Religious Kids Are More Selfish: The Stickers Don’t Lie

Early 1960s calendar print. Courtesy flickr user wackystuff via Creative Commons

Are kids from more religious families more or less altruistic than their peers from less-religious families?

That’s what a high-profile new study from University of Chicago neuroscientist Jean Decety and a global crew of collaborators sought to determine. In the course of the study, published in the journal of Current Biology, the researchers use something called the “children’s dictator game,” a.k.a. stickerpalooza. Here’s how it worked:

Step one. Go to an elementary school. Find a child. Place a set of 30 stickers in front of the child. Tell the child to pick her favorite ten.

Step two. Introduce a plot twist. Tell the kid that not everyone in school could participate in the sticker bonanza. Fortunately, there is a chance to share: the kid can pick between zero and 10 of her favorite, cream-of-the-crop stickers, and set them aside in an envelope. That envelope will go to another person in the school. Afterward, the kid will walk out with whatever stickers she chose to keep.

In order to keep things nice and relaxed, this sharing stage is anonymous. Nobody watches the kid set her stickers aside. She doesn’t know which classmate receives them, and the classmate doesn’t find out who donated them. But, later, the researcher can count the shared stickers, and have some approximation of the child’s moral fiber.

In their study, Decety and his colleagues gave the sticker test to 1,170 kids at schools in six cities—Amman, Cape Town, Chicago, Guangzhou, Istanbul, and Toronto. “Altruism was calculated as the number of stickers shared out of 10,” they write. The researchers also gave the kids another test, in which they watched videos of people hurting other people, and then judged (a) how mean the bullies were, and (b) how much punishment the bullies deserved.

Then Decety and his collaborators went to the kids’ parents and asked them questions about the religious identity and practices of the family, and about how moral they thought their kids were.

Here’s the zinger: according to Decety and his colleagues, kids from more religious households are less altruistic, and more apt to deal out punishment, than kids from non-religious households. Corollary zinger: on the punishment front, Muslim kids are even more vindictive than Christians.

“Nonreligious children are more generous,” explained a headline at Science magazine. “It’s not like you have to be highly religious to be a good person,” Decety told Forbes. “Secularity—like having your own laws and rules based on rational thinking, reason rather than holy books—is better for everybody.” Forbes headlined the article “Religion Makes Children More Selfish, Say Scientists.” (Decety tweeted a link to the piece). In the Forbes interview, Decety cautioned that there would be naysayers, at least among the anti-science crowd. “My guess is they’re just going to deny what I did—they don’t want science, they don’t believe in evolution, they don’t want Darwin to be taught in schools.”

The Cubit is all for science—and Darwin! In fact, that’s why we feel obligated to point out that Decety’s paper is deaf to interdisciplinary critiques, premised on an obsolete and misleading view of the world, fundamentally unable to acknowledge its own hubristic assumptions, and, consequently, unlikely to produce meaningful insights into reality and the human condition.

Making generalizations

Sometimes, late at night, drinking in pseudo-seedy college-town hipster dive bars, I’ll start complaining about bad social science, and midway through the rant, friends will adopt kindly, exasperated who-gives-a-shit? expressions.

It’s a fair response. But this stuff matters.

Western societies have developed specific tools with which to make generalizations about human beings. Some of the more traditional tools here include stereotyping and prejudice. Often, these get ideologized into more potent generalization-generators: racism, anti-Semitism, Islamophobia, sexism, and so on.

In many cases, social science—and especially experimental social science—tries to produce generalizations about human beings, too. Sometimes, it seeks to do so about a specific subset of the population (say, Christians or Muslims). I want to be 150% clear here: when researchers seek to find general principles or patterns in human society, that does not make them racist, or anti-Semitic, or Islamophobic, or sexist. It is very possible to seek rules for social interaction without falling into ideological or political traps.

At the same time, social scientists run a special kind of risk, because their work does seem, superficially, to share certain goals with the ideological generalizers. Researchers have to be extra-careful to ensure that the biases and preconceptions of the political space don’t bleed into their work. And they have to be sober and realistic about the conclusions they draw from their research. Additionally, it might not hurt to make sure that your methods are something other than totally shitty.

