Study: Religious kids believe the Bible and other ‘fiction’

Is this Clutching at Straws Month? Because I don’t know how to dress or what to buy for it. I do know how to celebrate, though. Just publish a study that counters traditional beliefs. And don’t ask questions that might uncover flaws.

The latest example emerged this week in the July issue of Cognitive Science. Three researchers alleged that young children who are “exposed to religion” — gotta love that wording — have trouble telling fact from fiction.

This claim is in an appallingly brief, 291-word article in the Huffington Post — which, true to form, swallows and regurgitates the stuff without chewing. We’ll get to that in a bit.

First, here’s how it went down:

Researchers presented 5- and 6-year-old children from both public and parochial schools with three different types of stories — religious, fantastical and realistic — in an effort to gauge how well they could identify narratives with impossible elements as fictional.

The study found that, of the 66 participants, children who went to church or were enrolled in a parochial school were significantly less able than secular children to identify supernatural elements, such as talking animals, as fictional.

By relating seemingly impossible religious events achieved through divine intervention (e.g., Jesus transforming water into wine) to fictional narratives, religious children would more heavily rely on religion to justify their false categorizations.

“In both studies, [children exposed to religion] were less likely to judge the characters in the fantastical stories as pretend, and in line with this equivocation, they made more appeals to reality and fewer appeals to impossibility than did secular children,” the study concluded.

Now let’s dismantle this, starting with the sampling. I don’t often resort to italics, but c’mon — sixty-six subjects? I saw several times that many kids yesterday at one Vacation Bible School. A sampling of 66 children is pretty small for an attempt to generalize to all children.

I’d also like to know how many schools, and what kinds, the researchers drew from. Yes, they included both public and parochial schools. How many? Where? And what kinds of parochial schools? Just Catholic? Any Jewish or Baptist or Episcopal or Muslim schools?

These questions might be answered in the study itself, which was unfortunately behind a paywall. But HuffPost didn’t answer the questions, and probably didn’t ask them.

The study link did have an abstract, which doesn’t answer the questions either. The abstract does reveal a key fact: that the study dealt not with real or imagined events but with characters.

“Children who went to church or were enrolled in a parochial school, or both, judged the protagonist in religious stories to be a real person, whereas secular children with no such exposure to religion judged the protagonist in religious stories to be fictional,” the abstract says. The study got similar results with “fantastical stories that included ordinarily impossible events whether brought about by magic or without reference to magic.”

Then the abstract takes its flying leap to the conclusion: “The results suggest that exposure to religious ideas has a powerful impact on children’s differentiation between reality and fiction, not just for religious stories but also for fantastical stories.”

Hold on there, scientists. For the stories or for the protagonists of the stories? You test for one, then draw a conclusion about the other? Exactly how is belief in a person related to belief in events?

After all, most of us believe that Muhammad existed, but not that he heard from Gabriel in a cave. We pretty much agree that Joseph Smith lived, but not necessarily that he dug up golden plates on direction from the Angel Moroni. And we don’t have to deny that Kim Jong Il existed to doubt that he hit 11 holes in one during a single golf game.

What do other scholars think of the conclusions in the Cognitive Science study? How about religious leaders or educators? HuffPost doesn’t say. Probably didn’t ask.

A longer story in The Raw Story offers a fuller look. It says the three researchers — Kathleen Corriveau, Eva Chen and Paul Harris — expected that religious children would judge whether characters were real or imaginative based on “fantastical elements within the narrative, such as ‘invisible sails’ or ‘a sword that protects you from danger every time.’ ”

And yet, “this prediction is likely to be wrong,” because “with appropriate testimony from adults” in religious households, children “will conceive of the protagonist in such narratives as a real person — even if the narrative includes impossible events.”

So, kids tend to believe religious stories — and the characters in them — if their parents seem to believe in them? What a shock. But the story goes further:

“Children with exposure to religion — via church attendance, parochial schooling, or both — judged [characters in religious stories] to be real,” the authors wrote. “By contrast, children with no such exposure judged them to be pretend,” just as they had the characters in fairy tales. But children with exposure to religion judged many characters in fantastical, but not explicitly religious stories, to also be real — the equivalent of being incapable of differentiating between Mark Twain’s character Tom Sawyer and an account of George Washington’s life.

Actually, no, it would be more like failing to tell between the biblical Samuel and the nonbiblical Harry Potter. Did The Raw Story ask about that? Or the questions above? Nope.

This whole matter follows another study earlier this month, purporting to show that children of same-sex couples are actually happier and better-adjusted than those from traditional families. As my GR colleague Bobby Ross Jr. said, that study was based on a mere 315 same-sex parents across Australia — parents who volunteered for it after being told what it was about. At least the New York Post included one skeptical viewpoint.

Both studies still amount to clutching at straws, largely with media approval. Anything, it seems, to chip away at traditional beliefs.

