Thinking About God

Does cognitive reflection lead to religious disbelief?

Posted Feb 09, 2020

"It is the heart which perceives God and not the reason." —Blaise Pascal

A prominent hypothesis in the psychology of religion predicts that analytic thinking leads to religious disbelief (Norenzayan & Gervais, 2013). Accordingly, those who have a strong tendency for reflection or those who are prompted to reflect on religious ideas will become more suspecting of these ideas. This view is consistent with the widely held opinion, epitomized in Pascal’s statement above, that religiosity is not a matter of reason.

Indeed, Pascal had claimed that his famous religious conversion was based on a “dreamlike or ecstatic” state that he experienced on the night of 23 November 1654 (Clarke, 2015). However, at least until his conversion, Pascal was both an influential mathematician and a philosopher—an analytic thinker par excellence.

Was his conversion an intuitive change of heart or was it based on his reflections on God’s existence?

Pascal’s statement is doubly insightful, as it anticipates the modern dual-process model of the mind: the intuitive, automatic and effortless Type 1 processes that we often rely on, and the analytical, controlled and effortful Type 2 processes that take over when necessary and possible (Evans & Stanovich, 2013).

The growing body of evidence for the so-called intuitive religious belief hypothesis is based exactly on this dual-process perspective.

An overview of the literature

Broadly speaking, two alternative methods are used to obtain evidence about the intuitive religious belief hypothesis.

The first method is to look for statistical associations between the tendency for reflective thinking on the one hand and religious belief on the other. Numerous studies, resulting in meta-analyses with large numbers of observations, have found a generally consistent pattern of negative association between the two variables, thereby providing correlational evidence for intuitive religiosity (Pennycook, Ross, Koehler, & Fugelsang, 2016). This association holds not only in Western, Educated, Industrialized, Rich, Democratic (WEIRD) and (we might add) predominantly Christian cultures but also non-WEIRD and non-Christian societies (Bahçekapili & Yilmaz, 2017; Stagnaro, Ross, Pennycook, & Rand, 2019). For example, despite cross-cultural variation, Gervais et al. (2018) found a negative (albeit small) association in the aggregate dataset of thirteen diverse cultures.

Secondly, various methods have been devised to manipulate analytic thinking in order to isolate its effect on religious belief. The intuitive religious belief hypothesis has first been tested in 2012 by two independent groups of scholars, with the common finding that analytic thinking tends to decrease religious belief (Gervais & Norenzayan, 2012; Shenhav, Rand, & Greene, 2012), which was later replicated in a non-WEIRD culture as well (Yilmaz, Karadöller, & Sofuoglu, 2016).

However, many of the later studies have shown the initial experimental findings to be less robust than their correlational cousins (Camerer et al., 2018; Sanchez, Sundermeier, Gray, & Calin-Jageman, 2017; Saribay, Yilmaz, & Körpe, 2020), and the experimental manipulations of analytic thinking used in the literature have come to be seen as weak and unreliable (Deppe et al., 2015; Meyer et al., 2015; Yilmaz & Saribay, 2016).

Note: Sample images of Rodin’s The Thinker (left) and a Roman copy of Myron’s Discobulus (right).
Source: Wikimedia

For example, Gervais and Norenzayan (2012) showed participants either a picture of Rodin’s The Thinker or that of Myron’s Discobulus to explicitly promote analytical thinking and found evidence for intuitive religiosity. Two large-scale studies failed to replicate this interesting but implausible finding (Camerer et al., 2018; Sanchez et al., 2017) and the authors laudably acknowledged their initial result as a false positive (Gervais & Norenzayan, 2018). To add salt to the methodological wounds, another reflection manipulation used by Gervais and Norenzayan, the difficult-to-read-font, has later been convincingly shown not to work at all (Meyer et al., 2015).

Although some may interpret this and similar recent episodes of replication failures as confirming their doubts about scientific practices, these experiences resulted in the open science movement emphasizing the transparency of methods and reproducibility of findings, which dramatically improved the experimental study of human behavior over a short period of time (Open Science Collaboration, 2015).

