Study Finds That Atheists Can Be More Closed-Minded Than Religious People (Sort Of)


Image via Wikimedia Commons

If you’ve followed my blog or Facebook at all, you have probably seen me rant about the poor reasoning from my own atheist community. Sometimes it seems like atheists are entrenched in their own ideology (i.e. anti-feminism) that made them appear just as dogmatic as theists. Well a new study provides some experimental data that atheists can be more dogmatic than theists. Let’s go through it.

First off, the study was published in a reputable, peer-reviewed journal, Personality and Individual Differences. That’s a good sign.

As we jump into the paper, the authors start to talk about how theists have typically found to be more dogmatic than atheists in the literature. However, recent research may provide evidence that atheists can be dogmatic in their own ways:

Just as religious beliefs increase when believers are confronted with adversity (Sedikides & Gebauer, 2014, for review), the belief in science has been found to increase when non-believers are experimentally confronted with adversity (Farias, Newheiser, Kahane, & de Toledo, 2013)

I’ve read a few of these papers and I think it is reasonable to bring up. People like meaning. People especially like adhering to meaning and belief systems when stressed. So far so good. The authors then argue that there may be different types of dogma that need to be parsed out:

Thus, dogmatism, defined as unjustified certainty (Altemeyer, 2002), should be more present among the religious, especially if this refers to cognitions relative to the existential and moral domains.

So while theists have higher “unjustified certainty” atheists may be less likely to entertain opposing viewpoints:

We argue that, in highly secularized religious contexts, non-believers, compared to their religionist peers, would be less prone to be interested in, consider, understand, and appreciate perspectives that oppose their own.

They provide some support for this (but it seems somewhat weak in my opinion).

Indirect evidence in favor of these expectations comes from a recent analyses of large international data by Gebauer et al. (2014). These authors found that the somewhat negative association between religiosity and openness to experience decreases, disappears, and may even be slightly reversed, when one shifts from religious to secular countries. They interpret this finding as reflecting the fact that, in the latter societies, religious believers “swim against the stream”, whereas non-believers “swim along the stream”. Additional indirect evidence comes from recent studies showing that those very low in religious fundamentalism, or very high in antireligious sentiments, have their own prejudices with regard to specific targets, that is religious people and moral conservatives (Brandt & Van Tongeren, 2017; Kossowska, Czernatowicz-Kukuczka, & Sekerdej, 2017).

In sum, we expected non-believers, compared to religious believers, to show, at least in a secular cultural context, a lower intolerance of contradiction as well as less readiness for perspective taking for positions alternative to their own. This should be the case even if religious believers score higher on dogmatism, defined as an unjustified certainty in one’s own beliefs.

So that’s the background. Let’s dive into the methods:

The participants were 788 adults (18–71 years old, M = 32.61, SD = 10.83; 41% women) recruited mostly through the crowdsourcing platforms Crowdflower and Prolific Academic, as well as through social networks. They were residents or nationals of the United Kingdom (n = 242), Spain (214), or France (332). The participants reported being atheist (302), Christian (255), agnostic (143), Muslim (17), Buddhist (17), Jewish (3), or “other” (51).

Not a bad sample size. Cool that it was international. They had three tools to measure dogma. A myside bias task, intolerance of contradiction task, and a dogmatism scale.

To measure myside bias as a low propensity to take a different perspective into consideration, we used a version of an arguments-generation task developed by Toplak and Stanovich (2003) and adapted by Van Pachterbeke, Keller, and Saroglou (2012). Participants were first asked to rate their agreement with three different opinion statements: (1) “Child adoption by homosexual couples is a positive advance for society”; (2) “The meaning of life is something entirely personal”; and (3) “In a house, rooms must be painted with light colors” (8-point Likert scales). The order of presentation was counterbalanced. Afterwards, in a separate screen page, participants were asked to generate as many arguments as they could both in support for and in opposition to the statements reflecting the above opinions.

