Sometimes my hunger for new data sets gets the best of me. After staring at Adrian White’s map of global happiness, I just had to graph some data, run some regressions and try to offer some speculation on why some regions seem happier than others.
But it’s always a good idea to know where that data came from before you spend too much time playing with it. I got an email this morning from Nic Marks at the Center for Well-Being at nef, the New Economics Foundation. Nic wrote:
Just so you know â€“ Adrian White took all the life satisfaction data from nefâ€™s Happy Planet Index â€“ published July 2007 and simply put it in mapping software and then passed it off as his own original research. See the links below for the HPI. If you read the HPI you will see that we had estimate life satisfaction for about 80 of the 178 countries â€“ as the data simply doesnâ€™t exist. It also was based predominantly on one life satisfaction â€“ not Ed Dienerâ€™s Subjective Well-being Scale you cite. You were not to know this from Adrian Whiteâ€™s paper as he obscured his sources â€“ just listing us as one of many â€“ but check the figures for yourself…
I did check the figures myself, using the data from nef’s Happy Planet Index. Burundi ranks last with a life satisfaction index of 3.0, Denmark the highest with 8.2. Multiply both numbers by 33.3 and you get the 100 score for Burundi and the 273 score for Denmark seen in White’s data…
That would be fine (though I’d understand why Nic and his colleagues were pissed off), except for the fact that White’s data doesn’t make clear that almost half the data set is interpolated. When you try to do linear regression (the stats tricks I was playing with Friday) on data that’s been generated via interpolation, you get utterly unreliable answers – you’re trying to demonstrate a relationship between two variables and you find one… because one variable was calculated based on the value of the other variable.
In other words, ignore everything I’ve written here. It’s an interesting train of thought, but not one that can be statistically verified at this point. I may try to graph just the data nef collected from various surveys, which leaves a smaller set of countries to play with, but I’m not going to be able to do that until my real job leaves me alone for a few days. Grrr…
The founding fathers of the United States declared independence from Great Britain with the memorable phrase “life, liberty and the pursuit of happiness“. The phrase was inserted by Thomas Jefferson as a departure from Adam Smith’s more capitalistic formulation, “life, liberty and the pursuit of property.” (The frequent blurring of property and happiness in American poplar consciousness may well trace itself back to this tension…)
In recent years, some governments – notably the government of Bhutan – have suggested that a measure of gross national happiness might be a better evaluation of national priorities than a purely economic measure like “gross domestic product per capita”. And academic journals have appeared, dedicated to happiness studies, or the more academic-sounding “Subjective Well-Being”. These journals produce lists of happiest and unhappiest nations, which are always good for a quick media story, proclaiming Denmark the happiest place on earth and Burundi the most miserable.
Adrian White from the University of Leicester has compiled a map of global happiness, using responses to the Satisfaction With Life Scale questionaire, a simple document designed to measure subjective well-being. It’s so simple that I can include it in its entirety:
Respond to these statements with a number from 1-7, where 1 represents strongly disagree and 7 represents strongly agree:
– In most ways my life is close to my ideal.
– The conditions of my life are excellent.
– I am satisfied with my life.
– So far I have gotten the important things I want in life.
– If I could live my life over, I would change almost nothing.
If you ask those questions of people in 180 nations and normalize the data so that your unhappiest country (Burundi) equals 100 and your happiest (Denmark) equals 273 and color code all countries by their happiness (darker equals happier), you get this lovely map.
A quick glance at the map tells you that Africa’s an unhappy place and that North Americans, Western Europeans and Aussies are happy folks. My first guess was that the distribution of happiness correlated closely with wealth. White asserts that the strongest correlation is to health, followed by wealth and access to education.
I played a bit with the subjective well-being data and graphed most of the data points from White’s work against life expectancy and saw a pretty some correlation (R2 = 0.3779, which is close to the R=0.7 White asserts…) What interested me was the fact that the countries seemed to cluster into three distinct areas: countries that were happy and healthy, countries that weren’t very healthy or happy, and a group of depressive nations that were healthy but unhappy.
(If I were a statistician, I’d do something clever like run an ANOVA test to demonstrate variance between the groups and some way to segment the set. But I’m a geek with too much time on my hands, so I just drew some circles and lines. If you’re a statistician and want to play with this data, lemme know…)
Eliminate that top cluster and you could get a pretty good equation to model the data from the black line and the lower red line. Eliminate the bottom cluster and you could use the top red line and black line. In other words, it looks a little like there’s two separate groups of nations here, one which has a strong relationship between health and happiness, another where that relationship is much less clear.
Click on image to enlarge
Most interesting to me are the nations that are outliers of the curve – nations that appear to be unusually happy or unusually unhappy as based on their life expectancy. Nations in the upper right corner of the graph are ones we’d expect to be happy, as their citizens have long lives (Denmark, Switzerland). In the lower left of the graph, we’ve got nations we’d expect to be unhappy because life is short (Zimbabwe, Burundi).
The other corners are the interesting ones. The upper left corner are nations that are unhappy despite long lifespans. You’ll note some common characteristics to these nations: they’re members of the former Soviet Union. (They’re also very cold, but other chilly nations like Canada, Iceland, and Scandinavia are quite happy…) Despite a long lifespan, Armenia is one of the unhappiest nations on earth (something I can confirm from my visits to the country.)
It’s harder to characterize the lower right corner, where nations are happier than we would expect. Bhutan lives in this corner, which we might expect from the country that invented gross national happiness. And nations that are both very happy and unusually happy include a number of tropical paradises, suggesting that if you, personally, would like to be happy, moving to the Bahamas might not be a bad start.
But moving further down the happiness scale, we see a number of nations that are happier than we’d expect based on their lifespans. Many of these nations are in sub-Saharan Africa, and a number of them (Ghana, Kenya, South Africa, Mali, Botswana, Namibia) are nations that people often point to when pointing towards the hope for African growth and development. Others include nations where civil war has settled into relative peace and prosperity (Sierra Leone, Guinea-Bissau, Mozambique.)
You could offer an interesting narrative based on this – the idea that nations are happiest when citizens think things are getting better, saddest in nations where things seem to be getting worse. Not all former Soviet nations are depressed, but those that are include some nations that fared well under the old regime and are struggling in a new economy. And the happy African nations are the ones where things are changing from very bad to not so bad, or have the potential to become leaders on the continent. As Laurie Anderson asked in her beautiful “Same Time Tomorrow”, “Are things getting better? Or are they getting worse?”
Unfortunately, there’s another way to explain the African results – life isn’t too bad in Namibia, for instance, except for one of the highest HIV prevalence rates in the world. (The same could be said for Botswana, Zambia and several other countries in this cluster.) HIV brings down the life expectancy creating a cluster of countries where life isn’t as hard as it is in Burundi, but it is tragically short. Find a way of meaningfully addressing HIV and these countries might join the family of happy, healthy nations instead of being a statistical anomaly and humanitarian disaster.
Is a post on happiness allowed to end on this sort of unhappy note? This one does.