Nevertheless, important questions remain: Does cognitive reflection promote religious disbelief? How can we explain the reliability of correlational evidence, when experimental results have been so fickle?

Recent developments

In a recently published article in Judgment and Decision Making, we sought answers to this question (Yilmaz & Isler, 2019). We argued that, since religious belief is often acquired during one’s early years of socialization, it is plausible for religious belief to become deeply entrenched, making it very difficult to reliably manipulate. People usually know, as a matter of fact about their individual identity, whether and how strongly they believe in God.

On the other hand, since thinking style is an enduring individual personality trait likely to be evidenced even during early formative years, correlational studies may be able to pick up this enduring association with relative ease. Indeed, we replicated the significant negative correlation between the tendency for reflective thinking—as measured by the verbal version of the Cognitive Reflection Task (Thomson & Oppenheimer, 2016)—and religious belief.

In two large-sample pre-registered experiments (N = 1,602), we also sought to test the causal link between reflection and religious belief. To do so, we chose a reliable and strong cognitive process manipulation, namely, time-limits. Both studies recruited U.S. residents from Amazon Mechanical Turk. The first study compared the strength of religious beliefs elicited under 5s time-pressure, those elicited after at least 20s of reflective time-delay and those elicited without any time-limits (the control condition). While the relatively more spontaneous responses elicited under time-pressure tend to reflect intuitions by limiting opportunities for reflection, responses elicited under time-delay tend to generate more reflective answers.

We found absolutely no evidence for the intuitive religious belief hypothesis.

To our surprise, we found exploratory evidence that reflection tended to increase belief in God among non-believers (i.e., atheists and agnostics). To check that our results were not a statistical fluke, we then ran another preregistered study. This time, however, we hypothesized that reflection will increase belief in God and that the effect would be stronger among non-believers. Our revised hypothesis went against the grain. We also used a so-called “within-subjects” design, where participants were first prompted to rely on their intuition and answer the belief-in-God question within 5s, and then had the opportunity to revise their answers after being prompted to reflect for at least 20s. This design was intended to increase the power of our test—its ability to detect systematic differences—and to provide further insights into belief change.


We did find significant evidence that reflection strengthens belief in God, especially among non-believers. Further analyses indicated that the effect stemmed from the shift in the opinions of non-believers from relatively more certain disbelief (“definitely does not exist”) towards more uncertainty (“probably does not exist” or “not sure").

In other words, non-believers became more skeptical about their presumably long-held belief about the non-existence of God as they were led to think through their spontaneous answers. We conjecture that non-believers are more open to self-questioning. In the context of religious belief, such self-questioning results in the acceptance of the possibility that there is positive probability that they may, after all, be wrong about God.

Pascal’s wager

Perhaps it was the calculated apprehension of being wrong about God that led to Pascal’s conversion. His famous wager clarified what’s at stake in religious disbelief. “Let us weigh the gain and the loss in wagering that God is”, he said. “Let us estimate these two chances. If you gain, you gain all; if you lose, you lose nothing.”

In essence, Pascal here argues that atheism or agnosticism can under no circumstances result in higher benefits than belief in God’s existence. Pascal’s Wager provides a plausible explanation for our finding that reflection increases religious self-questioning among non-believers. Needless to say, our novel finding begs further study, and we hope to experimentally test Pascal’s Wager in the near future. 


Bahçekapili, H. G., & Yilmaz, O. (2017). The relation between different types of religiosity and analytic cognitive style. Personality and Individual Differences, 117, 267-272. doi:10.1016/j.paid.2017.06.013

Camerer, C. F., Dreber, A., Holzmeister, F., Ho, T. H., Huber, J., Johannesson, M., . . . Wu, H. (2018). Evaluating the replicability of social science experiments in Nature and Science between 2010 and 2015. Nat Hum Behav, 2(9), 637-644. doi:10.1038/s41562-018-0399-z

Clarke, D. (2015). Blaise Pascal. In N. Z. Edward (Ed.), The Stanford Encyclopedia of Philosophy: Metaphysics Research Lab, Stanford University.