Remember that old newsroom slogan “If it bleeds, it leads”? Well, the new version may be “If it kicks religion, it’s a good decision.”

Members mourn Atlanta church; why don’t they talk?
Can a laundromat replace the traditional church?
Same as it ever was: It’s time for a new, old GetReligion
Members mourn Atlanta church; why don’t they talk?
About Jim Davis
  • MDevlin

    This study seems to be a clickbait-y way of saying children interpret stories in terms of their upbringing. Are there plans to follow these children for years and see how they develop? Based on my own experience, I think this fear-mongering is unwarranted and the kids will grow up just fine.

  • Darren Blair

    Let’s see:

    1. In real life, here are numerous instances in which the “fantastical” has actually taken place. For example, there are numerous recorded instances in which a person has had enough in he way of adrenaline or another such substance going through hem ha hey continued to function despite serious injuries; here are also instances of people getting enough in he way of adrenaline going to temporarily give them “super-human” strength. In that sense, we’d need to know he full list of he “fantastical” stories that the kids were asked to evaluate.

    2. What were he denominations and religions represented by these schools? We’re not told.

    3. I agree that a mere 66 kids is too small of a sample size for a study like his to have any merit.

  • fredx2

    Think about this for a while.

    If we can pick out the shoddiness of this research, why couldn’t the scientists involved in it? Why couldn’t their school or other organization funding this research do this? Why couldn’t “Cognitive Science” do this? Why couldn’t the academic reviewers do this?

    The answer, I am afraid is this: The social sciences are hopelessly flawed at this point. Jonathan Haidt once gave a speech at a gathering of about 1,000 social scientists. He asked them to raise their hands. How many are liberal? Almost every hand went up. How many are conservative? 3 hands went up.

    The intense bias revealed by this “research” is nothing more than a cover for a current political campaign. Studies like this are the ultimate in the abuse of science as a tool of political propaganda, It simply is anti-science.

    The Huffington Post is, of course, not a newspaper, but a deeply committed ideological gathering place. So, I don’t suppose we should expect anything approaching journalism to come from them.

    But in the past, we used to have standards of self respect, that said that if you published something, you at least tried to make sure it was accurate. No more.

  • helen

    I have sometimes wondered, in a family which regularly tells stories about Jesus as fact (and the children have never seen Jesus) whether it’s sensible to make quite such a big deal about “Santa Claus”.

    But that usually gets sorted out before 7 or 8, when, I have read sociologists say, children start differentiating between fact and fiction. Interesting that these picked children of 5 and 6 years for their little “study”!

    • IriathZhul

      My parents taught me to believe in God, and explicitly taught me not to believe in the tooth fairy (aside from sarcastic jokes about her car getting stuck in the mud every time I lost a tooth), Santa Claus, and, God forbid, the Easter Bunny.

  • Sabio Lantz

    You said, “A sampling of 66 children is vanishingly small. And the margin of error for such a study must be enormously large.”

    The article said: “The study found that, of the 66 participants, children who went to church or were enrolled in a parochial school were significantly less able than secular children to identify supernatural elements, such as talking animals, as fictional.”

    “significantly”, if reported correctly, means that the statistics done in the study account for n=66 and show that the chance of a difference between the two groups is OUTSIDE “the margin of error”.

    • Tim Worley

      Agreed, Sabio. And it’s also reflective of the difference between a cross-sectional, correlational survey (like the ones frequently used in opinion polling, which would have enormous margins of error at this sample size), and a carefully-controlled between-groups experimental design. To be fair, I don’t know how well the study was carried out, as I haven’t been able to read it yet, and it does appear on Jim’s analysis that some of the substantive interpretations of the results could be questioned. But it is QUITE possible to find significant differences (i.e., very small “margin of error”) in an experimental design with only 66 participants. Assuming balanced numbers within each condition, 66/3 conditions = 22 participants per condition. For comparing differences between groups, a within-condition sample size of 22 is often capable of uncovering genuine, statistically significant differences among the groups.

      I don’t think this necessarily takes away from the force of Jim’s concerns about interpretation, but it’s worth noting.

      • Sabio Lantz

        I am glad someone who understands statistics jumped in. Just as it irks authors here when someone who doesn’t understand religion writes badly on it, having someone review research who doesn’t understand statistics is ironic.

        I do think it takes away from the force of his post here because it reveals rhetoric and bias — I mean, look at his choice of words: “vanishing” and his comparison to his Sunday school size. Heck, he later said, “mere 315 same-sex parents across Australia” —- really, “mere” — looks like no numbers will satisfy him if the evidence points away from his bias.

        The only folks who will listen to this criticism will be believers — he is talking to those in the echo chamber. But if that is the intended audience, I guess that is fine. But I would think this blog is shooting for a wider audience.

