You may have seen in the news recently some reports about a new study with an intriguing finding. Here’s the Discovery Channel, for example, telling us that education may not dilute religious beliefs. Hmm. I think they’ve got their headlines wrong, but it’s not entirely the media’s fault in this case. I think that, in this case, some of the conclusions that the researcher has drawn aren’t really justified by the data.
Philip Schwadel, a sociologist at the University of Nebraska, has done a careful and rigorous analysis of data from the 1998 General Social Survey in the USA. He looked at the link between education and a variety of religious beliefs.
In many ways, he found exactly what you would expect: that educated people are less religious. For example, they are less likely to say that any one religion has ‘the Truth’, they’re less likely to think that the Bible is the actual word of a god. They’re less likely to think that you should follow religious teachings, and more likely to oppose praying and Bible reading in school.
On the other hand, educated people are more likely to read the Bible and to pray frequently. They’re more likely to go to Church and do voluntary religious work, and less likely to say that religious leaders should stay out of politics. And although they switch among religious affiliations, they are no more likely to switch out of a religious affiliation altogether. They still say they belong.
What this tells is is that educated Americans are more liberal and open-minded in their religious beliefs, but more dutiful and community oriented. That’s pretty much what you would expect.
But what about the nub of the matter – actual belief in God? Well, here’s where it gets a little bit more complicated – and where I start to disagree with Schwadel’s conclusions.
To understand why, you need to understand the difference between the size of an effect (so-called ‘effect-size’) and it its statistical significance. The bottom line is that if you look at very large samples of people, you can pick out even tiny effects (in this case, the amount of change in religion associated with an increase in education). Even though the effect is tiny, it could still be statistically significant.
On the other hand, if you have a small sample of people, then you might miss even quite large effects. It’s not that they’re not there, it’s just that there’s not enough data points in your sample to smooth out the background noise and let them shine through.
Take, for example, the question of whether people believe in God. According to Schwadel, educated people still believe, it’s just that they say they believe in a ‘higher power’ rather than in a personal god. That’s because his data shows that this link is statistically significant. But when you look at the effect size – i.e. the correlation between education and belief, well then the link is just as strong between education and ‘Don’t believe’.
In other words, we can’t really be sure that educated people aren’t less likely to believe. It may simply be that the sample of non-believers was not big enough for us to tell.
Now, this doesn’t mean that education is linked to disbelief, but it does suggest that any link is complicated. That’s not too surprising because there’s an awful lot of variability in what you get when you sign up for ‘education’, after all.
Some colleges are overtly religious, of course. But more important than the nature of the college is the nature of the subject and the nature of the experiences. Other studies have shown that the studying humanities is particularly likely to lead to loss of belief.
It ain’t that education leads to loss of belief. It’s more opening your mind to other perspectives that does it!
Philip Schwadel (2011). The Effects of Education on Americans’ Religious Practices, Beliefs, and Affiliations Rev Relig Res : 10.1007/s13644-011-0007-4