Field of Science

Why some countries are more religious than others

After nearly two years of blogging on other people's peer-reviewed studies, it's a refreshing change to now be blogging on one of my own! My paper, Is Personal Insecurity a Cause of Cross-National Differences in the Intensity of Religious Belief?, is now out in the Journal of Religion and Society (pdf).

The paper is a statistical analysis of the causes of religiosity at a national level (in other words, the core characteristics of a country that help to explain how religious its population is). The motivation for this was triggered by a couple of conundrums.

Firstly, studies done looking at people within a single country generally conclude that religious people are more pro-social. But back in 2005, Gregory S Paul published a study which seemed to show that, at the national level, the opposite happens. The more religious a country is, the worse is its 'Societal Health'.

Secondly, although it's commonly assumed (at least by atheists) that increasing wealth and all that goes with it (science, education, communication) is gradually eroding religion, there are some glaring anomalies. The most spectacular is the USA, which is both one of the wealthiest large nations and also one of the most religious.

Could these two be connected? They would be if societal ill health is an important factor in making people more religious, and if wealthy nations are not equally effective in improving the lot of their citizens.

There were some intriguing hints already out there. As I explained in an earlier post, it was already known that nations with higher welfare spending and lower income inequality were less religious.

So I set out to test whether this idea really could compete with the conventional theories on why nations differ in their religiosity. Here's the two hypotheses that the paper tests:
  1. Are the worst societies really more religious? And if so, is income inequality a thread that ties together the markers of societal health? If it does, then income inequality can be used as a kind of overall measure of societal health in the next step of the analysis.
  2. The second step was to work out how important income inequality is as an cause of religiosity, compared with the standard theories.

I should explain here that by religiosity I mean the intensity of beliefs, as measured by how often people pray, rather than how often they go to church or anything else. Religion is a complicated, multifaceted beast. So results that hold for one measure (frequency of prayer) might not hold for another.

Worse societies are more religious

I pulled together data on frequency of prayer from over 50 countries, and found that countries where people prayed more frequently had lower life expectancy and scored lower on the Peace Index. They also had higher infant mortality, homicide rates, and levels of corruption, and had more AIDS and more abortion. That's pretty conclusive.

What's more, countries with worse societal health also had more income inequality. In fact, the relationship between income inequality and societal health was similar to that between religiosity and societal health. Income inequality can indeed serve as a 'barometer' of overall societal health, as it relates to religiosity.

The next step was to compare the importance of income inequality as a cause of religiosity with the standard theories of why some countries are less religious. These are modernization (as measured by per capita GDP and urbanization) and regulation (both governmental and social pressures on the free expression of religious beliefs, as well as religious pluralism).

The pie chart shows the relative importance of these factors in explaining country differences in religiosity. As you can see, both modernization and societal health (income inequality) are powerful, whereas regulation is less so. Overall, about 60% of the variation between countries is explained.

How does this map out? Well, using this model you can plot the predicted level of religiosity, based on just 5 key factors (income inequality, GDP, urbanization, religious pluralism, and government regulation) against the actual levels.

It does a pretty good job. Across a wide variety of national cultures, from the Far East to sub-Saharan Africa to Europe, most countries have pretty much the same level of religion as the model predicts. The USA is still a bit of an outlier, but that's a topic for another post!

The take home is this. The standard theories of religion seem to work. The more you regulate religion, the more you turn people off it. The more you modernize the country, the more people abandon religion.

But there is a key missing ingredient that helps to explain why these standard theories are incomplete. And that missing ingredient is societal health.

Nations have choices over how to look after the people at the bottom of the social pile. Those nations that choose to make this a priority, which inevitably involves shifting money and resource from the rich to the poor, lower the overall levels of stress. And when you remove the stress caused by their social situation, people tend to lose interest in religion.
Rees, TJ (2009). Is Personal Insecurity a Cause of Cross-National Differences in the Intensity of Religious Belief? Journal of Religion and Society, 11

Creative Commons LicenseThis work by Tom Rees is licensed under a Creative Commons Attribution-Share Alike 2.0 UK: England & Wales License.


  1. Congratulations are in ordar.

    Is there data on Denmark? Just curious how much the Danes pray vs. what the model predicts.