For a case study of how this little balancing act might break down, let’s check out the Decety et al. paper. The study is trying to understand certain rules that might govern the relationship between religiosity and prosocial behavior, which is a technical term for stuff you do that benefits other people in your community. Prosocial behaviors range widely, from the dramatic (sacrificing your life for a team) to the mundane (sharing stickers).

I’d like to peel apart the methodological issues here with a cool and well-informed eye, except that the methods section of the paper is unusually vague, and Jean Decety isn’t responding to my polite emails. Nor my polite calls. Nor the polite request that I lodged with him through his lab manager, who confirmed that Decety knows that I’m trying to reach him.

Frustrated, I asked the editor of Current Biology if he could help me puzzle through some of the gaps in the methods section. He told me to email Decety.

Soldiering on through the silence, we can see a few key problem areas with “The Negative Association between Religiousness and Children’s Altruism across the World.” Basically, these problems have to do with the definition of religiousness, the metric for altruism, and the concept of world. Let’s take them one by one.

Problem 1: “The Negative Association between Religiousness….”

 Scholars of religion are fond of explaining all the reasons that religion is really, really difficult to define. The more cynical among us might point out that, in a crowded job market, academics distinguish themselves by explaining why everyone else’s categories are wrong, meaning that religion scholars have a strong incentive to expound, at length, about all the reasons that religion is really, really difficult to define.

But, look: they have a point. In most of the world, for most of history, cultures had very little notion of a discrete thing called “religion,” as something that you could then choose, reject, or petition for freedom of. The choice to constellate certain kinds of rituals, stories, propositions and epistemological modes into a single package called “religion” is a fairly recent, European, and Protestant phenomenon. Similarly, the idea that religiousness is separable from the rest of culture, in such a way that you can see it quantifiably motivating certain behaviors, would seem alien and weird to a lot of people, for whom Faith is not so easily distinguished from other strands of culture.

In the case of a cross-cultural religiosity study, such as Decety’s paper, this definitional challenge makes the work pretty tough from the start. Religiosity is entangled with all sorts of other cultural markers and experiences. The researchers control for one confounding variable here: socioeconomic status. Number of siblings? The researchers didn’t address that, even though religious engagement can be correlated with family size (thanks to a Reddit commenter for pointing this out). Amount of exposure to other kids outside of school? Ditto. Parents’ education levels? Missing. Heterogeneity of the surrounding community? Also left out. And so on.

More slippery, though, is that this kind of works requires some way of measuring religiousness, such that you can get a scale that works the same in mostly-Muslim Amman as it does in mostly-Christian Chicago, and also in Guangzhou, where concepts of ritual and religiousness are wildly different than those held in the Western world. In other words, you have to take this fragile Western construct (religion), put a numerical scale on it, and then apply that scale around the world, as if religion were a solid, universally measurable thing, like the force of gravity or the density of water.

Judging by the paper, Decety and his colleagues haven’t even recognized this problem, let alone addressed it. Instead, they just grabbed a handy global-religion-quantifier and went to work. The metric they chose, the Duke University Religiosity Index, or DUREL, is designed for epidemiological studies that take place within a single religious tradition, and for use with Abrahamic religions. In other words, it’s not at all adapted for cross-cultural research that includes East Asia.

Problem 2: “…and children’s altruism…” 

It’s unclear how well something like the sticker test measures the delicate, context-dependent applications of altruism within an individual child’s life. “One potential critique [of the study] is the artificiality of the situation” said Luke Galen, a professor of psychology at Grand Valley State University who reviewed a draft of the paper. But, he added, “that’s true of 90% of the studies in the literature.” In other words, things like the sticker test are comparable to other discipline-approved tools that we have to plumb the dynamics of human kindness.

In all fairness, social science is hard. Developing a metric for altruism is tricky work, and you have to start somewhere. In this sense, the sticker test probably isn’t a bad tool. The question is how far you’re willing to generalize about the qualities of billions of people based on its results. Judging by Decety’s comments about the nature of secularity and morality, the answer for some researchers is very far indeed.