Open Science Collaboration (2015). Estimating the reproducibility of psychological science. Science, 349(6251), aac4716.

Deppe, K. D., Gonzalez, F. J., Neiman, J. L., Jacobs, C., Pahlke, J., Smith, K. B., & Hibbing, J. R. (2015). Reflective liberals and intuitive conservatives: A look at the Cognitive Reflection Test and ideology. Judgment and Decision Making, 10.

Evans, J. S., & Stanovich, K. E. (2013). Dual-Process Theories of Higher Cognition: Advancing the Debate. Perspect Psychol Sci, 8(3), 223-241. doi:10.1177/1745691612460685

Gervais, W. M., & Norenzayan, A. (2012). Analytic thinking promotes religious disbelief. Science, 336(6080), 493-496. doi:10.1126/science.1215647

Gervais, W. M., & Norenzayan, A. (2018). Analytic atheism revisited. Nat Hum Behav, 2(9), 609. doi:10.1038/s41562-018-0426-0

Gervais, W. M., van Elk, M., Xygalatas, D., McKay, R. T., Aveyard, M., Buchtel, E. E., . . . Riekki, T. (2018). Analytic atheism: A cross-culturally weak and fickle phenomenon? Judgment and Decision Making, 13, 268-274.

Meyer, A., Frederick, S., Burnham, T. C., Guevara Pinto, J. D., Boyer, T. W., Ball, L. J., . . . Schuldt, J. P. (2015). Disfluent fonts don't help people solve math problems. J Exp Psychol Gen, 144(2), e16-30. doi:10.1037/xge0000049

Norenzayan, A., & Gervais, W. M. (2013). The origins of religious disbelief. Trends Cogn Sci, 17(1), 20-25. doi:10.1016/j.tics.2012.11.006

Pennycook, G., Ross, R. M., Koehler, D. J., & Fugelsang, J. A. (2016). Atheists and agnostics are more reflective than religious believers: Four empirical studies and a meta-analysis. PLoS One, 11, e0153039.

Sanchez, C., Sundermeier, B., Gray, K., & Calin-Jageman, R. J. (2017). Direct replication of Gervais & Norenzayan (2012): No evidence that analytic thinking decreases religious belief. PLoS One, 12(2), e0172636. doi:10.1371/journal.pone.0172636

Saribay, S. A., Yilmaz, O., & Körpe, G. Does intuitive mindset influence belief in God? A registered replication of Shenhav, Rand, and Nowak (2012). Judgment and Decision Making, conditionally accepted as Stage 1 registered report.

Shenhav, A., Rand, D. G., & Greene, J. D. (2012). Divine intuition: cognitive style influences belief in God. J Exp Psychol Gen, 141(3), 423-428. doi:10.1037/a0025391

Stagnaro, M. N., Ross, R. M., Pennycook, G., & Rand, D. G. (2019). Cross-cultural support for a link between analytic thinking and disbelief in God: Evidence from India and the United Kingdom. Judgment and Decision Making, 14(2), 179-186.

Thomson, K. S., & Oppenheimer, D. M. (2016). Investigating an alternate form of the cognitive reflection test. Judgment and Decision Making, 11, 99-113.

Yilmaz, O., & Isler, O. (2019). Reflection increases belief in God through self-questioning among non-believers. Judgment and Decision Making, 14(6), 649-657.

Yilmaz, O., Karadöller, D., & Sofuoglu, G. (2016). Analytic thinking, religion, and prejudice: An experimental test of the dual-process model of mind. The International Journal for the Psychology of Religion, 26(4), 360-369.

Yilmaz, O., & Saribay, S. A. (2016). An attempt to clarify the link between cognitive style and political ideology: A non-western replication and extension. Judgment and Decision Making, 11, 287-300