Field of Science

Less God and more democracy

A new analysis from the University of Zurich rates 30 nations on the strength of their democracy. All interesting stuff, but it's at times like this that I like to ask 'How does it correlate with religion'.

So, if you were wondering the same, here's the answer. Those countries whose population rated God as less important in their lives, were also the countries where democracy is strongest.

The relationship is statistically significant, with an r-squared of 0.16 (which means that about 16% of the variation in god-belief can be 'explained' by variation in democracy - or vice-versa).

Not a great surprise, of course. It's just the latest in a long line of similar results - the least religious countries are more peaceful, have less corruption, more telephones, do better at science, have less inequality and other problems, and are generally just less dysfunctional.

As usual, those dastardly Scandinavian countries, with their strong social welfare programmes, liberal morals, and strong social ethics, come out on top on both scores.

This is a nice sample of nations, though. Because they're mostly pretty wealthy (the poorest, South Africa and Costa Rica, have per-capita GDPs around $10k), it's not too badly distorted by wealth.

Just a note on the data. The religion numbers come from Waves 4 and 5 of the World Values Survey (I used Wave 5, unless a country was only represented in Wave 4). I used the "Importance of God" question because it's the only one asked consistently in both Waves.

The democracy number "uses 100 empirical indicators to measure how well a country complies with the three democratic principles of freedom, equality and control as well as the nine basic functions of democracy" (Science Daily).

So there you go. More evidence that the least religious countries are the best places to live. Who'da thunk?

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


  1. On your figure there, it looks like there are four points that could be mild outliers. They're cases where democracy is weak and religion is relatively unimportant. I was wondering what countries those are.

  2. I have to say it makes sense to me, as a lot of religion allows for distorted interpretation of its teachings, which can have a negative effect on a society.

  3. But in which direction is the causation? Perhaps people need religion to cope with dysfunctional societies :)

  4. I would guess the outriders are secular Arab regimes like Syria and Egypt.

  5. Maybe it's just because I'm an engineer and not a social scientist, but an R^2 value of 0.16 is piss-poor to me. Yes, it means that 16% of the data is explained by this simply regression, but equivalently it means that 84% of the variation can't be explained by it! I'd say the model is pretty worthless and you can't really draw the conclusion that there's a relationship between democracy and religiosity in this manner.

    And Anonymous' point is a very good one as well.

  6. By the way, can you link me to the original article you have this r-squared parameter and figure from? I'm having a hard trouble finding it. Thanks!

  7. Andy, those three outliers with low religion and relatively weak democratic structures are Great Britain, France, and Japan.

  8. Waldheri, yes - I think that's the reaction of most physical sciences. But an r^2 of .16 is actually pretty good for the social sciences. Most of the studies I report here come in a lot lower than that. The reason, of course, is that you can't control the environment in the way you can in other scientific fields, and so there's a lot of noise.

    Also, I wouldn't make a big deal out of this correlation. It's only an n of 30, and it isn't even an attempt at a model. It's just an interesting observation.

    On it's own, it wouldn't amount to much. But the interesting thing is that it's a fairly consistent pattern, regardless of which measure of 'societal health' you look at.

    As for the figure and r-squared, well the reason you cannot find the source is that they are mine! I did 'em myself, in excel.

  9. wald

    in physical sciences, you can control every variable you can identify, take as many readings as you want, and the variables of interest are intrinsically interval-ratio in type. none of these conditions is true, even remotely, in the social sciences.

    also, you can't *possibly* think anonymous's point is valid. you just disputed that there is a (linear) relationship at all, you can't then turn around and say consistently "but maybe y causes x".

  10. I understand that the subject matter of the social sciences is vastly more complex than that of the physical sciences, and that proper testing is very hard in that field. If an R^2 of 0.16 is pretty good for the social sciences, I fear for it - and I don't mean that in a cynical way.

    Looking just at the scatter-plot of your data points, nobody would have been able to guess the regression line. It's all noise: no clear pattern whatsoever. To then start jumping blindly through statistical hoops hoping to find a pattern shows a lack of critical thinking.

    You argue that maybe the relationships are weak, you find ones in the same direction. You could also say that there's consistently not been found a clear relationship between the two. Beware of your cognitive biases, because this smells of confirmation bias - you left out those pesky other studies that paint a different picture (if you accept that they paint a picture at all), for example the success of religious communes over secular ones.

