February 07, 2006

Bad News: Statistics Show Strong Correlation Between Islam and Authoritarianism; Good News: Muslim Nations Becoming More Liberal After Bush

[Ed note: this post is a response to Dean Esmay's post which shows that at least two Muslim countries are considered 'liberal democracies', the conclusion being that Islam is not antithetical to liberalism, and Rusty's response which argues that a more meaningful measure would be to correlate the number of Muslims in a country with that country's freedom index. Demosophist and Rusty hash out the methodological problems associated with this statistical analysis here.]

OK, I did some coding last night and obtained the latest Freedom House Indices for 2006 as well as some numbers for the percent of the population that's Muslim for 164 countries. Although I have some fairly consequencial objections to Rusty's methodology regarding this data, I thought I'd go ahead and run some correlations and a quick regression using both of the composite Freedom House indices that I calculated. (Basically it's just the mean of civil liberty and political freedom, so doesn't include press freedom.) I'm not sure how to present the regressions, but since they show essentially the same picture as the Pearson correlation coefficients I'll just post those first.

Correlation between % Muslims and the 2001 Freedom House Index = 0.6044
Correlation between % Muslims and the 2006 Freedom House Index = 0.5650

For those not familiar with correlation, anything over 0.5 is considered large. But things at least seem to be moving in the right direction. As one might expect since the regression is on only one variable it shows pretty much the same pattern as the correlations. The raw coefficient for the percent Muslim for 2001 is 0.031. That means that for each increase of 1% in the percentage of Muslims in the population the level of freedom goes down by 0.031 points on a scale of 7. (Roughly 1 in 200.) The relationship is also highly significant.

In other words, the relationship has positive slope. (Remember that the dependent variable isn't freedom, but repression, because the higher the score the less free the society.)

Now, using the 2006 index the coefficient for the percentage of Muslims goes down a bit, to 0.029. However both numbers are within a 95% confidence interval. For those used to thinking in terms of beta coefficients, the betas are mathematically identical to the Pearson coefficients above, for a simple regression like this. Unlike the raw coefficients these are scaled to variation, which is why they're called "standard coefficients." They provide a little better sense of what's going on: about 0.04 for 5 years, or about 0.01 per year. (I guess it depends on when you start counting.) That's not very much in absolute terms, but it'd be interesting to know whether it's greater or less than the previous 5 year period. Is the trend toward freedom in the Ummah accelerating or decelerating?

I used only one set of numbers for the percent Muslim, because it was all I could find. For anyone who'd like to duplicate this effort, and possibly retain a few more cases, the data are here. They're for 2005 so the change in percent Muslim from 2001 to 2006 probably doesn't explain why the coefficient has dropped, since the percentage of Muslims has been growing. For the 2001 regression the percentage of Muslims is overestimated, so the actual coefficient would he greater relative to the 2006 number that this analysis shows. In other words the resistance to freedom in the Muslim world may be dropping faster than this suggests. It's hard to say how much greater unless one finds the percent Muslim data for 2001, which I don't have. But assuming the drop is real and significant (the coefficient for a "dummy variable" for 2006 is negative and almost significant at the 90% confidence level with a coefficient for the percent Muslim of 0.30) it's reasonable to suppose that the change is either part of a long term trend toward freedom, or it's a result of policies followed by the US. At any rate this analysis certainly doesn't support the Left's notion that Bush is making things worse. (We sort of knew that though, right?)

The bottom line is that Islam puts up considerable resistance to civil and political freedom, but that resistance is at least not increasing over time, and it is probably decreasing.

Well, make of it what you will.

(Cross-posted to Demosophia)

By Demosophist at 08:25 AM | Comments |