Sunday, November 17, 2013

Why Politics Make Us Stupid

November 17, 2013, 11:20 a.m.
Yes, Liberals, This Means You, Too

It turns out that the half-jocular line, "My mind is made up, don't confuse me with the facts," is worth more reflection than we may have realized.

Yale Professor of Law and Psychology, Dan Kahan, heads that school's Cultural Cognition Project, a national group of academic researchers who study how one's political or other group affiliations and opinions can alter an individual's perceptions of relevant data concerning, say, climate change or gun control. Nicole Ng, "Kahan Responds to Media Storm," Yale Daily News, Oct. 29, 2013.

This first came to my attention as a result of one of my many informative BBC program podcasts, this one from the program "More or Less: Behind the Stats" ("Numbers are used in every area of public debate. But are they always reliable? Tim and the More or Less team try to make sense of the statistics which surround us.").

This episode, titled "Does politics make us get our sums wrong?" was a exploration of Dan Kahan's research.

Kahan didn't take the easy way out, which would have been to choose a random or representative sampling of subjects from across our socio-economic-educational classes. Instead, he choose those with outstanding mathematical ("numeracy") skills. As he explains it, "We did tests and found that people who are more science literate and better able to make sense out of scientific data tended to be more polarized along cultural lines on issues like climate change or guns or nuclear power, not less. That’s not what you would expect if the problem were that people had a deficit in rationality — in that case, the people who are the most science comprehending among those different groups would be converging on their views consistent with the best evidence."

He began with an exercise that was political-value-neutral: two skin creams, one of which smelled like strawberries, the other like bananas. He told the participants that the strawberry cream was tested on 300 subjects, 200 of whom saw improvement in their rash, and 100 of whom did not. The banana cream was tested on a smaller group; it produced only 80 whose skin improved and 20 whose skin did not. Virtually all of those with high numeracy skills got the correct answer to the question, "Which was the better skin cream?" (If you're having trouble with this one, although 200 got improvement from the strawberry version, and only 80 from the banana, 200 out of 300 (200 plus the 100) is 66.6%, whereas 80 out of 100 (80 plus the 20) is 80%.)

On the other hand, when the exercise involved statistics regarding a politically loaded issue, such as gun control, the numeracy-gifted subjects came up with remarkably different rates of success. The subjects were divided into two groups, those who believe that crime rates come down when more people have guns, and those who believe that crime rates come down when there are bans on gun ownership. Both tended to respond according to their beliefs.

Let's assume, hypothetically, that the statistics again involved the 200-100 and 80-20 splits: out of, say, 100 cities 80 would have seen crime rates decline, and 20 would have seen no change.

When the correct answer supported the view of those who believed that more guns mean less crime, this high numeracy group scored about as well as they had on the skin cream exercise. However, when it did not, when the results seemed to show that gun bans did a better job of reducing crime, the percentage of those who came up with the correct answer dropped to as low as 35%.

I hasten to add, those who thought that banning guns would reduce crime produced exactly the same results. When the "correct" answer conflicted with their ideological opinion, only 35% of them came up with it.

We've seen reports of surveys that involve the provisions of a supposed legislative proposal. When Democrats are told that the provisions are in a bill from President Obama, the percentages that think it's a good idea are far higher than if they're told it comes from the Tea Party Republicans in the House. Republicans' responses show a similar disparity depending on the provisions' sponsor -- even though the specific provisions are precisely the same in all four tests. [Photo credit: Stan Honda / AFP-Getty Images.]

But Professor Kahan has now provided us with some scientific data that helps to explain how this happens. It turns out that the power of our predispositions is such that they can, even for the most mathematically gifted, scientific and data-driven among us, block out their ability to do simple math.

Oh yes, some liberals have read Kahan's data as supporting their assumption that conservatives are more stupid than liberals. Alas, not only are the liberals wrong about that, they have simply provided more evidence in support of his findings!

"My mind is made up, don't confuse me with the facts"? Ah, if only that were possible. Apparently, once our minds are made up it's highly unlikely that anything, including persuasive facts, will ever confuse us.


k9mensan said...

Doesn't seem like fair comparison either way because the sample sizes are so different. I am mathatically badly challenged so this is more a question than an argument. I get the point that evidence be damned and people chose what supported their beliefs — but what about the sample size?

Critic46 said...

There is also the problem that those who are scientifically aware have already formed opinions about guns based on existing data. They have already "made up their minds" because they have other data that shows a result differing from the hypothetical exams. For the test to work, they would have to find liberals who are uninformed.

Nick said...

Advertising Notice

Notice Regarding Advertising: This blog runs an open comments section. All comments related to blog entries have (so far) remained posted, regardless of how critical. Although I would prefer that those posting comments identify themselves, anonymous comments are also accepted.

The only limitation is that advertising posing as comments will be removed. That is why one or more of the comments posted on this blog entry, containing links to unrelated matter, have been deleted.
-- Nick