    Anonymous: I meant that it was a good point if there wás a clear correlation, but upon second reading Tom already mentions this in the article ("... which means that about 16% of the variation in god-belief can be 'explained' by variation in democracy - or vice-versa") so I retract that comment.

  11. Curious about this sentence:
    "uses 100 empirical indicators to measure how well a country complies with the three democratic principles of freedom, equality and control as well as the nine basic functions of democracy"

    Are the defining "Democracy" as "Socialist Democracy"? Like "Democracy", all those defining words are weasel words and could be made the flags of many political systems.

  12. Waldheri, Sosis' study on communes (which I have blogged about) is interesting but not relevant here. It suggests that religion can make small groups more cohesive, especially when the 'entry fee' (in terms of ritual demands) is high.

    What the current post relates to is actually a fairly widely recognised phenomenon (at least, among sociologists of religion), and it goes by the name of 'modernisation theory'. Wealthier, more stable nations with higher quality of life tend to have populations that are less religious. Quite why that should be is hotly debated, but not really disputed.

    Lastly, confirmation bias it two-edged. You need to be careful about rationalising away evidence that conflicts with your preconceptions!

  13. Sabio, I dunno about "Socialist Democracy". I think their point simply is that there's more to democracy than voting for a president once every few years. Incidentally, the USA scored quite well - higher than France, Germany, UK etc. The link I gave has the full rankings.

  14. Waldheri, you're missing the point.

    You can only find or would test against an R2==1 if you'd suspect a 1:1 causation between two variables. But everyone knows there's a lot of variables influencing how perfect the state of your democracy is, not just religion. So there's simply no way to find an R2==1, no surprise or data cooking there.

    Tom just looks at the data and finds a 0.16 R2, so there's some relationship between democracy and religiosity - within this particular data sample. That's all. It's not a model, it's just one tiny data point about how religiosity and modes of society correlate.

    P.S. Tom, congrats on making it on Galileo's official list of (increasingly desperate) science deniers. I hear the company is nice there.

  15. Sorry, I wrote Galileo, but I meant Michael Blume. I always confuse the two.

  16. @ tarik

    Could you spell out your allusion to the Galileo allusion for those of us of less subtle minds -- it sounds interesting.

    Are you calling Tom a science denier? Are you calling Michaele Blume a science denier?

    Sorry, I couldn't follow.

  17. Sabio, check Michael Blume's new laments at Like that:

    And it is quite ironic to see very unlikely alliances emerging i.e. among religious fundamentalists such as answersingenesis and some antitheist scholars such as Richard Dawkins or Tom Rees jointly (and increasingly desperately) ranting against evolutionary studies of religion in order to protect their respective worldviews. But others such as Susan Blackmore showed the integrity and courage to accept the empirical evidence. It seems that this time, the scientific field cannot be muzzled again as it has been during the first half of the 20th century.

    In Germany that goes on for a long time: Michael doesn't address arguments (as he did not adress Tom's), he just feels attacked by anti-religious science deniers.

    Regards from Berlin.

  18. @ tarik

    Ah, thank you -- that helped. You are saying that Blume is accusing Tom of anti-science thinking.
    Both Tom and Blume see themselves as pro-science. What do you think an intelligent approach to mediating this would be?

    Do you feel Blume is just outright wrong?

  19. Sabio, There's a lovely quote from Keith Parsons, a philosopher of religion who recently announced that he was giving up philosophizing about religion (the quote is in the comments):

    "First of all, let me make clear that personal asseverations about what one does or does not find vacuous--whether asserted by me, Richard Purtill, Peter Kreeft, or anybody--have zero epistemic significance. I take that to be obvious ... Had Purtill or Kreeft made such an announcement, my response would have been "meh," and a shrug of the shoulders. That is exactly what I would expect from anyone hearing me assert that I regard the "case for theism" as vacuous if that person disagrees."

    In other words, who cares what my or anyone else's opinions are on this topic - although I'm always happy to hold forth :)

    For the record, though, I think that most of what is written on this topic is way too simplistic. I've touched on some of the complications in earlier posts. But there are many other challenges, particularly from the perspective of genetic fitness.


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