    I have been saying for years that when the U.S. gets national health care for everybody, society will transform. I predict that in a single generation, political and religious attitudes will change markedly for the better (from my point of view). And with the U.S., so will the whole world.

  2. Very nice. Thanks for linking me to this. I suppose that it does make a certain level of intuitive sense that the more precarious a person's livelihood the more likely that person will turn to prayer. After all, it seems to me that people typically pray when they want to alleviate some undesirable situation, no?

  3. /humor mode on/

    So, perhaps all the social ills you listed are caused by praying. So, praying is not only ineffectual it is harmful. God gets ticked off by being prayed to. So, to improve a society all we have to do is to get them to stop praying. Sounds like a plan.

    /humor mode off/

  4. I wonder at the frequency of prayer stat for Islamic countries. For example given that Iran is a theocracy and there are 5 official prayers per day I wonder just what the respondents meant. Did they consider the daily prayers as part of a religious ceremony (thus lowering frequency)? And if so, isn't this a rather distorted picture of religiosity?

  5. I love the approach and believe you are finding valid associations with religious behavior at the societal level. However, your make some statements in conclusion that imply a causal direction, that restricting religion causes a decrease in religious behavior and modernization causes a decrease in religious behavior. Causality can't be inferred from your study. It could well be that abandoning religion by people leads to more interest in modernization and more willingness to impose restrictions on religion. Alternatively, there could be a third factor influencing both. For instance, a perception of international competitive forces could lead to people both abandoning religion and being interested in modernization (witness the US reaction to the launch of Sputnik, for instance). Still, I like the analysis.

  6. Thanks Bjørn :) Denmark is the dot between Russia and Sweden. The dot near the Czech Republic is France. I think that you're right, with regards to the future of the US.

    There are feedback in the reverse direction - with religion damping down social welfare. The thing that interests me is whether there are multiple stable points, which you might expect with a feedback loop. In which case, a society with a given level of religion and welfare would resist being moved until it reached a tipping point. Then there would be a shift to a new set point.

    I did quite a lot of research on this. My original intent was to look at both directions of causality. Unfortunately the data just aren't strong enough to do that. But I do have a nice analysis that shows religion seems to be a factor in increased income inequality. A topic for another post!

    Coming back to the US, there are several structural factors that work against social welfare including, bizarrely enough, the electoral system. The mechanics of First-past-the-post systems means that they are less likely to result in leftist governments. PR makes it easier for liberals to get in.

  7. jdhuey, good point. I should've made it clear that these data are for praying outside of religious services. So formal prayer like the Islamic Salat don't count.

  8. UtahGamer, you're absolutely right that the evidence for causation doesn't come from this study. The evidence comes from a very large range of studies which show that people put in stressful situations have increased religiousness.

    Building on this, I tried to find out whether this is a plausible explanation for a previous finding that unhealthy societies are more religious. My study supports that, although of course it is not the final word.

    But! It is also very likely that there is some reverse causality. Other studies have shown that religious people place less value on social welfare programmes. So there is likely to be a vicious circle in operation.

  9. This is so amazingly akin to my intuitions that religiosity as a social phenomenon tends to be highly connected to stupidity and that has been linked already to several other indices such as crimminality, family violence, jail issues, less income, less life expectancy and so on.

    I live in Mexico, can you tell me where, if it is, is our dot?

    Thank you Tom and congratulations for making public this interest of you!...

  10. Hi Victor, if you start from the 'S' in South Sorea, and travel left and slightly downwards... that's Mexico!

    I'm going to post a table showing how all the countries in the analysis score - next blog post!

  11. Nice analysis.

    I wonder about Israel (my country), which seems to be less religious than the index implies. I would guess that its reported levels of religious pluralism are higher than actual fact, and of religious regulation lower than actual fact; it's more religious that it appears in official statistics, due to its desires to appear Western. I would also have guessed the "other factos", mostly historical-cultural ones, to push it towards religion. All that would have meant having an even higher expected religiosity, however, wouldn't it? And yet it has less religiosity. Strange.

  12. Ah, looking at the table I see I misread the data. (I was just guessing being on the other sides of the USA meant *less* religiosity.) Israel is more religious than the index predicts - which makes sense in light of the biases I noted.