As Galen points out, the fact that the researchers did any kind of rigorous altruism test, instead of just asking kids and parents to self-report about their moral feelings and behaviors, sets the Decety study apart from many others in the field. Wisely, Galen adds that the study should be read in the context of a larger set of recent studies finding that religion might not be the magic prosociality-booster-pill that some other generalizers have claimed it to be.

These are important caveats to my snark.

Problem 3: “…across the World.”

There are a few reasons, though, to wonder whether these results can be generalized very far. Does a population of kids from a handful of major urban centers really tell us much about the world as a whole? Additionally, the researchers don’t explain how they recruited the kids. Omitting recruitment data doesn’t invalidate the findings, but it makes it hard to gauge the generalizability of the results, Galen said. Were the kids selected randomly? Did families have to volunteer? It’s not clear.

One other, even bigger problem: most of the non-religious kids probably come from China.

Not that you would know that from reading the paper. In a strange omission, Decety and his colleagues do not provide any kind of country-by-country breakdown of where the Muslim, Christian, and non-religious kids come from. As a result, when you see a graph like this, which is the linchpin of the paper—


—you can almost imagine that it refers to a robust global sample of non-religious kids, stacked up against their faithier peers, and not a bunch of kids from Guangzhou, getting compared to children from five other cities, all of which lie West of Mecca.

How do I know all of this? Educated guesswork. I could be wrong. But here’s how the numbers break down: globally, 323 families in the study identified as non-religious. And 219 kids in the study came from China. It is extremely unlikely that more than a handful of the Chinese families identified themselves as Christian or Muslim, and we know for sure that they mostly avoided identifying as Buddhist, because just 18 families in the whole 1,170 kid dataset did so. Nobody identified as Confucian. Just six families, worldwide, said they were “other.”

By elimination, that leaves around 200 Chinese kids for the non-religious side of the ledger, or around 60% of the total non-religious pool.

So, how do we know that this study is picking up on something unique to religiosity, instead of the difference between Chinese kids and non-Chinese kids, or Western kids and non-Western kids, or Guangzhou kids and everybody else? Well, you can analyze the stats enough to pick out religiousness as one factor, distinct from country-of-origin, that seems to be driving some of the result. That’s important. But it’s not clear from the numbers provided in the study that non-religious kids, globally, share fewer stickers in a way that’s separable from other ethnic markers.

A divergence

In the past few decades, there has been a sharp divergence between those who study religion from within sociology and the humanities, and those who approach it from the side of social and evolutionary psychology. The humanists and sociologists have moved toward more and more granular snapshots of religious life, leaving behind the old, sweeping Religion is x, y, and z formulations that defined the good old days, when a dude in an office at Oxford could comfortably sketch out a theory of ritual based on secondhand ethnographies from remote tropical islands. Meanwhile, the social and evolutionary psychologists seem to be flying full-tilt in the direction of more and more grand theories of The Role of Religion in All Humanity.

From my semi-neutral post as a journalist who covers both fields, I’d like to suggest that the social and evolutionary psychologists are more full-of-shit than the humanists. The fact that someone like Decety feels comfortable taking his sticker games and making public comments about the fundamental nature of morality and secularity feels slightly surreal. (It’s not just in Forbes interviews. Here’s the final line of the paper: “More generally, [our findings] call into question whether religion is vital for moral development, supporting the idea that the secularization of moral discourse will not reduce human kindness—in fact, it will do just the opposite.”)

The problem is not that Decety and his colleagues’ results aren’t interesting, or even that they’re wrong—for all I know, all the world over, kids who engage more with certain ritual experiences are less kind to their peers.

The problem is that, absent robust evidence for his generalizations about the Nature of all Christians and Muslims, it is difficult to tell where Decety’s grand claims emerge from actual evidence, and where they may owe a debt to politicized beliefs about how religion in general, or specific religious traditions (i.e. Islam), motivate people to do bad things.

Also on The Cubit: Lying About Our Religion, and Other Problems with Polling