Custom Search

Saturday, January 24, 2009

Economic Freedom Correlates with Peace

[ Hat Tip Toro’s Running of the Bulls ]

Speaking of economic freedom and global unrest, an excellent post over at Division of Labour nails it: Economic Freedom Correlates with Peace.

Related Posts:
Global Violence, Gold: But Not Yet

4 comments:

Unknown said...

Statistically, that correlation is pretty weak.

Gnostic Neuron said...

Actually, as it accounts for a substantial amount of the variance (39.7%), I'd say the correlation is fairly strong (r = -0.630).

That said, I'll be the one to point out that correlation does not equal causation. :P

Unknown said...

How can you say it's fairly strong? .39 is pretty weak. I bet you could get similar (weak) values of correlation for per capita GDP or number of McDonalds per square mile or per capita internet bandwidth. It doesn't show anything, even before you apply the correlation and causation caveat.

I see near that range of R-squared on random couplings of variables all the time- average number of children per family vs local oil prices, for example. Even with a random selection on variables you have no suspicion of correlation, you don't necessarily push much closer to zero.

If you want to see highly correlated data on that data set, just looking at it, I'd say that the peace values of proximate neighbors have a lot more to do with peace values in a given country than economic freedom.

I'm not saying there's *no* correlation (or even causation), but the data on that chart doesn't support a very strong statement on it.

Gnostic Neuron said...

It might just be a matter of perspective... in psychology, a correlation of .63 is generally considered fairly strong.

Of course, you're right- you can correlate any two variables to some extent, and even statistically significant correlations are not necessarily meaningful. I certainly agree that further exploration of the data set is warranted to provide a fuller perspective, lest we are misled. As Benjamin Disraeli warned, "There are three kinds of lies: lies, damned lies, and statistics."