Intelligence, religion, and the twisted-​​up mind-​​game that is statistics

Photo via Thinkstock.

Photo via Thinkstock.

Poten­tially sur­prising fact: there’s a pos­i­tive cor­re­la­tion between ice cream con­sump­tion and drowning deaths, as one goes up, so does the other. At first glance this sounds pretty hor­ri­fying, but, as my stats-​​teacher-​​husband loves to remind me, they have nothing to do with each other directly. Instead, their cor­re­la­tion is gov­erned by a third “con­founding vari­able,” which is the weather. As it gets hotter, more people eat ice cream. Like­wise, the heat causes more people to go swim­ming, thereby increasing the like­li­hood of a drowning death.

Since I was a kid, my dad has enjoyed reminding me that “cor­re­la­tion does not mean cau­sa­tion.” The ice cream-​​drowning story was the first to drive the notion home for me.

A new paper from  North­eastern psy­chology pro­fessor Judith Hall and her col­leagues at the Uni­ver­sity of Rochester brought the idea to mind once again. The paper, “The rela­tion between intel­li­gence and reli­giosity: A meta-​​analysis and some pro­posed expla­na­tions,” reports the team’s analysis of 63 sep­a­rate studies car­ried out over the last cen­tury, and con­firms a long-​​examined hunch that more intel­li­gent people tend to be less religious.

I’d been puz­zling over how to write about the article for a while, when I found myself eating sushi with that stats teacher I men­tioned ear­lier. This is the sort of sub­ject that is bound to get people’s knickers into twists so I wanted to be careful.  I kept returning to the same phrase: “The studies they reviewed show that intel­li­gent people tend to be less reli­gious,” I kept telling the teacher-​​husband. I couldn’t dis­en­tangle that truth from the ques­tion of cau­sa­tion. It sounded so much like they were inex­tri­cably linked. Then the husband-​​teacher asked me a couple impor­tant ques­tions: “Does being intel­li­gent cause people to be less religious?.…Does being reli­gious cause people to be less intelligent?”

The answer to those ques­tions seemed unlikely, to me, but very dif­fi­cult to prove with any amount of cer­tainty. If you wanted to prove that reli­gion causes lower intel­li­gence, or vice versa, you’d have to do some kind of double-​​blind, ran­dom­ized trial, con­fer­ring intel­li­gence on some people and looking at whether they became more or less reli­gious, or vice versa. Despite the obvious chal­lenges with this setup, I’m sure no insti­tu­tional review board would ever go for it.

This all then brought to mind a phe­nom­enal story I heard on NPR the day before on the TED Radio Hour. Mar­garet Hef­fernan, a British writer and busi­ness­woman, told the story of a researcher named Alice Stewart, who, in the 1950s, wanted to figure out why the rate of child­hood cancer was increasing, par­tic­u­larly among affluent fam­i­lies. Not knowing at all where to begin, she sent a ginor­mous ques­tion­naire to par­ents of chil­dren with and without cancer. She asked them every ques­tion she could pos­sibly think of. One of these, was a ques­tion to mothers: did they have an obstetric x-​​ray while pregnant? It turned out there was a serious sta­tis­tical cor­re­la­tion between answering “yes” to this ques­tion and having a child with cancer.

At this point, Stewart’s data was akin to what Hall and her team have got now: Responses from people of varying levels of intel­li­gence about their reli­gious beliefs and prac­tices. But lucky for Stewart, she would have an easier (if not at all easy…it took another quarter cen­tury before clinics would stop giving preg­nant women x-​​rays) go at proving a cau­sa­tion behind the cor­re­la­tion she’d observed.

With the intel­li­gence and reli­gion ques­tion, things are less straight­for­ward. “The present find­ings are cor­re­la­tional and cannot sup­port any causal rela­tion,” Hall and her col­leagues write in the article. But they do present a few hypotheses as to what might explain the sta­tis­ti­cally sig­nif­i­cant cor­re­la­tion they and their pre­de­ces­sors observe. In one sce­nario, they note that intel­li­gent people, who also tend to sway towards non­con­for­mity, may be less reli­gious because reli­giosity is a soci­etal norm that they are trying to steer away from. Another pos­si­bility, they sug­gest, is that the cog­ni­tive style of people with high IQs is typ­i­cally ana­lyt­ical. And when it comes to spir­i­tual beliefs, there’s simply no way to empir­i­cally test the ques­tions at hand.

Finally, the authors sug­gest that some­thing called “func­tional equiv­a­lence”  makes reli­gion fun­da­men­tally unnec­es­sary to people with higher IQs: Other studies have shown (through sta­tis­tics of course) that reli­gion helps people make sense of the world around them, offering a sense of order and external con­trol. It helps us feel better about our­selves and others, helps us delay grat­i­fi­ca­tion, lowers our sense of lone­li­ness, con­fers safety and secu­rity in times of distress.…Other studies have shown all of these things to be true of high IQ as well. So, the authors sug­gest that per­haps where reli­gion serves a par­tic­ular pur­pose for some people, intel­li­gence does the same for others.

I should also men­tion that the meta-​​analysis mostly looked at studies of Chris­tian reli­gions in western societies.…because that’s were most of research has been done in the field so far.  Looking at other cul­tures and reli­gions could obvi­ously open a whole bunch of doors for more expla­na­tions and hypotheses.

So, what does all of this tell us? Basi­cally, don’t jump to conclusions: