The Origins of Religious Disbelief: A Dual Inheritance Approach, now in press at SPPS.

Me, @Nava_Caluori , @MNajlePhD . Preprint at https://psyarxiv.com/e29rt 

Tweetorial about it

Atheism makes a nice test case for different theoretical approaches to religion. We wanted to see which types of things predict atheism. Different theories say different things about which predictors oughta carry the load.

Following exemplary work by @aiyanakoka in Czech/Slovakia, we focused on three main theories:

secularization
cognitive byproduct
dual inheritance https://www.researchgate.net/publication/312207939_Testing_Theories_of_Secularization_and_Religious_Belief_in_the_Czech_Republic_and_Slovakia

Secularization theories from sociology and social psych variously say that religions thrive where life is rough and atheism might emerge under conditions of existential security. Primary predictor = various social conditions

Ev Psych and cognitive science of religion favor a cognitive byproduct account. Basically, the idea is that religions are good fits for our adapted minds.

What predicts atheism? Lower mentalizing, but *especially* cognitive reflection. They straight up say this is final boss

Finally, an emerging dual inheritance approach that says cognitive byproducts are all fine and well, but they’re basically content biases. But cultural things like religions need more than content biases. things like credibility enhancing displays are important (see @JoHenrich )

3 different approaches. Different hypotheses about what ought to be the best predictors of atheism.

Years ago now, we paid GfK a boatload of money to get a nationally representative sample in the USA with measures of these key constructs, plus all the usual suspect covariates.

The main analysis predicted religious disbelief with theory-favored measures (mentalizing, security, CREDs, reflection), and we saw which ones popped.

We have a pretty clear winner: CREDs.

Cognitive reflection is there, but a lot weaker. Mentalizing too. Security meh.

To illustrate the degree to which CREDs trump cognitive reflection, we imagined a hypothetical person who effectively maxed out each variable while being otherwise typical. Odds of atheism were about 90% higher for someone at floor on CREDs than for someone ceiling on reflection.

Interestingly, cognitive reflection and CREDs interact. Basically, cognitive reflection mainly/only predicts disbelief among folks low on CREDs. Or to flip that, sufficient exposure to CREDs inoculates people against the putatively corrosive influence of reflection on faith.

So, the data were (to me at least) way better a fit for the dual inheritance approach than for secularization or byproduct alone. And IMO it wasn't especially close.

But, at always, caveats apply. For instance, our measures were shitty in unequal ways across theories.

Overall, this fits really nicely with Aiyana’s work in Czech/Slovakia, which found really similar patterns within and between countries. Such synergy.

CREDs, y'all. Kind of a big deal

The project itself was a bit of a ride. I designed it in 2013, in my first big grant application. 7 year lag is a damn long time to complete a project, but the end result is better for the wait. I’ve gotten a lot better at this science schtick since the initial proposal.

Enjoy!

As always, the paper is better off thanks to insightful reviews by 3 anonymous sets of eyes. Also big thanks to guest editor Bill von Hippel whose steady hand guided us to a more nuanced paper.