This is the second post inspired by a short essay in the New York Times, “When beliefs and facts collide,” by Brendan Nyhan, whose essay itself discusses a 49-page paper by Dan M. Kahan that’s in press in Advances in Political Psychology. Kahan is a professor of law and psychology at Yale. (Kahan’s paper, in advance form, can be downloaded free at the link at bottom). All the figures and quotes in this post, save for one at the end, come from the longer paper.
The subject of both pieces is why so many Americans deny palpably true science, in particular evolution and human-caused global warming. Both of these “theories” are supported by mountains of evidence (no rational scientist would deny evolution, and something like 97% of climate scientists also accept that human activities are making the Earth warmer). My previous post, a few days ago, was on evolution; today’s is on climate change, the major topic of Kahan’s analysis.
In this case, Kahan administered a questionnaire to about 2000 American, asking them not what they thought about climate change (and by that I mean anthropogenic global warming), but about about what they thought scientists believed about climate change. But they also asked people a few questions about their own opinions about global warming, and also some questions to determine whether the respondents fell on the liberal or conservative side of the political spectrum. The graphic analysis is the same as in the post on evolution I published earlier.
Here are a few of the questions in the assessment of climate-change knowledge (the correct answer is underlined, and the table also gives percentage of people giving the correct answer as well as the question’s source):
The first question was the hardest; the answer is “no rise in sea level”, as the Arctic ice cap floats on the sea, like an ice cube in a glass of water, and so the water it displaces will be exactly replaced by its melt. Like water and ice in a glass, the melting ice won’t change the water level. The Antarctic ice cap, however, does not float—it rests on a continental land mass. So when the Antarctic ice melts, as with any continental ice, the sea levels will rise.
The percentage of correct answers was combined into an index of “Ordinary Climate-Science Intelligence” (OCSI), which measures how much people know about what “most scientists believe.” And, as with the evolution question, for any given question the proportion of people giving the correct answer to a single question is correlated with the overall OCSI score. With harder questions, though, only those who know the most get them right. You can see this in the following example of four questions asked, with the percentage of people getting that question “right” (i.e., demonstrating what climate scientists know) shown on the Y axis, plotted against the overall OCSI score, on the X-axis:
The last two questions are hard ones for the layperson (the next-to-last was hard for me!), and so scores rise exponentially as general knowledge increases; i.e., the savvier people are the only ones who get them right.
And as with evolution, the more one knows about what scientists think (i.e., the higher your OCSI score), your expectation is that your own opinon would increasingly concur with the scientists that Earth is getting warmer because of human activities. But that wasn’t the case! When people were asked their own opinion about whether the warming was human caused, naturally caused, or whether there was no warming at all, the average OCSI score was virtually the same in all classes:
The lesson here is the same as Kahan reached for evolution: how much you know about science doesn’t necessarily affect your opinions on science-related matters related to public policy. In the case of evolution, this was because no matter how much people know about science (and about evolution), they tended to reject evolution if they were religious. For global warming, though, although religion plays a role (Christians tend to deny climate-change more often than do nonbelievers), politics is probably more important. Kahan therefore divided up his results by the political self-identification (based on answering another set of questions) of the respondents.
If you divide up the probability of getting answers correct for a given question by political affiliation, you see that it doesn’t make much of a difference: for some questions liberal Democrats do a bit better, for others conservative Republicans do better (!). Here are the same four plots as above, but divided by political grouping: red for conservatives, blue for Democrats (note that the bottom two plots have been switched compared to the four shown above):
And in those bottom two plots, the Republicans have a higher probability of getting the two questions right for a given OCSI score than do Democrats. (As with the evolution questions, the scores are lumped into 21 groups, and the size of the bars are the 95% confidence intervals.)
The graphs below show that there’s not much difference in the percentage of people getting answers correct based on political affiliation, though for five of the questions it looks as if Democrats did statistically better than Republicans (bars are two standard errors of the mean):
Finally, there is one plot giving peoples’ personal opinions on climate change divided up by political affiliation, and this one shows the real divide:
What we see, of course, is that the proportion of conservative Republicans giving the correct answer (that Earth is warming mostly because of human activity) is much lower than for liberal Democrats at any level of OCSI score. That’s regardless of the fact that both groups know pretty much the same thing that scientists accept. So their beliefs, then, do not rest on knowing the scientific consensus, but on one’s political affiliation.
