Field of Science

Live long and be atheist

The splendid World Happiness Database has released a new analysis of their 2009 data. Basically what they've done is to multiply happiness scores in each nation with the life expectancy. The idea is that what most people want is a life that's both long and happy.

Costa Rica came out top, followed by the usual gaggle of Northern European Countries (and Canada).

Now, there's an 'ecological' problem in this analysis, in that the people with long lives in a nation aren't necessarily the happiest. What's more, happiness might be very unevenly distributed in some countries.

And being grumpy might have its plus side - in the news yesterday was an Australian study which claimed that grumpy people are less prone to errors of judgement!

Be that as it may, whenever these national statistics come out I always like to correlate them against religion, to see how they stack up. So here's the results for this one.

What I've done is plot the percentage of hard-core non-believers in each country against the 'happy-life-years'.

In the top graph, it's the percentage of people who say they are 'confirmed atheists'. In the bottom graph, it's the people who say that religion is 'not at all important' to them.

There is a weak, but statistically significant relationship - especially with the unimportance of religion. What's more, the correlation is about 50% stronger than with happiness alone.

However, digging around in the data shows that the is mostly driven by life expectancy. Average happiness, by itself, is not related to the number of atheists, and only marginally related to the number of non-religious.

Now the interesting thing is that happiness is strongly correlated with life expectancy (as you might expect). So you also would expect a correlation of happiness with atheism - simply because they both correlate with life expectancy.

The fact that this does not happen suggests a negative interaction. What may be happening is that some countries with short life expectancy are particularly religious. That makes them happier than you would expect, and confounds the straightforward link between long life expectancy, happiness and atheism.

To put it another way, turning to religion has the effect of increasing happiness. But good life expectancy is more important, and countries with good life expectancy are the happiest and least religious. 
Creative Commons License This article by Tom Rees was first published on Epiphenom. It is licensed under Creative Commons.


  1. This post has been linked for the HOT5 Daily 11/4/2009, at The Unreligious Right

  2. I suspect that Costa Rica coming out on top is a statistical anomaly based on some questionable data published by a single person.

    Google for Rosero-Bixby. He is a Costa-Rican demographer who has identified what he believes is a "Blue Zone", an area where life expectancy is unusually long, in part of Costa Rica.

    It appears he has worked rather hard to control for the usual data problems that give the appearance of Blue Zones (the most common being that people are just lying about the age of village elders -- and maybe not even consciously lying, it is sometimes a cultural thing) but IANAS (I am not a statistician) so I have no idea whether his methodology is valid or not.

    What sets off my skepticism detector, though, is that the only people who have published papers on unusually long life expectancy in Costa Rica are Rosero-Bixby and his team. Maybe it's just not a high visibility topic, but the lack of independent confirmation of his data makes me mighty suspicious.

    So, if you are multiplying life expectancy by a happiness index, and you have an anomaly in your data on life expectancy...

  3. Cista Rica does have a high life expectancy, according to the data at least. 78.5 years - not quite the highest, but nearly there.

    But it also has very high life satisfactions score - 8.5. I think that's the highest of any country surveyed!

    I forgot to put a link to the data in the post:

  4. What makes you think there's a true correlation between the two factors? I'm not a religious person either, but I can't see that your conclusion is anything other than the usual crunching of numbers. Have you controlled for other possibilities, or did you just plot the two variables against each other. I'm always doubtful when I see statistics like this, they just seem like someone played around, but not like serious scientific calculation. And I'm even more doubtful when someone transforms a small statistical significance (how high is it?) into a result that sounds like yours. That is just plain unethical.

  5. Well, as I said in the post the correlation is weak and there's an ecological problem that means any kind of causality has to be taken with a hefty pinch of salt.

    This clearly isn't an academic exercise, just a quick and dirty exploratory analysis.

    Having said that, there is good reason to suppose (based on what been found in published analyses) that atheism is linked with life expectancy. Basically, people who live in countries that have good healthcare provision and low infant mortality are less religious.

    There is also good evidence that, within a given cultural environ, religious people tend to be happier (well, maybe not compared with atheists, but certainly compared with less religious believers).

  6. The high life expectancy is true, I'm a costarican and I can say that our socialist health care system... founded by the unholy alliance of the church, the elite and the local communist party, are responsible of that.
    And with the happy thing it's easy, we are a chauvinist group of people who never leave the country, so we think we're the best in the world... and we don't have army, so we have a lot of money to be happy.

  7. Totally agree with the above comment and very nice article...!


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