I don’t particularly like this measure (and the authors admit at the end that it could be limiting). There is scientific evidence that gay marriage is not harmful. So to argue against it is to argue against science. So being supportive of gay marriage is not an unjustified belief at all. But the other statements provide some room for “both sides.” Next measure, intolerance of contradiction:

This construct was measured using three pairs of short statements, each pair presenting two seemingly contradictory scientific findings (statements selected from Peng & Nisbett, 1999). Participants were asked to rate, on a 9-point Likert scale, the extent to which they thought each of the six findings was true. The rationale behind this measure is that people who are intolerant of contradiction will have more difficulty in accepting the seemingly contradictory findings as equally true (or false). Thus, if they evaluate one scientific finding of the pair as true, they will tend to judge the other as very false.

This measure seems better, but can still have some issues. Sometimes science is relatively decided on issues and sometimes it is not. I can see having a difficult time viewing both sides as equal in certain topics. For example, one of the statements (found here) was about climate change, something that is clearly settled. So it would be hard to argue in a justified manner for the other side.  But I guess you could see how well people can entertain both sides in some of the other statements that are less clean cut.

The last construct is general dogmatism:

We administered six items from Altemeyer’s (2002) Dogmatism scale (6-point Likert scale, α = 0.75). This scale measures unjustified certainty, specifically in one’s own beliefs. A sample items is: “There are so many things we have not discovered yet, nobody should be absolutely certain his beliefs are right” (reverse). People scoring low on this measure show openness to the possibility of reassessing and changing their opinions.

That one is pretty straightforward and what we may think of as unjustified belief. So what did they find?!


The results were consistent with their hypotheses! (They also assessed attitudes towards religion, which I didn’t mention yet as it wasn’t the focus).

As far as attitudes toward religion were concerned, the expected hierarchies were observed: in religiosity, Christians were higher than agnostics (p b 0.001), who were higher than atheists (p b 0.001). The opposite was observed with external critique: both atheists and agnostics were higher than Christians on symbolic unbelief (ps b 0.001), but did not significantly differ from each other.

Regarding the measures of closed-mindedness, significant differences between the convictional groups were observed, though not all in the same direction.5 Christians were significantly higher than atheists in self-reported dogmatism (p = 0.017). However, atheists were higher than Christians on the intolerance of contradiction, myside bias in arguments, and myside bias in conviction (ps = 0.037, 0.071, 0.013). Similarly, agnostics were higher than Christians on myside bias in arguments (p = 0.004), but did not significantly differ from Christians on intolerance of contradiction or myside bias in arguments, and from atheists on the myside bias indexes

Okay so yes, from this study it does seem like atheists had higher levels of intolerance to contradiction and myside bias. But theists had higher levels of dogma.

Importantly, the authors talk about several key limitations (that undoubtedly will be ignored by most media reporting on this study).

  1. Small effect size. In the table about we see statistical significance (the likelihood the finding did not occur purely by chance). But effect size (the magnitude of the effect) was small. So this means that we should be careful if there really was much of an effect here.
  2. Dogmatism scale was unrelated to the other two measures. So it seems like they are tapping into different constructs and shouldn’t all be labeled “dogma.” Close-mindedness seems more accurate for an umbrella term.
  3.  Again, those measures that found atheists are close-minded have problems. Being “dogmatic” for something that has clear evidence (climate change) is different than being “dogmatic” about something that does not have clear evidence (existence of god).

So in conclusion, this paper is interesting and provides a great avenue for future research. I definitely think atheists can be biased and close-minded about topics. Atheists are humans too. But this study, as the authors admit, has limitations. I don’t think it’s fair to equate the close-mindedness of theists and atheists, but both groups definitely can be dogmatic in their own ways. I look forward to future research on this topic!

PS: I now have a Patreon if you’d like to support my writing and podcasting. 

Keep up with my writing, activism, and research by following me on Facebook!