        Yes, as you said, even 22 participants could give the study power. I remember from Statistics 101 when we learned how n=30 is more impressive (depending on design) than we’d imagine. But then statistics is counter intuitive — thus mistakes are the norm for untrained folks.

        • Tim Worley

          And to be clear, I’m a “believer” (yes, a Christian) and a quantitative social scientist. And I generally really appreciate the work of Get Religion.

          • Sabio Lantz

            No need to be “clear” — not sure it matters, though I guessed you were Christian since you went out of your way to support Jim’s argument in spite of these points.

          • dpfavand

            I agree that it is possible to get statistically significant results in an experimental comparison with 66 participants. In that way, the results might be (and probably are) completely valid. The question though is whether that comparison can then be applied to the broader population as if it is a cross-sectional, correlational poll. The study is likely to be fine – if one takes it for what it is. The problem (as it almost always is with statistics) is when the analysis is applied inappropriately. Sixty-six is, as Tim notes, a bit small of a sample for a broad survey. But it’s not a broad survey.

            As far as the 315 same-sex couples, a ‘mere’ 315 sounds (to me) like a good sample size – if the sample is chosen well. But having a sample made of volunteers who know what the survey is about would raise questions to me. If you were gay or lesbian and your kids were unhappy/not well adjusted, would you be more or less likely to volunteer to participate? I don’t know about any or all of the individuals asked, but I could see myself being less likely to volunteer. Of course our samples are almost always volunteers, so really it’s a hard nut to crack. And I don’t know how else one would go about getting the sample. But that doesn’t mean I don’t have some concerns about it, as I do with many surveys.

          • Sabio Lantz

            Right, generalizability is always an issue.
            The misunderstanding of the technical use of the statistical word “significant” by Jim Davis seems to have been the crux of the problem with his article. He needs to do some stats reading.

    • Jim Davis

      Math was never my strong suit, Sabio; thank God I’m good at writing. :-) But as I think I understand social science, the smaller the sample, the larger the chance for error. If Pew or Gallup made sweeping conclusions for a survey of only 66 people, who would pay attention?

      Actually, some media might, if the conclusions criticized some traditional belief.

      • Sabio Lantz

        Jim, you are good at writing.
        I suggest you talk to your stats buddies before criticizing research analysis which you don’t seem to understand when it comes to “significance”. You would ask the same of those writing on religion. For both you and Tim to resist this admonition is odd, given your site’s calls for others to do the same.

        • Sabio Lantz

          In other words, an “Good point. Thanks, I should correct that.” would be what you expect of others and I had hoped from you.

          • Jim Davis

            I don’t mind agreeing with good points, but can you really expect me not to defend my writing? :-) Significance would seem to vary with the size of the sampling, no? Can you really draw conclusions about all children, everywhere on the globe, by the responses of a mere 66?

          • Sabio Lantz

            I will let Tim explain “significance” and p-values and such to you. “Significance” in scientific language is a technical scientific word. Without understanding of scientific research, it is difficult to explain simply.
            You need to ask someone to help bring you up to speed on this issue, it would help you from making further “writing” errors in the future.

          • Jim Davis

            FYI, I’ve just deleted that sentence about the margin of error. A colleague convinced me that even in small samples, there can be significant
            variations. Thanks for pointing it out.

          • Sabio Lantz

            You are welcome. That was a long thread to get to that point. This process points out how hard it is for people to seek to understand something that contradicts their desired conclusions. It takes someone inside our circles to usually convince us. All of us do this.

    • AuthenticBioethics

      My guess is that statistical significance was not achieved. In my line of work, I deal with statistical significance all the time (clinical studies of pharmaceutical products and the diseases they treat). The word “significance” was used by the news report, not the study authors. The abstract conspicuously avoids the word, preferring terms like “varied sharply” and “suggest that exposure…has a powerful effect.” Abstracts that use the word generally tend to include minimal statistical evidence to support its use, such as a brief explanation of the statistical methods used and the “P” value, which has to be at most less than 0.05 – which is to say, a 95% confidence that the finding is valid. This abstract does not provide any such information, and again, does not use the word “significant” to describe the results. Given the size of the study and the absence of “significant” and any P value, I would remain skeptical until I see the whole article. In my business, a study of only 66 subjects requires a very dramatic difference between study groups, as well as clear control of demographic differences beyond the variable being studied.

      • Sabio Lantz

        Great points, AB! Thanx.

  • Fascination Wolves

    It blows my mind that 5 and 6 year olds were tested on whether something was fictional or real, considering most 5 and 6 year olds are perfectly capable of believing a story to be both at the same time. Nobody had to explain to me that the Neverending Story or E.T. were fictional when I was 5 and 6 years old. I understood, on some level, that they were just stories. And yet that did not stop me from being terrified of the monsters or of feeling deeply sad about the events depicted as I watched them. They were very real to me. I didn’t fully grasp the concept of “actors” at that age.

