Field of Science

Why non-religious Americans die younger

This post was chosen as an Editor's Selection for ResearchBlogging.orgIf you're 55 right now and living in the USA, this graphic shows you how much longer you have left to live!

The data come from the Health and Retirement Study, which was started back in 1992 and has been following a group of 18,000 Americans ever since. Over that time, just over 4,000 have died.

In a new analysis, Allison Sullivan of the Population Studies Center at the
University of Pennsylvania has looked how religiosity in 1992 was linked to early deaths later on. The top-line results are in the graph - Jews lived the longest, and Black Protestants died youngest. The non-religious didn't do too well - less well than mainline Christians.

In fact, in each year the non-religious were 28% more likely to die than were mainline Protestants.

Digging into why this might be, Sullivan found that one important factor is probably the divorce rate. On average, being divorced or separated raised the risk of death by 60%, and being never married raised it by 45%. The non-religious were around 50% more likely to be either divorced/separated or never married, and when she factored this into the stats, it explain a large chunk of the mortality differential.

Another important factor was religious attendance. In a model that just looked at religious affiliation after controlling for attendance, there was no difference between the non-religious and mainline Protestants.

The importance of attendance is a familiar finding. It seems to be particularly important for women, and seems to work by keeping people happier. Religious belief, and other religious practices, don't seem to have any effect on health.

In fact, what this study does is further underline the importance of social integration for happiness and health. Divorce is well known to have large and long-lasting effects on happiness. And having a group of people to meet with once a week provides people with a ready-made social group.

Of course, these effects are not all that large. If your objective is to keep healthy, there's a whole load of other things you can do that will have a much greater effect.

But I think humanists, and others who would like a society without religion, need to ponder these findings. Is there a realistic secular alternative to church going that can provide the social integration once supplied by churches? I think there probably is.

In the UK, the current (albeit unpopular) buzzword is 'The Big Society'. Nobody quite knows what it means, but in like to think that it means government providing the infrastructure to help people to come together and build a stronger community. I think it's this sort of vision that humanists need to encourage.

ResearchBlogging.orgALLISON R. SULLIVAN (2010). Mortality Differentials and Religion in the United States: Religious Affiliation and Attendance Journal for the Scientific Study of Religion, 49 (4), 740-753

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


  1. I think these are the questions humanists should discuss if they want to be an alternative to religion. Criticizing fundamentalism when it's destructive is all very well, but just criticizing wrong supernatural claims without having a community of your own to offer is just pointless. People with big social needs (the ones that will make groups stronger and thus more influential) will just ignore it and go somewhere else. Also, humanists often use the power of religion by getting members from criticizing it, rather than promoting humanism. It's understandable but I don't think building a community around a negative interest/goal is particularly attractive for most people. Making and offering secular ceremonies is a good example of offering something positive, although it's hard to build a community around that.

    Didn't CFI do a study showing that their members in some area met less than some church? Meeting frequency and attendance are really crucial. Humanists should focus more on how to get the members to meet (unless they're less social by nature - that's what you might get attracting critics and contrarians).

  2. Why do non-religous people OVER THE AGE OF FIFTY FIVE die earlier than the religous.

    You can't generalize from retirees who lived out the bulk of their lives literally in last century to all 'non-religous people'.

    In other words, congratulations on creating another internet fake-science meme, in your search for a link baiting title. One more kick at the wall, and no one will believe science reporting ever again. Which is a good thing, if the alternative is sexing up studies to the point where the news article and the study have nothing to do with each other.

  3. Aside from Blck Protestant, the "range" over all of these is only 2 to 4 years. That doesn't seem like a huge difference, or even anything really worth considering. The chart makes it look like big changes, but it's anotehr attempt to make something out of nothing. Seriously, why is anyone paying any attention to this?

  4. Luke Galen did a study showing that those who attended a church and those who attended a secular organization (I think he was looking at a CFI branch) were equally well off, and it was just those who didn't do anything who were worse off. So I think the attendance thing is right, but I think it is misleading to throw all the none's together into one category to be compared to the religious. The more equal comparison is between the religious and the atheists/agnostics/humanists who are members of organizations for that worldview.

    I think the age thing is also significant. There are major generational changes within the nontheistic community. And the younger generation is getting more organized than young nontheists ever have before. Just look at the growth in SSA affiliates (generally college atheist groups in the US, though it is now expanding into high schools). They now have over 250 affiliates, up from 100 a mere two and a half years ago.

  5. Any post correlating religiosity/church attendance with life expentancy is incomplete without mentioning the ongoing work by Darren Sherkat and Ben Moulton which shows that there is a second-order correlation with level of education. Short summary: The less educated you are, the stronger the correlation between life expectancy and religious involvement. At the highest levels of education, the correlation actually reverses itself: The less religiously involved you are, the lower your mortality rate!

    Joacim asserts:

    but just criticizing wrong supernatural claims without having a community of your own to offer is just pointless.

    I beg to differ! So if somebody is clearly blatantly obviously painfully wrong, it's not okay to say so unless you can also invite them to a party? That's crazy talk...

    It would be useful to have a good secular alternative to the role of religion (though I would argue 90% of that is already fulfilled by the modern welfare state, but anyhoo...) but the fact that such a thing has value does not mean that pointing out the epistemological, moral, and intellectual bankruptcy of religion does not also have value.

  6. James: Well, it's not exactly pointless and I do those things myself. I rather ment to not waste energy on things that don't give anything in return. Of course, if you like to do those things and get a kick out of it, then it's meaningful for you of course, but it's not like many people will care in the long run (depends on the size of the audience of course).

  7. Sullivan did a great job in her article, but one key point of focus should be on the relatively high rates of mortality of "evangelical" white sectarian Protestants. Indeed, their mortality is the same as the non-affiliated.

    And, regarding the substantive importance of this, I think living a few years longer is rather important! This is not mere statistical significance.

  8. Regarding the quality of this study:

    1. It's entirely appropriate to study older people. Younger people don't die very often in the USA (and those that do die in traffic accidents), so it's hard to pick up mortality differentials. Drug trials (statins etc) are usually done in older people for this very reason.
    2. This isn't the first or only study to find that religious people live longer. It's a fairly common finding.
    3. An increase of a few years may not sound like much, but if you found a drug that could do that in the general population it would be pretty impressive.
    4. The study adjusted for education and wealth (in fact, in this sample the religious were as well educated and often wealthier than the non-religious).

    So the findings are real and meaningful, all though not all that surprising. Just another piece reinforcing the similar findings from other studies.

    In fact, I wasn't going to blog this one until I saw the curious effect of marital status.

  9. The high mortality among evangelicals is interesting - although from Sullivan's model it looks like that's mostly explained by low education and wealth.

    I do find the divorce link fascinating because I don't think that divorce is intrinsically linked to worse life expectancy.

    Rather, I think it's another symptom of the fact that our society is in transition. We have a structure based on religious assumptions (marriage for life, nuclear family) but a changing belief system that no long fits the old social structures.

    Which is the point I was trying to make in the post. Not so much that we should create a parallel 'Humanist' church to cater for atheists, but rather that we need to find a new way of organising so that we insure people are fully integrated even without old-fashioned societal ties. Loneliness is a terrible blight of modern society.

  10. This is probably the easiest factor to impact as well - publicity alone could get a lot of folks out to already organized events - schools, libraries and community groups hold many walks, talks, lunches and etc. - probably easier to motivate people to than changing their diet (or divorce plans).


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