Not only that, but the higher the Republicans score on climate-science knowledge, the less likely they are to believe in anthropogenic climate change! (Democrats showed the oppposite trend.) I’m not sure exactly why that is, and Kahan’s discussion is unclear, but it may be that the “smartest” Republicans are those most likely to seek conformity to their group identify. The group-identity explanation is how Kahan interprets why Republicans and Democrats show such a disparity in personal beliefs despite roughly equal knowledge of what science says (my emphasis):
This pattern, moreover, characterized the responses of subjects of both left- and right-leaning po-litical outlooks. Left-leaning subjects were somewhat more likely to select the correct answer when the items correctly attributed belief in a climate-change risk proposition to climate scientists, and right-leaning ones somewhat more likely to do so when the item incorrectly attributed belief in such a proposi-tion to scientists. But both classes of subjects were substantially more likely to indicate that climate sci-entists believe in global warming will cause some specified harm regardless of whether that response was correct. Right-leaning and left-leaning respondents alike, one might infer, were responding to the OCSI items on the basis of an affective orientation that disposed them to credit responses attributing high risk to climate change.
. . . The settings in which people’s beliefs in, and assessments of in-formation relating to, global warming are characterized by “motivated System 2” reasoning are the ones in which they are applying their reason to answer the question who are you—whose side are you on? In that setting, the answer that expresses and reinforces their cultural identity is the right one given what they are best understood to be trying to achieve when conveying their group allegiances is at stake. It is thus the reasoning proficiency of their identity-protective self that is being measured by such items.
But those same individuals are also collective-knowledge acquirers. When asked not whose side are you on but what do we know from science, they apply their reason to that question, and if they are for-tunate enough to be superb reasoners, then regardless of their cultural identity they get the answer right more often than other people regardless of theirs. If one wants to measure what people have used their reason to discern about the science of climate change, the one has to be sure to ask them in a manner that does not threaten their identities. The OCSI shows that this can indeed be done.
Kahan’s last sentence gives a clue to how he, and Nyhan, think this divide can be breached: through accommodationism. For both evolution and climate-change, we need to convince people that accepting the scientific data does not threaten their religious or political identities.
Nyhan’s piece in the NYT suggests some solutions based on Kahan’s work and the recommendations in that paper, but they seem pretty lame. Breaking the association between one’s “community” and one’s beliefs about science be they affected by politics or religion, is very hard. It hasn’t worked well for religion, as the failures of BioLogos have shown. (They’ve tried to convince evangelical Christians that evolutionary biology does not violate their faith.) How do you convince a Republican to separate his/her group-bonding beliefs from their politics? Many political beliefs are held with the same tenacity as religious beliefs.
Nyhan gives one example of how it might work for climate change. A diverse group in South Florida came together, avoiding the “is this happening?” scenario and simply assuming it is happening and trying to fix it. And apparently they got something done. But in that case everyone in Florida is aware of the incursion of seawater into fresh-water areas, so there is a palpable and immediate danger that affects both Democrats and Republicans. As Nyhan points out, this “experiment” didn’t work in North Carolina, where sea-level rise really threatens the Outer Banks, and yet people remain polarized.
Finally, Nyhan makes this recommendation:
But we also need to reduce the incentives for elites to spread misinformation to their followers in the first place. Once people’s cultural and political views get tied up in their factual beliefs, it’s very difficult to undo regardless of the messaging that is used.
Well, I’m not sure what he’s talking about, since the incentives to spread misinformation are the same incentives that lead to climate-change denialism, it seems. If he has concrete suggestions, what are they?
Greg Mayer has questioned the soundness of Kahan’s research on other grounds, so he may want to weigh in below. But if we take Kahan’s study at face value, it paints a dismal picture for the value of science education. Of course that education is essential in keeping science as a going concern in our country—for producing young scientists. But as far as educating the public, well, it doesn’t seem to work if what is being taught goes against the grain of one’s political and religious beliefs. Fortunately, religion seems to be on its way out in the US, though it’s abating very slowly. But politics will be with us forever.
Kahan, D. M. 2014 (in press). Climate science commuication and the measurement problem. Adv. Pol. Psych.