Field of Science

Religion and health - a double edged sword

Long-time readers of this blog may remember a post from 2009 on religion and health. It summarized the findings of nearly 60 studies, and concluded that while going to church seemed to be linked to better health, religious beliefs per se were unimportant.

Studies that have been published since seem to support this notion, but what nobody's been quite sure of is why church going should be linked to better health at all. Two recent studies have shed some light on the issue.

The first used data from a survey of 35,000 Norwegians, and found that regular churchgoers had lower blood pressure - after taking into account other factors that can affect blood pressure, such as age. Blood pressure is a good general indicator of your overall cardiovascular health, and so picks out those people most at risk of having a heart attack or stroke.

Now, there have been a lot of previous studies that have found similar things, but this is the first large study in a mostly non-religious country. Given that only 4% of Norwegians attend church regularly, it's fascinating that the link between church and blood pressure still holds.

The second study is even more interesting.

Darren Sherkat, a sociologist at Southern Illinois University, and Ben Moulton (a data analyst), wanted to find out how religion, health, and education interact. They used data from 22,000 Americans who participated in the National Health Interview Survey in 1987, and whose deaths were recorded over the next 20 years.

What they found (after crunching the numbers through a model to account for a host of other factors) is best understood by taking a quick squint at the graphic below. Higher bars indicate greater likelihood of dying.

You can see that, for the less well educated, the risk of dying goes down as church attendance goes up. As you would expect.

Surprisingly, however, for the educated the effect is exactly opposite! Educated people who go to church often are actually more likely to die young!

Sherkat puts this down to the double-edged effects of religious teaching on healthy behaviour.

On the one hand, it encourages abstinence from harmful drugs like tobacco and alcohol. That's great for the ill-educated who may not know any better (or be motivated to abstain). But not so much use for the educated, who have had the risks drummed into them and who are unlikely to overdo the booze and fags in the first place.

On the other hand, it tends to undermine science and evidence-based medicine. That might be particularly a problem for the educated, who might otherwise be expect to know better.

Overall, because there are more ill-educated people than educated ones, religious attendance has a beneficial effect. But the benefit really accrues to the ill-educated. For the educated, religion (at least, the kind of religion widely practised in the USA) might actually be harmful.
Sørensen, T., Danbolt, L., Lien, L., Koenig, H., & Holmen, J. (2011). The Relationship between Religious Attendance and Blood Pressure: The Hunt Study, Norway The International Journal of Psychiatry in Medicine, 42 (1), 13-28 DOI: 10.2190/PM.42.1.b

Moulton, B., & Sherkat, D. (2012). Specifying the Effects of Religious Participation and Educational Attainment on Mortality Risk for U.S. Adults Sociological Spectrum, 32 (1), 1-19 DOI: 10.1080/02732173.2012.628552

Creative Commons License This article by Tom Rees was first published on Epiphenom. It is licensed under Creative Commons.


  1. " Higher bars indicate greater likelihood of dying."

    For any given human being the risk of dying is always 100%. We all die at some point.

    I can see from the 2nd paragraph after the graph that you mean the risk of dying "young". But then what does that mean? Define "young". How much younger and against what scale?

    Why did the researchers postulate a causal link between being educated church goers and lack of faith in medicine? At what confidence level? And how does this translate into higher risk of dying "young". What is the causal mechanism? What other possible causes were considered and how were they eliminated?

    What kind of population was being measured? I find it very difficult to believe that mainstream Christianity undermines science based medicine amongst the educated. Does the paper say more about how they came to this conclusion? Does the paper speculate on the effects of undermining science based medicine in the uneducated (who are presumably more susceptible - or is this just a conceit?).

    Also if this is science where are the error bars on the measurements. What is the margin of error when measuring 300 million people with a sample size of 22k?

    Also looking at the abstract the contrast is with "those with the highest levels of educational attainment." Which means what? PhD? Or just a graduate qualification? You ignore this distinction! What proportion of people at the highest levels of educational attainment attend church anyway? And why?

    The effect on the graph appears to be quite large (no explanation of the units or scale so this could be deceptive). So what you appear to be saying is that the effects of abstinence from alcohol and tobacco, the two common substances which have the greatest impact on health (in the UK at least), are outweighed by a reduction in faith in medicine? That's some wild effect that can neutralise the benefits of not smoking!

    Finally since this result to some extent conflicts with your earlier result, what are related studies saying? Presumably correlations between educational level and smoking have been studied, for example. What do those studies say?

    The result is potentially interesting, but I'm far from convinced that the information you're presented supports the conclusions, or is consistent with the conclusions in the paper (a little fudging has gone on, but without reading the paper we won't know how much).

    This is a fragmentary, and not entirely coherent, report of an isolated and decontextualised paper which raises more questions than it answers. Seriously if you're going to use scientific papers in your discussions you need to do a lot better job of presenting it.

