

In the NPR segment, the specific issue of gas prices was discussed. I think it's disingenuous to say that Democrats thought Bush could "do" something about gas prices... it's probably more accurate that a survey or question asked if people thought Bush (and his war in Iraq and subsequent instability in the middle east) was responsible for the higher gas prices. I'd love to see the actual question that was asked and the numbers of people who were polled. If you're comparing perceived responsibility with perceived ability to remedy (which is what the Republicans are saying Obama could do if he wanted to - just flip some switch and lower the price of gas), then I think you're comparing apples to oranges and it's a weak argument.

[bnyhan]

Hi Leah - here are the poll questions: "Is there anything the Obama administration/Bush administration reasonably can do to reduce gas prices?" (http://www.washingtonpost.com/politics/2012/03/21/gIQAk0IeSS_graphic.html). It's not a question about responsibility.

[Judi Knott]

Heard the morning edition story today and I am interested in better understanding your insights on how to use positive imagery to help individuals digest 'difficult' facts. I work in the retirement industry and believe your insights might be leveraged for many of our challenges as well...can you direct me to more information?

[David Jerde]

When I listened to the NPR segment yesterday, my immediate thought followed the vein of Leah's thinking. Bush had chosen to pursue a path of war, which affected the supply-side of the equation. To that end, I think it was probably right to assume that Bush had more responsibility over the price of gasoline. Not making a value judgement on the war, just observing that the war changed the supply side of the equation.
---------------------
Fast forwarding to the Obama administration, the economics of the world have changed. Of course, the Arab spring in general and the revolution in Libya in particular have created some supply-side issues (esp because Libya has so much sweet crude which is desired for gasoline production). More importantly, the demand side has changed drastically, with the emerging juggernaut of China consuming more gasoline than many were willing to guess c2006.
---------------------
Fiscal conservative, social moderate...

[Tim Ene]

I wonder what your study has to say about devices such as debate flowcharts (used by generations of high school forensics students):

A congressional candidate debate on climate change
http://www.persuasionpoints.com/CO2CD-EnergyClimate.htm

A philosophical debate on the role of government between progressive and conservative research organizations
http://www.persuasionpoints.com/coloradobudget.html

(double-click on individual arguments to see video of points being made)

To use studies such as yours as the basis for a social intervention, the real trick will be getting opposite sides to engage in way that leads to an (incremental) cultural shift.

[David in Cal]

I continue to think that Brendan would be better off if he referred to "beliefs" and not to "facts." First of all, "beliefs" is a more general category, and I think Brendan's results would apply to all beliefs, whether or not they're "facts".

OTOH referring to "facts" opens a can of worms.

1. Some of Brendan's examples concern beliefs, rather than facts. E.g., consider the question, "Is there anything the Obama administration/Bush administration reasonably can do to reduce gas prices?"

The time period for the price reduction is unspecified as is the amount of reduction. Also, there's the weasel word "reasonably". Thus, there's no right answer to this question. Brendan's research results are interesting, but they concern beliefs not facts.

2. The global warming questions are particularly weak. Using CNN as a model, one optiion offered was, "Global warming is a theory that has not yet been proven." What a bizarre choice! There are many theories relating to global warming. E.g.: the globe has been warming. The warming will continue. The warming will be catastrophic. The Hockey Stick theory. The cause of the warming is primarily man's activity. Etc. In particular, none of the choices get at the primary area of contention, namely, the degree of uncertainty.

Without specifying which global warming theory is being referenced, the question is meaningless. It marks the questioners are ignorant. How can Brendan and his co-author expect to change people's minds about global warming, when they begin by demonstrating that they're ignorant of the topic?

3. Cherry-picked questions. Brendan might have asked about job gains or losses from January 2008 to the present, rather than 2010, and global temperature change over the last 15 years, rather than 30 years. Had he done so, the factual answers would have been different.

4. Arbitrary definitions. Brendan wrote, "We define beliefs that global warming is either unproven or caused by natural factors as misperceptions..." Does Brendan have the right to make such a definition? I don't think he does. Although belief in AGW is widespread, it isn't necessarily that deep. One can see the lack of certainty in the response to actual policies. While most people may say they believe in CAGW, most people oppose the policies supposedly necessary to prevent he supposed catatastrophe.

As I said, if Brendan simply focuses on beliefs rather than facts, all these difficulties disappear and his general result still stands.

[David in Cal]

I can illustrate the difference between a very wide belief and a very high degree of belief using an urn model (a type of model loved by probability folks.)

Suppose an urn has 2 red balls and 1 black ball. One ball is to be drawn at random. If 10,000 experts are polled, most or all of them will naturally guess that the randomly drawn ball will be red. However, just because over 99% of the experts choose "red" as their best guess, that doesn't mean that the drawn ball is 99% likely to be red. On the contrary, the chance that the drawn ball will be red is only 66.7%.

In short, the percentage of experts favoring a certain view is NOT necessarily equal to the probability that this view is correct.