Religious children are less altruistic… or maybe not…

Image courtesy of Wikimedia Commons

A recent study published in Current Biology claims to have demonstrated that children raised in religious households are less altruistic and more vindictive than their peers raised in non-religious households. Using two different tests — a Dictator Game conducted with stickers and a task that measured reactions to watching interpersonal harm — the researchers showed that children raised in religious backgrounds shared fewer stickers in the Dictator Game and judged interpersonal harm as “more mean” than their non-religious peers. The authors of the paper conclude:

Not surprisingly, this paper got quite a bit of media attention, as it seems to contradict the running narrative that religion is a source of morality and altruism. But does it?

I would be the first person to admit that scientists are subjective. What we try to achieve in our science is objectivity, but both the way we plan our science (what question? what hypotheses?) and the way that we interpret our science (what does it mean?) will be subjective. The goal is for “the science” — the investigation that produces results — to be objective. How objective was the design and execution of this study?

Although the study claims to use a “large, diverse, and cross-cultural sample”, the breadth of this sample may be more of a liability than an asset. As others have pointed out, the fact that this study relies so heavily on different regions to provide different religious backgrounds is highly problematic. Basically the religious diversity covaries with regional diversity, which means that a whole host of other confounding factors accompanied variation in religion. It is probably impossible to eliminate all confounding factors in such a study, because religious affiliation is going to covary with some other social factor, but a less cross-cultural study would have better established that religion was a major driver of these results.

Beyond the methodological problems with this study, I am also pretty shocked by the methods section that it provides. I know that there is a premium these days on brevity, but many of the methods are so brief that it seems highly unlikely that anyone could faithfully replicate this study. It is hard to judge whether the study was designed properly to answer the questions at hand, because the methods section just sort of vaguely alludes to other studies without clarifying how those previous methods were applied in this study. There’s a bit of a “trust us” implication to the way methods are presented in this paper.

Given the way these authors interpret these results, it is a bit hard to trust them. As indicated above, I expect no one to be completely objective in their interpretation of results, but the interpretation of these results seems so biased. The results of the sticker Dictator Game are pretty straight-forward; if you believe that willingness to give away stickers provides a good approximation of overall altruism, it does appear that non-religious children from their particular countries (mostly China?… the authors don’t share the breakdown of participants by region) are more altruistic than their religious counterparts. But the effect size here is small, so small that one could imagine that a whole host of other factors besides the religiosity of these children could explain their difference in altruism.

But what about the second interpretation of religious children being more harsh in their willing to punish? The authors project their own negative interpretation of the increased perceived meanness displayed by children from religious households. But is being more judgmental of behavior that causes personal harm really negative? Of course the answer depends on who you ask, and what culture they come from. But punishment has been shown to both increase (Rockenbach & Milinski 2006) and decrease (Rand & Nowak 2011) cooperation depending on the nature of social interactions, so a harsh judgment of behavior could be associated with altruism. These authors give none of this context, instead providing a more polemic frame: religious kids are less altruistic, more punitive.

There’s a lot to be learned about what the real role of religion is in our complex, cooperative societies. It is fascinating to wonder whether secular values might enable altruism and cooperation as well — or even better — than religious values. But let’s show our real curiosity about the issue by conducting valid and nonjudgmental studies of how religious and non-religious people differ in their social behaviors.

This post is based on Current Biology “The Negative Association between Religiousness and Children’s Altruism across the World” (Decety et al. 2015)