Some data on education, religiosity, ideology, and science comprehension
Tuesday, October 15, 2013 at 1:22AM
Dan Kahan

No, this blog post is not a federally funded study. It's neither "federally funded" nor a "study"! Doesn't it bug you that our hard-earned tax dollars pay the salary of a federal bureaucrat too lazy to figure out simple facts like this?

Because the "asymmetry thesis" just won't leave me alone, I decided it would be sort of interesting to see what the relationship was between a "science comprehension" scale I've been developing and political outlooks.

The "science comprehension" measure is a composite of 11 items from the National Science Foundation's "Science Indicators" battery, the standard measure of "science literacy" used in public opinion studies (including comparative ones), plus 10 items from an extended version of the Cognitive Reflection Test, which is normally considered the best measure of the disposition to engage in conscious, effortful information processing ("System 2") as opposed to intuitive, heuristic processing ("System 1").  

The items scale well together (α= 0.81) and can be understood to measure a disposition that combines substantive science knowledge with a disposition to use critical reasoning skills of the sort necessary to make valid inferences from observation. We used a version of a scale like this--one combining the NSF science literacy battery with numeracy--in our study of how science comprehension magnifies cultural polarization over climate change and nuclear power.

Although the scale is designed to (and does) measure a science-comprehension aptitude that doesn't reduce simply to level of education, one would expect it to correlate reasonably strongly with education and it does (r = 0.36, p < .01). The practical significance of the impact education makes to science comprehension so measured can be grasped pretty readily, I think, when the performance of those who have and who haven't graduated from college is graphically displayed in a pair of overlaid histograms:

The respondents, btw, consisted of a large, nationally representative sample of U.S. adults recruited to participate in a study of vaccine risk perceptions that was administered this summer (the data from that are coming soon!).

Both science literacy and CRT have been shown to correlate negatively with religiosity. And there is, in turns out, a modest negative correlation (r = -0.26, p < 0.01) between the composite science comprehension measure and a religiosity scale formed by aggregating church attendance, frequency of prayer, and self-reported "importance of God" in the respondents' lives.

I frankly don't think that that's a very big deal. There are plenty of highly religious folks who have a high science comprehension score, and plenty of secular ones who don't.  When it comes to conflict over decision-relevant science, it is likely to be more instructive to consider how religiosity and science comprehension interact, something I've explored previously.

Now, what about politics?

Proponents of the "asymmetry thesis" tend to emphasize the existence of a negative correlation between conservative political outlooks and various self-report measures of cognitive style--ones that feature items such as  "thinking is not my idea of fun" & "the notion of thinking abstractly is appealing to me." 

These sorts of self-report measures predict vulnerability to one or another reasoning bias less powerfully than CRT and numeracy, and my sense is that they are falling out of favor in cognitive psychology. 

In my paper, Ideology, Motivated Reasoning, and Cognitive Reflection, I found that the Cogntive Reflection Test did not meaningfully correlate with left-right political outlooks.

In this dataset, I found that there is a small correlation (r = -0.05, p = 0.03) between the science comprehension measure and a left-right political outlook measure, Conservrepub, which aggregates liberal-conservative ideology and party self-identification. The sign of the correlation indicates that science comprehension decreases as political outlooks move in the rightward direction--i.e., the more "liberal" and "Democrat," the more science comprehending.

Do you think this helps explain conflicts over climate change or other forms of decision-relevant science? I don't.

But if you do, then maybe you'll find this interesting.  The dataset happened to have an item in it that asked respondents if they considered themselves "part of the Tea Party movement." Nineteen percent said yes.

It turns out that there is about as strong a correlation between scores on the science comprehension scale and identifying with the Tea Party as there is between scores on the science comprehension scale and Conservrepub.  

Except that it has the opposite sign: that is, identifying with the Tea Party correlates positively (r = 0.05, p = 0.05) with scores on the science comprehension measure:

Again, the relationship is trivially small, and can't possibly be contributing in any way to the ferocious conflicts over decision-relevant science that we are experiencing.

I've got to confess, though, I found this result surprising. As I pushed the button to run the analysis on my computer, I fully expected I'd be shown a modest negative correlation between identifying with the Tea Party and science comprehension.

But then again, I don't know a single person who identifies with the Tea Party.  All my impressions come from watching cable tv -- & I don't watch Fox News very often -- and reading the "paper" (New York Times daily, plus a variety of politics-focused internet sites like Huffington Post & Politico).  

I'm a little embarrassed, but mainly I'm just glad that I no longer hold this particular mistaken view.

Of course, I still subscribe to my various political and moral assessments--all very negative-- of what I understand the "Tea Party movement" to stand for. I just no longer assume that the people who happen to hold those values are less likely than people who share my political outlooks to have acquired the sorts of knowledge and dispositions that a decent science comprehension scale measures.

I'll now be much less surprised, too, if it turns out that someone I meet at, say, the Museum of Science in Boston, or the Chabot Space and Science Museum in Oakland, or the Museum of Science and Industry in Chicago is part of the 20% (geez-- I must know some of them) who would answer "yes" when asked if he or she identifies with the Tea Party.  If the person is there, then it will almost certainly be the case that that he or she & I will agree on how cool the stuff is at the museum, even if we don't agree about many other matters of consequence.

Next time I collect data, too, I won't be surprised at all if the correlations between science comprehension and political ideology or identification with the Tea Party movement disappear or flip their signs.  These effects are trivially small, & if I sample 2000+ people it's pretty likely any discrepancy I see will be "statistically significant"--which has precious little to do with "practically significant."

Update on Saturday, October 19, 2013 at 7:10PM by Registered CommenterDan Kahan

Yes, the tea party is just like everyone else -- which is to say, highly vulnerable to the reason-effacing consequences of our polluted science communication environment.  

More here.

Article originally appeared on cultural cognition project (
See website for complete article licensing information.