Tuesday, June 19, 2018

Tara does the Stats - is there an Association between Religion and Freedom?

It is sometimes suggested that Islamic countries are oppressive[1].  I seek here to test whether this suggestion has any basis in fact, and whether such associations can be found for other religions.


I have looked at the freedom scores of countries as assessed by Freedom House[2] and whether there is any apparent relationship with the religions of those countries as documented on Wikipedia[3].

Freedom House scores countries on a scale of 0 (least free) to 100 (most free).

The Wikipedia article breaks countries down by religion in seven categories, which it calls Christian, Islam[ic], Irreligion (atheism), Hindu, Buddhist, Folk Religion, Other Religion and Jewish.  The accuracy of these classification might reasonably be questioned – the appear to be cultural designations rather a measure of active faith – but for the purposes of this study they are used as provided. 

For each religion, I have plotted the freedom scores against percentage adherence to that religion for all counties for which the data was available, with a line of best fit and a confidence region.


The graphs below plot freedom scores against percentage of adherents to religions, with each point representing a country, the total country population displayed as the size of the point and the various continents in different colours.

Working our way through the columns in the Wikipedia table:

There is a positive correlation between Freedom Score and Christianity, although there are non-Christian countries with high freedom scores (Mongolia, Northern Cyprus, the Czech Republic) and (at least nominally) Christian countries with low ones (Central African Republic, Democratic Republic of the Congo, Equatorial Guinea, Burundi, Swaziland).

There is a negative correlation between Freedom Score and Islam, although there are non-Muslim countries with low scores (Central African Republic, Equatorial Guinea, South Sudan, North Korea, and many others) and Muslim countries with reasonable ones (Northern Cyprus, Senegal).

There is a positive correlation between Freedom Score and what Wikipedia calls Irreligion, though there are few countries where Irreligion is very high (the Czech Republic being the exception) and this creates significant uncertainty in the relationship.

There are only two countries where Hinduism is common (India and Nepal), so the data do not provide useful evidence about its relationship to Freedom Score (we can draw a horizontal line in the grey area).

There are likewise few countries where Buddhism is common (Burma, Cambodia, Thailand), so the meaning of the data is again unclear.

We see the same issue with Folk Religion – no data at the high end (although what we can see doesn’t look good).

Likewise, there is little data for Other Religion,

or Judaism (with Israel out on its own).


So what can we conclude from this, and what can’t we conclude?

Clearly, there does appear to be a real basis for the perception that Islamic societies are oppressive, while Christian societies are not and societies with a high proportion of atheists are not.

It does not follow that Islam causes oppression – oppression could lead people to Islam (though it’s not clear how) or some other, confounding, characteristic could both cause oppression and lead people to Islam (though it’s not clear what).  But the most obvious explanation may also be true.  Similarly, it may be that atheism promotes freedom, but could also be that free societies are the ones that allow people to identify openly as atheists.  Studies using whole-of-population statistics like this one cannot answer these questions.

It does not follow that the Muslims are the oppressors, though this is the most obvious explanation and may well be true.  An examination of Islamic history may help clarify whether this is plausible[4].  Similarly, it does not follow that Christians or atheists champion human rights, though they may.

It could be argued that Islamic societies sacrifice freedom in the pursuit of some more important good.  It is not clear what this might be, but the reader is invited to offer suggestions.  I intend to address such possibilities in future investigations.


The study shows an association between Freedom Score and Islam, which is negative.

The study shows an association between Freedom Score and Christianity, which is positive.

The study shows an association between Freedom Score and Atheism, which is positive, though less clearly so.

The data are not useful for examining other religions, so I will not do so in future investigations.

[1] See, for example, https://en.wikipedia.org/wiki/Human_rights_in_Muslim-majority_countries.
[2] https://freedomhouse.org/report/freedom-world-2018-table-country-scores
[3] https://en.wikipedia.org/wiki/Religions_by_country
[4] Unsheathed – the Story of Muhammad springs to mind, at least for me.  See http://www.answeringmuslims.com/2018/06/unsheathed-story-of-muhammad-now-free.html.

