This may get a little long, so here’s the bottom line:
While Kahan has very handily problematized the notion that liberals and conservatives differ in what is called “motivated reasoning” across the board, there is nevertheless too much suggestive evidence of various kinds of left-right asymmetries—most recently, here—to dismiss the idea outright (although it certainly may have to be refined). Moreover, broader personality-based and moral differences between left and right are well documented, contrary to Kahan’s critique. While we don’t always know how these differences play out in the real world, they are certainly affecting how left and right respond to political information. Kahan’s is therefore a very helpful critique of the book, but I’d advise taking your seats and ordering lots of popcorn—a lot more science is going to be coming in on this topic and I’m pretty confident that the argument isn’t over yet!
In my last post, I previewed a debate I’ve been having with Yale’s Dan Kahan. Basically, here’s the gist: Kahan doesn’t believe that liberals and conservatives differ in their tendency to engage in what is called “motivated reasoning.” Motivated reasoning, in case you’re new to the term, is letting your gut emotions shape your thinking or arguing in an identity protective fashion—to protect who you are, what group you belong to, your religious beliefs, your political views, etc. This is an extremely important phenomenon, incidentally, because whenever we see large scale politicized denials of reality—for instance, the conservative rejection of the science of global warming—it is likely that motivated reasoning is involved.
There are a variety of controlled psychological studies on motivated reasoning, and in some of them (e.g., here), there is at least a hint of conservatives doing it “worse.” However, Kahan points out methodological problems in these studies—for instance, if scientifically literate conservatives respond differently to the science of global warming (more dismissively) than scientifically literate liberals do to the science of nuclear power (less dismissively), that is not necessarily proof of any inherent difference between the groups. Because after all, the issues aren’t precisely the same, nor are the underlying emotions necessarily the same.
Kahan then goes on to show, in a new experiment, that when you set up a motivated reasoning test correctly (in his view) and with adequate controls, you do not get a big difference in left-right responses. Fair enough--but the problem for Kahan is that shortly after he did that experiment, a new study came out (see here; my reporting here) that once again appeared to capture a significant left-right difference in motivated reasoning. In this case, the study examined how much liberals and conservatives align their views of what is factually true with their morals or values on four separate issues—the death penalty, enhanced interrogation (e.g., torture), whether we should teach kids about condom effectiveness, and embryonic stem cell research. And in every case, conservatives aligned the facts with their values more than liberals did.
So who's right, and what’s going on here?
Let’s back up. First of all, the central argument of The Republican Brain is not that conservatives engage in more motivated reasoning than liberals, what Kahan calls the “asymmetry thesis.” I discuss “motivated reasoning” a great deal in the book, and acknowledge that both sides are doing it. I certainly let on that I suspect that conservatives are doing it “worse” in some way, but also admit that there is much uncertainty here—and in fact, an experiment run for the book itself failed to show worse motivated reasoning on the right across the board (though we did detect a hint on a few issues).
So why, then, am I not right alongside Kahan, simply saying that we’re all biased, just in different directions?
To put it simply, I think there is too much evidence that people who opt for the left, and people who opt for the right, are just different along any number of personality and moral dimensions that may skew how they respond to political information. The book is centrally devoted to reviewing and rehearsing this evidence; and its possible implications do not just include a potential difference in motivated reasoning. They are far broader than that.
For instance, perhaps an even stronger case exists right now that conservatives may engage in more “selective exposure” to information sources—consuming information that they already agree with, rather than sampling widely. (See, for instance, here.) That’s not exactly the same thing as motivated reasoning, but it has been captured in several studies and also fits theoretical expectations: Conservatives are known to be higher on the “need for closure,” and we know that if you have a high need for closure you tend to limit your information search once you’ve found what you need to feel secure in your beliefs.
If I have a real bone to pick with Kahan, it’s his apparent dismissal of much of the evidence—a large body, spanning decades—on left-right psychological differences. I want to quote what I think is his most problematic statement:
The studies Mooney assembles are not all of a piece but the ones that play the largest role in the book and in the literature correlate ideology or party affiliation with one or another measure of cognitive processing and conclude that conservativism is associated with “lower” quality reasoning or closed-mindedness.
These measures, though, are of questionable validity. Many are based on self-reporting; "need for cognition," for example, literally just asks people whether the "notion of thinking abstractly is appealing to" them," etc. Others use various personality-style constructs like “authoritarian” personality that researchers believe are associated with dogmatism. Evidence that these sorts of scales actually measure what they say is spare.
What is striking about this is that Kahan himself bases his research paradigm—“cultural cognition”—on self-reports (for the scale, see this paper) of whether people are “egalitarian” or “hierarchical” in outlook, and whether they are “individualist” or “communitarian.” So surely he knows that just because you are using self-report measurements doesn’t make your work invalid. What matters is how much various scales have been validated in experiments, and if you talk to the researchers who pioneered the study of need for closure, need for cognition, authoritarianism, and other measures, I think they will tell you that there is ample experimental evidence attesting that they are indeed measuring what they say they are measuring. Ditto for personality: The “Big Five” Personality scale, which is used to show major personality differences between the average liberal and conservative, is very well accepted.
In place of self-report measures, Kahan extols a test called the “Cognitive Reflection Test” or CRT, which he says has been better validated. Well, perhaps it has, but it is not necessarily the right test for what we are looking at. Measures like “need for closure” do not necessarily detect an inclination toward poor reasoning as the CRT does, but rather, deep-seated motivations in how people process information. The CRT is different: It detects how likely you are to follow your gut and give the objectively wrong answer on a series of tricky math problems. In other words, it detects a more reflective and less intuitive cognitive style—and in a recent study of libertarians, it turned out they were better than both liberals and conservatives on the test (though liberals were also somewhat better than conservatives).
Fine—but this is really not what we’re talking about. Or at least, not what I’m talking about. The CRT is not about your emotional or moral motivations or your commitments, but rather, your ability to not fall for a gut response and instead think through the problem to get the right answer. However, nobody has a strong personal stake in the outcome of these math questions. That’s very different from the recent study that showed conservatives aligning their factual beliefs with their values more than liberals did—here, these are clear moral issues (like the death penalty) where beliefs are very strong and emotions run high.
So, I am still not convinced that liberals and conservatives are just equally biased, but in different directions. For the record, here is how I am now thinking about things—an argument that was not entirely explicit in The Republican Brain, because my thinking has advanced further since then.
We know conservatives score higher on the need for closure. We also know, from Jonathan Haidt’s research, that they are more loyal to their group, tribe, or team. And these are not unrelated: Belonging to a group provides epistemic closure—a much desired sense of certainty, because you adopt that group’s views and membership in it provides you with a sense of stability and belonging.
So is it very surprising, then, that the views of the group get strongly defended, even if they may be factually wrong?
And it is just as important that liberals are the opposite—less loyal, less likely to affirm central dogmas of the group.
I think this may explain why, in some motivated reasoning tests where there isn’t much emotional at stake, you tend to find equal left-right biases. But in others, where there is something moral, emotional, or group oriented involved—or something political—you sometimes find conservatives more wedded to their beliefs.
But I fully acknowledge that there is much research remaining to be done here, and I expect further scientific clarity in the future. In the meantime, I want to thank Kahan for his helpful critique. I'm glad that we both agree that there is a lot of scientific uncertainty here, which can only be dispelled by further research—and until such time, we'll both have to advance our views tentatively and critically, in the best spirit of science.
Finally: I'll be on a panel with Kahan at the CSICON conference in Nashville on Oct 27. Like I said...bring popcorn.