    The US is less religious than the model predicts?! That's weird. I was always under the impression it was far more religious than the indexes predicted.


  13. Hi Yari: high numbers = less religious (sorry, that wasn't at all clear from my post). So the US is more religious than the model predicts, and Israel is slightly less religious.

  14. Ah, then my initial problem remains - Israel is less religious than the index, despite my analysis of the bias leading it in the other direction. Intriguing.


  15. Yair, I think that the difference is small and it's difficult to read too much into it. But my guess would be that Israel is less religious than you might expect because when politicians start using religion as a prop it tends to turn people off it. Also, the survey was conducted (I think) just before the start of the second intifada. The intifada might have been expected to increase the numbers of people turning to religion on both sides.

  16. I think praying is most importantly an attempt to "do" something about your life, your intention is to change/improve your situation. So I'd suggest a connection as follows: in a relatively free and egalitarian society, you have the freedom and capacity (including wealth and security) to make decisions about your life. If you don't have these opportunities (of feel you don't), prayer is the most powerful resource you can turn to (since there are no other choices).
    Of course the sad thing is the vicious circle you mentioned in one of your comments.
    As to the US, just some guesses (I live in Europe): some local communities seem to be quite closed with members supervising each other, which may reduce levels of (perceived) freedom in making personal choices (especially if they are perceived as deviant). Also the lower level of social security may entail more anxiety and a higher necessity to rely on these closed community networks, ergo less freedom, and prayer for divine intercession instead of solving the problem yourself.

  17. I have just read your first and second posts. Noting your predictive numbers I found the assumptions all of us would normally make, proved true for most of the countries except the USA. Obviously, there is more prayer here than any of us expected. At about 2 on a 1-7 scale, I would say that this result belies anecdotal observation. It is somewhat heartening that people in general have a deeper self than is revealed in our ever distracting environment.

    I noticed that some of the comments assumed maybe less "easy" lives produced the prayer. That is somewhat superficial given that a prosperous country also has a similar "prayer value".

    Many environmental influences can impact as you have stated,but what struck me is the way USA jumps out. I am sure some prayer is what is in it for me...relief of pain, stress, straits, economy, food, etc. But there are many forms or styles of prayer...meditation on a nature level, meditation on a mystical level, or just adoration and worship.

    The types of prayer may affect the frequency as well. For example, prayer of petition, may be in times of need and lack of the stuff of life, but it might also be a want, not need. Or it could be altruistic for another or unknown persons, etc. [A blind study was done in a NY hospital, where patients suffering from diabeties with similar progressive stages were put into 2 control groups...1 group received prayer from strangers and on a daily basis for their recovery. The other did not. Neither group knew they were in a study at all. A majority, 80% of the prayed for group improved through stabilization of sugar, or healed from "diabetic wounds" twice as fast. But I digress. Just thought you might find that interesting.]

    The other types of prayer imply the disposition of the person praying: Prayers of Praise, prayers of forgiveness, prayers of thanksgiving. So when speaking of prayer in direct conclusion can be made as to why some pray more than others. The difference between taking/wanting for self and giving of self.

    The environ impacts can include everything from family, upbringing, education, community support and/or encouragement, work, play, etc. Your premise is thought provoking and well done empirically.

    I would like to see your "another post" for the outlier America. As for me, I prefer to think of us as a generous and thankful people. :-)

  18. The idea that societies with more inequality tend to be more religious is consistent with Pierre Bourdieu's theory of religion. Bourdieu argues that one of the main functions of religion is the legitimation of inequality. See:

    Terry Rey, Bourdieu on Religion: Imposing Faith and Legitimacy (London: Equinox, 2007).

    Or this early article:

    Pierre Bourdieu, “Legitimation and Structured Interests in Weber's Sociology of Religion,” in Max Weber, Rationality and Modernity, ed. Scott Lash and Sam Whimster (London: Allen & Unwin, 1987), 119-136.

  19. This is an incredibly poor paper and I am surprised it has achieved publication. You cannot attribute the individual level effects to aggregate country-level effects. That's just plain ridiculous.

  20. The links to the study don't seem to be working anymore.

    1. Ahh, looks like they've rejigged the website again. Links are fixed now!


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