    The trouble is that you'll have a curve like that one in the post: the more educated will look at is and say "I doubt it"; while the less well educated don't have the background to assess your report, or the paper itself (or even have access to the paper to check if you've reported it well, which I frankly doubt); and the uneducated aren't reading the blog at all.

    You seem to be on safer ground with your more philosophically oriented speculations.

  2. So if I start dumbing-down by smoking lots of dope and then attend church regularly, I may live forever? Yeah!!

    On a serious note, I must say, I agree with Jayarava when he says, "I find it very difficult to believe that mainstream Christianity undermines science based medicine amongst the educated."

    Well, not enough to affect their health. In fact you reported a study where believers are more likely to hang on to Modern Medical intervention in their last days, if I am correct.

  3. We need a control - the effect of regular attendance at events other than church. Regular gym attendance is probably not a good control, but maybe regular attendance at a non-religious group (craft club? reading group? coffee shop? volunteer work?).

  4. Jayarava's attitude is intresting. In study there is clear link. Perhaps the "story is invented" and linked with booze. But the correlation between knowledge is still valid.

    Definition level is alo intresting. If there is something like age or education, you normally have smooth and gray-lined effect. And you just generally siplify your saying with words like "young" or "educated". People normally know what is more and less - and that is enough, usually.

    So, when the definition game and "playing with words" come along, I can not seriously think that is anything else than a "red herring".

    "those with the highest levels of educational attainment." Which means what? PhD? Or just a graduate qualification?" I can not do anything else than laugh. (Sorry that I say it loud.)

    In other hand in religious side those stories are taked extremelly seriously when there is correlation with health and religion. That proves that religion is superior way of living - some may explain health with happiness or even "healing by miracle".

    And if you citique those pro-religion stories inaccuracies, and start yelping with proof and margins of error, it is not rare at all to get answer which say something like "in humanist approach it is impossible to do study which give you what you demand - so you have impossible exeptations and you are driven by your worldview" (or something like that)

    So I can link the word-gaming attitudes and fixation to the stories beside correlation with certain atitude towards religion. So critique says perhaps a bit too much about the complainer, not the topic.

    But in other hand it is good to be accurate. So in this part it is "good for you".

  5. Jayarava, the risk of dying applies of course to the 20-year follow-up period of the study. Mean age at entry was 40.

    Sherkat cites his previous research showing that religious fundamentalists have lower scientific literacy. He comments that "This pattern is suggestive of a religiously-inspired polysemy of educational resources (Darnell and Sherkat 1997; Sewell 1992; Sherkat 1998), whereby the cognitive resources provided by education are reinterpreted in religious communities, and seen as oppositional and less valuable guides to health-related behavior."

    There's a fair amount of evidence that religious people do not take up preventative medical advice. I summarised some recent research in my round up of 2011: "...people who have faith in god are less likely to take their prescribed medicine, are less likely to get vaccinated, and more likely to have unprotected sex. Perhaps this is why neighbourhoods in the USA with a lot of Pentecostals also have the highest infant mortality."

    The effect was strongest in those with graduate education, which is what's shown in the graph that I selected. This means MSc, Phd etc. They get a similar effect splitting by <4 or >=4 years college, although the interaction was only significant for graduate education (the paper is 14 pages, so there's a lot I didn't cover).

    The scale on the graph is relative risk. So graduates who go to church more than weekly are 16% more likely to die than the average graduate, those who never go to church are 20% less likely to die than the average.

    Other studies have simply controlled for education, when looking for the effects of religion. This is the first study (to my knowledge) that has looked at an interaction. The two studies I wrote about here don't contradict. It's perfectly possible for religion to be associated with overall benefit, but a negative effect in a subset.

  6. Tuomo, yes, thanks for that comment. When I write these posts its a constant challenge to get the level of detail right. Some people want lots of detail, at least if the subject interests them. But I need to keep the posts fairly light to avoid making them too tedious. This blog is supposed to have a pop science feel!

  7. I read in a study years ago about how religious people live longer and have healthier lives but what they came up with was that it was the human connection on a social level each week is what they contributated to the longer and healthier lives of religious....I don't remembee the details of the study it was years ago when I read it.

  8. Maybe the more educated believers die younger because the more likely to experience large amount of stress because they probably spend more time defending their religious beliefs and from stronger adversaries. It would be interesting to see if this effect changed depending on the field of study.

  9. This is just an untested hypothesis, but I do think two thoughts might be relevant here.

    First, we know that the more educated you become, the less likely you are to attend church frequently. Second, knowing that, we might start to look for reasons why that small portion of educated people do not follow the trend. In my personal experience, there were a lot of people with MS and other degenerative diseases in church congregations looking for some miraculous healing, comfort, etc.

    So, it's possible the study is showing something like that - the more educated you become, the less likely you are to attend church regularly unless you are part of an unhealthy population. Just a thought.


Markup Key:
- <b>bold</b> = bold
- <i>italic</i> = italic
- <a href="">FoS</a> = FoS