    The study seems to say more about how reductive social scientists can be when it comes to talking about human children. The conclusion about non-religious children seems far more important. They assumed all characters they read about were fictional, even those from religious stories. Was this, perhaps, because the religious stories that contained “fantastical elements” were placed alongside fairy tales? Perhaps religious children had a better grasp at age 5 and 6 of concepts such as the miraculous, the divine, or the allegorical in literature. Unreligious children tend to view all narratives as fictional? What does that say about the study of history, or really of anything?

    But of course it appears that most of it boils down to: religious kids will believe anything, because religious people are backwards and irrational! Derp.

    • Sabio Lantz

      Though a former Christian, I raised my children religion-free and also no stories of santa, tooth fairy or easter bunny without being joking about it and being clear it was make-believe. Anecdotally I found them to question fact vs. fiction much more quickly than kids their age.

      • Guest

        You pointed out something that I thought of, but didn’t add to the article. The researchers’ abstract suggests that religion inherently blurs the line between reality and fantasy. But it would seem that if religion is a matter of parental teaching, then telling real from unreal would also come down to parental teaching — as it did in your household.

        • Sabio Lantz

          Not sure I follow that.
          If a religious teaching, told over and over by parents, peers and teachers, speaks of snakes talking or burning bushes talking or walking on water, and all sorts of things that contradict real experience, then wouldn’t that make the child more ready to not doubt fantastic stories (fiction)?

          Isn’t that what the article is implying?

      • IriathZhul

        Missing the point. If you teach your children that this specific thing, this Biblical teaching, is fact, then that doesn’t teach them that facts are fluid. In most religions, children are taught to reject many things as falsehood. Whether or not those things are in fact false has no bearing on whether the child lacks an ability to distinguish between fact and fiction.

  • Sabio Lantz

    BTW, I must say, I don’t think there is any harm in a child not being able to tell fiction from fact. I don’t think amazing stories are a problem. It is only later in life that it can be an issue. For instance, when a person believes God will cure a child by magic prayer and medicine is not needed.

  • AuthenticBioethics

    Kids are very trusting. I have raised 5 of them. If you give them a story, they will believe it to be historical, simply because they trust you. They will also believe hollywood special effects as real. But they also learn to play “make believe” early on. And as they get older, they become better able at making the distinctions. With only the abstract to go on, it is difficult to assess the study in any detail.

  • Kristi Winters

    Jim, you’re an idiot. You need to learn about what constitutes a valid sample size for the purposes of social science – when researchers want to generalize about lots of causal relationships to a wide and diverse population – and of psychology which looks very narrowly and a specific relationship over small samples across studies.

    Especially for first time research, a small sample is completely appropriate as a preliminary test of a causal theory. After all, why spend money studying something you aren’t sure even exists. Start your remedial education here:

    In short, your unhelpful appeal to scepticism has made your readers dumber, when you could have used this as a chance to learn more about how social research is produced and pass that knowledge on instead.

    • Darren Blair

      If you’ll read through the rest of the comments, you’ll see folks with actual experience in statistics arguing both the appropriateness of the sample size and how valid the results are because of it.

      I myself am an MBA who did coursework on statistics and survey design, and so you can add me to the list.

      Rudeness like what you’ve shown isn’t appropriate.

    • Guest

      Normal people spell skepticism with a K. You’re obviously a British barbarian and not fit to live among civilized people. This is to be considered fact since I know this one British guy, and he’s a real jerk, always spelling skepticism with a c, therefore all British are mindless savages.
      Because.. science!!!

      • Sabio Lantz

        LOL — sarcastism as a lesson in logic. hmmmm

  • Guest

    Anything you read in the Huff-puff Post is going to be nonsense.
    You should know that.

  • cdmihai
    • Darren Blair


      They deliberately excluded Jewish children from this study?


      Way to invalidate your own results…

      • cdmihai

        They picked children that were exposed to the christian bible dogma, since they derived their characters from the Bible. So non-bible religions would have represented a bias towards their experimental setup. But sounds like a good future research topic.

        • Darren Blair

          Judaism and Islam both follow at least portions of what we know as the Old Testament, and the Baha’i faith is an offshoot of Islam.

          So to say that Judaism is a “non-Bible” religion is grossly inaccurate.

          • cdmihai

            It was a “technical” limitation, not a philosophical one. Since they derived their experimental stories from the christian bible, they included kids from christian religions. For other religions they would have needed to derive stories from those religion’s texts. So the research results generalize only to the bible based religions and no information is given for the other ones. Future research can of course make a religion agnostic variant of the study. That would be interesting.

  • Angel Ollerdisse

    I enjoyed this well written and insightful article. I agree that the small group of 66 children hardly points to any concrete proof of anything other than the fact that children do not have the cognitive abilities of adults. :-)