A Tale of Two Studies: Media Filtering of Science Narratives


In April, 2012, a Science article suddenly caught an extraordinary amount of media buzz: Will M. Gervais and Ara Norenzayan’s “Analytic Thinking Promotes Religious Disbelief” (Science 336, 493, 2012).  The L.A. TimesScientific American, Huffington Post, ABC News, Atlantic, Panda’s Thumb, and other prominent venues immediately drew attention to the study (of which an admirably balanced account can be found here).  It looked only at questionnaire responses given under the short-term influence of certain stimuli (e.g., viewing an image of either Renan’s analysis-suggesting The Thinker or Myron’s merely athletic Discus Thrower) . . . but screw the details, it was time to go ape on tendentious paraphrases.  “Even the Religious Lose Faith When They Think Critically” was the Atlantic’s take, and things didn’t go up from there.

But when the same two researchers published an article just one month later finding that autism tends to impede belief in a “personal God”—“Mentalizing Deficits Constrain Belief in a Personal God”—it got almost no buzz.  So far as I can tell, not one of the venues named above covered the second study.  Total blackout.  The obvious interpretation: it did not appear to pump up anybody’s culture-war narrative about Reason vs. Religion, and so was not News.

It was the best of news, it was . . . not news

There were, of course, some commentators—including some atheists—who refused to join the superficial hoot-fest over “Analytic Thinking Promotes Religious Disblief.”  One fact too little noted even by the skeptics, I think, is that the study had nothing to do with conscious analysis of religious beliefs: it looked strictly at mood-shifts below the threshold of awareness.  Such shifts, as earlier studies had already shown, can be triggered by cues as subtle as making people read an ugly, difficult font rather than an easy font.  People forced to turn on their “analytic system” in order to read crappy-looking text tend to perform better on analytic tasks shortly afterward.  Gervais and Norenzayan used ugly font as one of their stimulus sets and found that it tended to suppress religiosity in follow-up questionnaires, as expected.  Conclusion: not that religion is irrational on its face, but that we are profoundly suggestible creatures.  Which we already knew.

Also, perhaps, the experiments add support to the dual-process theory of human cognition, which Gervais and Norenzayan employ.  The idea is that we all have two cognitive systems or processes, intuitive and analytic, and activities that turn up the power on analysis tend to suck it from intuition.  So, to the extent that religion runs on “intuition,” its plausibility is weakened when cognitive resources are diverted to the analytic system, even if the religious stuff isn’t subjected to explicit analysis.  Which, in this case, it wasn’t.

What the Science study certainly does not show, or even tend to show, is that when people begin to reason about their religious beliefs they instantly see them for the idiotic confabulations that they are.  If anything, it tends to show that movement away from religious belief can be driven by subconscious shifts having no more explicit logic content than counterposed mood-swings toward belief.  That’s old news to sociologists and psychologists of religion, but contradicts any simplistic “reason sees through religion” reading of the study.

It is best to put aside all hope that either religion or atheism can be meaningfully promoted or demoted by experimental methods.   Gervais and Norenzayan’s data are about certain very narrowly-defined aspects of the cognitive psychology of certain humans selected from a culturally narrow sample population: they reveal nothing about the nature of the God, the Universe, and Everything.  We cannot decide such questions by such means.  Ditto for their autism study.

As an Episcopalian Christian with a doctorate in Engineering Sciences, my own background is a vanilla-fudge swirl of religious ideation and hard-core analytic training. I’m no super-shakes at either, but my experience of this twofold path, as far as it goes, is that considered merely as moods, analysis and religious awe (and kindred ecstatic states) are indeed often in tension, even at war.   There really does seem to be a tendency for the analytic mood to dilute or diminish the religious or ecstatic mood.

On the other hand, I do not find that all religious beliefs simply crumble away when subjected to the bright, clean light of logic.  Of course they don’t.  As a pre-teen, my first serious encounter with emphatically analytic, logical writing was in the religious essays of C. S. Lewis: I ate it up.  Since then, I can say of the tension between the analytic and the ecstatic, as Scrooge’s nephew said of Christmas, “I believe that it has done me good, and will do me good; and I say, God bless it!”  It is what provokes all good theological writing, which attempts to bring the two ends together in a coherent, fruitful way.

It would not surprise me if a future study found that religious believers who engage consciously and analytically with their faith in an ongoing way are less likely to give it up, in the long run, than those who do not.  I also suspect—the hypothesis follows trivially from the dual-process theory of cognition that Gervais and Norenzayan favor—that experiments similar to those described in the Science paper would show that subconscious, pro-analytic mood shifts, or outright exercise of our analytic abilities, tends to suppress (at least briefly) not only our religiosity but our ability to engage with beauty, morality, joy, empathy, and all other “intuitive” aspects of the self—aspects that even evangelical atheists would not be so anxious, I am sure, to mock or ditch.

Of my experience of the darker side of this basic tension, I will say more next week.