Education and religious commitment

The BUGB news sweep picks up another forthcoming piece of social scientific research (the report is from the Daily Mail, but as far as I can tell nonetheless fairly accurate), this one using data from the American General Social Survey. The headline is that there is a pronounced positive correlation between years in education and likelihood of attending religious services – more colloquially, the more educated you are, the more likely you are to go to church (or synagogue, or mosque – but given it is American data, church is the real point) – more precisely, each additional year of education makes an American 15% more likely to attend worship.

This is valuable in that the narrative of ‘only ill-educated people are religious’ is still out there in the culture, and hard data to rebut it is useful. That said, it is not a surprising result, nor should we be too quick to claim that education makes people more likely to believe; there is no indication in the report as to whether the statistics have been normalised for other factors; if not, it may well be that religious observance and educational longevity are both related to another variable. (Amongst British students, to take a slightly parallel example, there is simply no question that evangelical Christianity is on average much stronger in elite universities than in others; I am fairly sure that the explanation for this is less ‘cleverer kids are more likely to be evangelicals’ than ‘middle class kids are more likely both to be evangelicals, and to be accepted by elite universities’.)

The report goes on to make claims about the nature of the faith of educated Americans: educated people are more likely to read the Bible (again – is this normalised? ‘educated people are more likely to read’ is not news; do they turn to the Bible disproportionately?); less likely to agree that ‘the Bible is the literal word of God’; and less likely to agree that ‘only one religion is the true religion’.

These latter results strike me as complex; the researcher, Philip Schwadel (U of Nebraska-Lincoln), narrated them in terms of more educated people being more ‘open-minded’ but not less ‘faithful’; I suspect that this is wrong. As any of us who have been involved in survey design and interpretation know, at the point where you are asking people to agree with simple statements about complex issues, you get into difficult areas of interpretation. For instance, a recent survey showed that British Evangelical church leaders were less likely to agree with the statement ‘abortion is always wrong’ than their church members; in discussing this, the people who did the survey made the point that this was probably an unfortunate artefact of the question, in particular the inclusion of the word ‘always’: anyone who has a measure of theological education these days will have been offered ‘abortion to save the life of the mother’ as a test-case in ethics (I’m not saying it’s a good one, but everyone uses it, including me…); with this background, the word ‘always’ in the question sounds like it is a test on where you stand on this, somewhat obscure, debate on the limits of ethics, not a test of your basic ethical position concerning abortion.

Similarly, ‘the Bible is the literal word of God’ is a very complex statement: someone with a measure of theological knowledge will immediately recall debates over ‘is’ vs ‘contains’ in this context; the use of the word ‘literal’ is at least obscure, and probably actually misleading (if one holds to Calvin’s doctrine of accommodation, say, can one say that the Bible is the ‘literal word of God’? The answer is not immediately clear; Calvin’s doctrine of Scripture was hardly cautious). I reflect on what happens in my own mind when confronted with such a survey question: I first note that the question is complex enough not to admit of the yes/no answer required; then reflect on which response more adequately reflects my view; then reflect on what the question might be used as a proxy for (if I say yes, will someone assume that I am closed-minded/a fundamentalist/a six-day creationist/…?); then I probably tick the ‘don’t know’ box… It seems likely that there is a simple correlation between educational achievement and an unwillingness to indicate agreement with a broad-brush statement on a complex issue.

In passing, this illustrates the difference between two types of data: church attendance and educational achievement are ‘hard’ data: people may lie (in the aftermath of the Thatcher government, political pollsters found they had to correct for ‘shy Tories,’ people who voted Conservative but would not admit it, even in an opinion poll!), but if you can get good data, you can be confident of what it means. ‘Do you agree with the following statement…’ is (almost) always softer data, needing interpretation and unpicking.