Freedom map from Wikipedia.


Di Wang said...

Koran is a falsified document by roman emperor Constantine Constantine beholds goth goth

Di Wang said...

Koran is a falsified document by roman emperor Constantine Constantine beholds goth goth

Ed said...

Useful research. Thanks.

I'll add this.

Three major facts dovetail:

1) Of all the major regions of the world, the core Islamic region, Middle East/North Africa, has the worst human rights record. Worse than Asia, worse than Sub-saharan Africa, worse than South America. See https://freedomhouse.org/report/freedom-world/freedom-world-2018

2) global public opinion polls show that the majority of Muslims support some totalitarian aspects of Islamic law.

3) The core Islamic texts contain a totalitarian expansionist doctrine.

Three items that dovetail with each other to show that no, Islam is not a relgion of peace.

MCPlanck said...

I would like to suggest a follow-up study. Perhaps you could apply (as far as possible) the same methodology and standards to the world circa 1000 AD. This would establish a valuable, even necessary control point, since it would correct for cultural change vs religion (assuming Islam and Christianity are taken to be largely invariant over time).

Absent that kind of control, you're simply cherry-picking a data point. Sort of like calculating German compassion using 1944 as your only data source.

Tara MacArthur said...

Yes, this is a snapshot, taken now because now is what most interests us and because that's the data that is readily available. So it could be argued that Islam is going through a bad patch and Christianity a good patch.

That's one for the historians.

Oh, wait...

MCPlanck said...

It's not so much a snapshot as a collage. If you look at the comments you can see people taking it as evidence for universal conclusions.

This study does not prove that Islam as a doctrine is worse for freedom than Christianity as a doctrine. It simply notes that is the case today. But imagine if you did a study of wealthy, free countries and divided it up by race. Obviously all the majority black nations would be at the bottom of the scale, and places like Norway would be at the top. Would you then feel comfortable in concluding that black people are worse at making wealthy and free countries?

No, of course not. Correlation is not causation. You have to do more work to eliminate the other variables, otherwise the signal you are looking for gets completely overwritten by some other signal.

In particular it seems unfair to compare Islam against post-Reformation Christianity, since Islam has not had a reformation. I think if you study Christianity and politics before the Reformation you will see a lot more similarities to current Islam. This would be a valuable insight, because it would tell us one way to advance freedom is to encourage an Islamic Reformation.

The current study seems to imply that the only way to advance freedom is to discourage Islam in favor of Christianity. Not only is that remarkably presumptuous, it's essentially impossible and definitely counter-productive. Starting a holy war of conversion is not going to improve anyone's freedom.

This is why science doesn't cherry-pick data; because it leads to conclusions that are false, and false conclusions lead to tragedy.

MCPlanck said...


1) Go back a thousand years and Christianity is right up there with the human rights abuses. Are you asserting that the core doctrines of Christianity have changed over the last thousand years?

2) American opinion polls show that the majority of Evangelical Christians support aspects of totalitarian law. (https://www.vox.com/identities/2018/4/20/17261726/poll-prri-white-evangelical-support-for-trump-is-at-an-all-time-high). Historically, the Southern Baptist Convention was created expressly to justify slavery on Biblical grounds.

3) Christianity has been sending missionaries around the world forever. Christianity has a historical record of eliminating native religions in the countries it expands into.

If these three items prove Islam is not a religion of peace, then they also prove that Christianity - the religion of the Crusades, the Inquisition, the slave-holding Confederacy - is not a religion of peace. All of those were consistent with Christian doctrine in their time. Being Christian did not prevent people from doing those things.

To assert that the Christianity that could support those things can no longer support them is to assert that the doctrines of Christianity have changed. Is that what you are asserting?

There is another possible conclusion, which is that religious doctrine is largely irrelevant to political doctrine (and thus freedom is controlled by other factors). Or even worse, that religion is subservient to politics. Both of these are supported by the facts at hand; but the conclusion you are trying to reach simply is not.