Here I will analyze a recent paper by Kathleen Corriveau et al., published in Cognitive Science that has been heralded as evidence that “Children Exposed To Religion Have Difficulty Distinguishing Fact From Fiction.” First, I will look at what the researchers say the study says, then look at what the media claims the study finds, and examine the two experiments conducted in this study to see what the study actually finds. In short, this study does not provide evidence that kids from religious households have more difficulty distinguishing fiction from fact. More research is needed to rule out alternative interpretations of what the authors are claiming.
Kathleen Corriveau has done extensive research on how children learn from testimony. She finds that children are not passive recipients of testimony, but that they modulate their trust. For example, she finds that preschoolers place less trust in people with a different accent from the child’s social milieu (a sobering finding), but overall, accuracy trumps accent in deciding whom to trust. One of her co-authors, Paul Harris, has recently written a very accessible and informative book Trusting What you’re told about the profound influence of testimony on children’s worldviews (e.g., their receptivity to scientific unobservables). The present study aims to examine the role of religious testimony in a child’s receptivity to stories with unusual occurrences.
In order to appreciate what the authors aim to find out, it’s important to get clear on what the received view is in cognitive science of religion. The received view is that people are “born believers”, and that religious belief is natural. On this view, children are naturally receptive to extraordinary claims of religious traditions with special elements such as people who turn water into wine, or who can calm a storm by commanding the sea to be still. Corriveau and colleagues challenge this claim by this set of 2 experiments, arguing that children from religious households are more open to extraordinary claims (such as one can find in religious traditions), whereas children of secular households would be more skeptical.
Now, the popular media version glossed this study as follows: children exposed to religion would have “difficulties distinguishing fiction from reality”, for instance “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.” [note, this study did not test whether children believe talking animals could exist, so this claim is rather misleading]. If this were true, it would certainly confirm various preconceptions about religious believers as more gullible. It would also raise discussion about the possible harms of a religious upbringing. However, as it stands, the paper does not support the stronger claim the popular media make, and I think it does not adequately support the more nuanced claim of the researchers either.
Let’s look more closely at the 2 experiments that are discussed:
Experiment 1 finds children from religious households know their Bible
Experiment 1 presented five-and six-year-olds from secular and religious families with stories drawn from the Old and New Testament. They were asked whether the protagonist in the stories was real or fictional. An example is the following prompt
This is Joseph. Joseph was sent to a mean king in a land far away. However, God sent Joseph many dreams warning about terrible storms, and Joseph used those dreams to tell the king how to protect his kingdom from the storms. The king was so amazed by Joseph and they became friends.
They had three types of scenarios featuring Joseph (and other characters from the Bible): fantastical (Joseph used his magical powers to see into the future), realistic (Joseph used natural means to foretell the storms) and religious (prompt above). Unsurprisingly, children from religious households were more likely than secular children to say that Joseph was a real person in the religious scenario, and slightly more likely to say so in the fantastic scenario than secular children.
The bar chart above shows the difference between children from secular household (striped bars) and from religious groups (other three bars) – the bars indicate the percentage of children judging Joseph was a real character. Note that the endorsement of Joseph as real in the fantastic scenarios is also lower in the religious kids, but less so than in the secular kids.
Given the familiarity of children with these Bible stories, this result is unsurprising. It does not show children from religious households are less able to distinguish fact from fiction, rather it shows that they know their Bible. The Bible characters are presented to them as historical, so of course they would be more likely to judge them as historical than children who didn’t hear about these characters. Indeed, even the fantastical stories are strongly reminiscent of the original Bible stories. Would young children be prone to judge George Washington is a fictional character, even if supernatural properties were attributed to him?
Experiment 2 finds children from religious households are more likely to judge that a character in stories that are closely analogous to Bible stories are real than children from secular households
To control for these familiarity effects, experiment 2 changed the names of the characters to unfamiliar names, e.g., “John” instead of “Moses”. Also, in some versions, the stories were slightly altered; instead of the sea being parted, one condition included a mountain that is being parted so that people can escape their pursuers:
This is John. John led his people when they were escaping from their enemies. When they reached the mountain John waved his stick. The mountain separated into two parts, and John and his people escaped through the pathway in the middle.
This experiment also found that children from religious backgrounds were more likely to judge that John was real than children from secular households. There were almost no religious justifications (appeal to God) when children from religious households judged John to be real. Now, the authors take the fact that children from religious households are more likely to judge both the parting mountain and parting sea scenarios as true “undermines the hypothesis that religious children associate fantastical stories with familiar Bible stories. That is, if familiarity with particular biblical stories were critical for religious children’s judgments of fantastical stories, then the difference between secular and religious children ought to have been greater for stories that included familiar events from the Bible compared to those that contained only unfamiliar, albeit equally impossible events.”
Given that the stories in study 2 were still closely analogous to Bible stories, I am not convinced that these events would strike children as unfamiliar. In order to establish that kids from religious households are more receptive to claims about impossible events, we would need to have widely divergent scenarios. At the onset of experiment 2, the authors report a study where they asked children to categorize 3 familiar characters (historical) and 3 characters that appear in stories (fictional stories). For each correct categorization, children received 1 point. Interestingly, there was no difference between religious and secular children. Secular children had a mean score of 5.75, and religious children had a mean score of 5.88. Unfortunately, there are no further details about this study, but it does suggest that children from religious households are just as able as children from secular households to distinguish familiar fictional characters (e.g., Cinderella) from historical characters (e.g., George Washington).
To see whether the slightly altered Bible scenarios support the authors’s claim that religious children are more receptive to impossible events, consider the following analogy: suppose we were to test preschoolers who were exposed to lots of scientific ideas from an early age onward. Harris found earlier that children are very receptive to the testimony to invisible scientific entities. Now, suppose we compared such children with children from such households to normal children with prompts like “Phlogiston is a kind of stuff that is inside things like wood, and it evaporates when you burn the wood” – Is Phlogiston a real thing or is it pretend? Suppose furthermore that kids from science-intensive households were more likely to judge phlogiston is real. Would it support the claim “Children exposed to science early are less able to distinguish science from pseudoscience?” Perhaps it would support something along the more modest claim Corriveau et al are making, namely that children from such households are more receptive to claims about invisible scientific entities. But perhaps it just shows that children are reminded of closely analogous scientific entities (like oxygen) and make their claims based on that. Note that earlier work by Harris and colleagues found that even preschoolers from religious backgrounds are more likely to think uncontested scientific unobservables (germs) exist than widely endorsed imaginary beings (Santa Claus).
In closing, I do not think the claims that Corriveau et al. make are prima facie implausible. Perhaps a theistic worldview does make one more receptive to claims about unusual events that defy ordinary causal occurrences. For instance, Keith Ward writes that if theism is true “[i]t will be possible for miracles to occur and to be so reliably recorded that expectations are upset. The existence of such reliable records would, indeed constitute a worthwhile argument for the possible action of a god, even though it may be far from decisive on its own. Whether reliable enough records actually exist is a different question; the point is that they certainly could exist” (Ward 1985, 138).
However, stories that diverge more sharply from the Bible would need to be used in order to establish this. It would be worthwhile to do a follow-up study recounting Hindu, Buddhist and Sikh miracle narratives, and see whether children from Christian households would be more likely to think those miracles occurred than children from secular households.