It sometimes seems like evolution and religion have had a bad relationship almost from the start. In 1874, the Princeton theologian Charles Hodge asked, “What is Darwinism?” His answer: atheism. To this day, people on both sidesof the issue continue to think that believing in God means not accepting evolution, and vice versa. This antagonism has bled into the cognitive science of religion. There’s a growing body of literature that’s saying – not always outright – that religion and science use different ways of thinking. Can people’s performance on an analytical thinking test predict if they’re creationists or if they accept evolution?
Readers of this blog – or anyone who has read almost anything in the cognitive science of religion – are familiar with the cognitive theory that’s used in these studies. It’s called dual-process theory. The theory says there are two, separate ways in which we think. They have the Dr. Seuss-like names of System 1 and System 2. System 1 is the fast, intuitive, fly-by-the-seat-of-your-pants way of thinking. There’s quite a bit of research that says religious thinking is built on System 1. Examples include kids’ essentialism about animals, people’s tendency to find a purpose for everything (which has the fun name of “promiscuous teleology”), and the tendency to detect personal agency behind anything complex. System 2, on the other hand, is slow, analytical, and quite frankly hard. It’s what philosophy and science use. And when people want to, they can use it to override intuitions from System 1.
Will Gervais, an evolutionary and cultural psychologist from the University of Kentucky, was working in dual-process theory when he submitted a small set of questions to his school’s annual psychological testing survey that hundreds of UK undergrads take each year. Gervais’ hypothesis was that the better someone did on an analytical thinking test, the more likely it was that student accepted evolution. The analytical thinking test was the Cognitive Reflection Task (CRT). It’s comprised of only three questions, but it asks them in such a way that an intuitive but wrong answer immediately springs to mind for most people. People who typically think more analytically are supposed to resist giving the intuitive answer and get more right. The questions about evolution are the exact same ones that Pew and Gallup have been using for decades. Both had answers that corresponded with evolution by completely natural means, theistic evolution, and creationism.
Finally, Gervais wanted to control for the influence of culture on beliefs about evolution, since a devout creationist could be a highly analytical thinker. So the final questions on the survey asked about participants’ belief in God, and either 1:) whether they had a religious upbringing or 2:); how often they went to religious services and questions about credibility enhancing displays(CREDs) – – cultural activities that foster the adoption of religious beliefs in kids and newcomers. Gervais thought that CREDs might give creationism more credibility, so he wanted to be able to statistically control for this.
Gervais’ first finding was quite sad: most of the undergrads got not one item of the CRT right. I’m going to choose to believe that people fell into the intuitive trap with the CRT because they just wanted to get done with what was probably a very long survey, since it included questions from the school’s entire psych department. But even with this abysmal performance, Gervais gathered significant results. The second finding looked at the data by whether the participants answered the Pew or the Gallup evolution questions, but the results were the same for each. In both cases, Gervais found that for every question on the CRT that somebody got right, the likelihood – or odds ratio – that that person accepted some type of evolution went up by about 30%.
But Gervais didn’t just want a correlation model. He wanted a predictive statistical model that controlled for demographics, religious belief, and religious history (a religious upbringing or religious attendance and CREDs). Using such a model, Gervais found about a similar increase in odds ratio of believing in evolution as what he got when he looked at just the CRT and evolution poll questions. For example, among those who answered the Pew questions, only about 60% of people who got no CRT questions right accepted either naturalistic or theistic evolution. However, for that small group that got all three right on the CRT, 75% accepted some type of evolution. The trend was similar for the Gallup questions. (However, the percentages were different – which could be a product about how the Gallup and Pew questions were worded differently. There’s a truism in polling: wording matters.)
The design for Gervais’ experiment was very basic, because he had to share space in a large survey with lots of other researchers. This is a shortcoming. He also wanted to set up a neat binary between System 1 thought/creationism and System 2 thought/evolution. But this ignores theistic evolution – which for the most part Gervais just lumped in with naturalistic evolution. How belief in theistic evolution (that is, the claim that evolution proceeds by natural selection more or less as scientists describe it, but under God’s planning or guidance) matches System 1 vs. 2 thinkers isn’t clear. This is to say, there is a lot more to the data than discussed. Moreover, a majority of students who got zero CRT questions right still believed in some type of evolution. A high aptitude for analytical thought based on the CRT cannot explain these findings.
Gervais wisely used cultural controls in his models, but with neither model for the Pew or the Gallup questions did those controls add a significant contribution to the model. In other words, Gervais’ measure of religious upbringing or CREDs were not significant predictors of evolutionary – and by extension, creationist – belief. This doesn’t mean that controlling for cultural effects on beliefs isn’t worthwhile. It does mean that we need measures for those effects that are more sensitive than anything currently out there. As for now, it could be that what we know is that the people who accept evolution are the same ones who pay better attention to the questions – and thus do better on long, probably boring surveys with tricky, analytical thinking questions.
For more, read “Override the Controversy: Analytic Thinking Predicts Endorsement of Evolution” from the journal Cognition.