Religion and Statistics

Imagine that while reading Richard Dawkins' book The God Delusion, you come across a rather bold claim: "As it happens, in Britain, the more educated a person is, the more likely he or she is to be an atheist." Of course, such a claim would need to be supported by some sort of controlled scientific survey. But imagine that Dawkins didn't even provide a citation; he made the claim without any benefit of actual evidence outside his own experience. Would it be a big deal? After all, Britain is his home.

We can be confident that had Dawkins made this mistake, somebody would have noticed and called him out on it. Perhaps that person would have been Christian apologist David Marshall, who has made an effort to debunk Dawkins on several points. (He even has a list of "160 errors" in The God Delusion.) But imagine further that, when confronted with this problem, Dawkins offered the following response:

"Years ago I saw a study conducted in Liverpool which informed me of this correlation, and ever since then I've personally noticed that this is the case throughout the rest of Britain. I don't know where you can find that study, but you can take my word for it. I'm a generally perceptive person, and I wouldn't be making this claim unless I had a wide exposure to British people."

Would you be satisfied? Would David Marshall be satisfied? As it turns out, I don't think Dawkins did make this mistake in his writings, so "his words" above are completely fictional. But it's interesting that David Marshall, in his book The Truth Behind the New Atheism, makes a parallel claim and leaves it unsupported. On page 41 he tells us that "in East Asia, there was a strong positive correlation between education and the Christian faith. The more books you read, the more likely you were to be a Christian." This claim was not followed by a note citing any valid scientific surveys; it appeared to come straight from Marshall's own observations. When I pointed this out to him, here is the reply he gave:

"I would have liked to cite the study of faith and education in Singapore that I mentioned. The numbers were quite dramatic. They also corresponded with what I've seen all around East Asia: the more educated a person is, the more likely to be a Christian. However, I read the paper in Taiwan, and cited it in a paper I wrote in seminary, and haven't seen it since. The reader is free to disbelieve me if he likes, or suspend judgement until hard stats turn up. But I've been involved with Asian Christianity for 24 years, both on the ground and academically; it would be wise to take my opinion on this seriously. (Alternatively, check into almost any Chinese church in America, and see how many people with Phds in science, mathematics, and engineering you can find.)"

Is it alright for Marshall to put this fact in his book if it's merely based on an old study from Singapore (that he no longer has access to) and his own experience with East Asia? Perhaps it would be more responsible of him if he altered his wording to offer the information not as a categorical fact, but as his belief based on his experience.

In The God Delusion, Richard Dawkins offers some statistics (based on an actual study!) of the percentage of scientists that believe in God. Marshall appears to distrust Dawkins' cited statistics, writing that it's "a little hard to jibe with experience," since he sometimes meets devout Christians in church who also happen to be scientists (p. 39).

But consider what Marshall is suggesting to his readers here: that the statistics seem suspicious because he meets Christian scientists in church. Isn't this precisely where we would expect to find Christians (of all professions) on Sunday mornings? Clearly such an observation does nothing to deflate statistics from a properly-conducted survey. When I mentioned this to Marshall, he responded as follows:

"Your complaint about my response to the claim that relatively few scientists believe in God seems overly argumentative. My point wasn't that the data was wrong; only that based on my own experience, it seemed fishy. What I'm offering is a prima facia reason to look more carefully at the data. I didn't say only that scientists who are Christian are in church on Sunday. I said I often find the most committed Christians in many churches are scientists. But later I concede that the incidence of atheism among scientists is probably higher than the general public, and then offer some non-rational reasons why that might be so."

But it's hardly clear how meeting Christian scientists in church constitutes even a prima facie reason to be skeptical of the data. The claim made by Dawkins was not that all scientists are non-Christians; his claim was that a survey reported a particular correlation between being a scientist and being an atheist. Besides, do we need this sort of information in order to convince us that we should carefully evaluate the data?

There is at least one more incidence of this sort of problem in the book. On page 30, Marshall writes that "most atheists in modern times have been Marxists." But again there is no citation supporting the claim, it is just a bare assertion. Elsewhere, I believe he has pointed to the Soviet Union and China as virtual oceans of atheists. But even so, how many of these people were really atheists (by whatever definition of the word Marshall is working with)? And how many were really Marxists? Presumably, it must mean something to be a Marxist (perhaps a belief in some set of doctrines laid out in the writings of Marx himself, or an avid reader of those writings, or even just somebody who claims to be a Marxist). Whatever Marshall means by these words, we cannot simply assume that everyone living in a communist country is or was a Marxist. And even if we could, this still doesn't tell us anything about the number of atheists among these people, or how these numbers compare to non-Marxist atheists in modern times. When confronted with this challenge, Marshall responded:

"You complain that I don't cite the claim that "most atheists in modern times have been Marxists." But later in the book, I describe the role atheism played in Marxist ideology in some detail. There are only some ten million atheists in the United States, but hundreds of millions in communist China, who learned disbelief from Marxist teachers. (How do I know? I've been studying communist history, and talking to Chinese young people about their beliefs, for decades. Take my word for it. Or check my sources.) And many atheists in Western Europe and America during the 20th Century were also Marxists.... And my estimate that most modern atheists have been Marxists, at least in the sense that they imbibed Marxist ideology from the same source that they learned atheism, is almost certainly true."

I responded by pointing out that he still had not provided a study or survey which proved his claim, and that merely living in a Communist country probably does not qualify one as a Marxist. I further wrote:

"If Marshall has some evidence that most people who have claimed to be atheists in modern times have also claimed to be Marxists, he should have cited it when he made his claim. Either we take his word for it, as he suggests (since he's been studying this stuff and talking to some Chinese youngsters for decades), or Marshall can just point us to a study/survey which proves his point. I can't check all of his sources all the way through to see if they may have happened to include such a statistic."

And Marshall offered a longer response:

"A 2007 survey by scholars at East China Normal University of about 4500 people showed that a projected 300 million adults in China have some sort of religious belief. (Reported by BBC, Washington Post, etc.) That would leave some 700 million adults without a religious belief. How many are atheists? Given that atheism is what students have learned in school for the past 60 years, it would seem fairly safe to assume -- quite a few. My own small-scale survey supports this conclusion. When I ask Chinese to pick among a thirteen options (including three forms of Buddhism, two forms of Taoism, two forms of Christianity, Islam, "don't know," "am not sure" "other," and "atheism"), so far about 30% have indicated "atheist," and about the same percent was unsure, maybe agnostic. I haven't surveyed enough people yet for these numbers alone to be a strong argument, but this fits my general impression, in 23 years of talking with mainland Chinese. (Perhaps since before you were born.) Furthermore, religion has been growing rapidly -- before 1979, the percent of those who saw themselves as atheist was almost certainly a majority of the Chinese people.

You'll be hard-pressed to find 300 million atheists in all of the never-communist world. Four percent of Americans call themselves atheists; that's 12 million. About 18% of EU citizens call themselves atheists, according to a Eurostat poll reported on Wikipedia; that's perhaps 70 million. A lot of them are in post-communist countries, or have been Marxists in Western European countries. (French and Italian communist parties often received a quarter or more of the votes; German and other W European communist parties have also done quite well at different periods, drawing of course from atheists in particular.) Marxism in the former Soviet Union created hundreds of millions of avowed atheists over its run.

Atheists probably exist in small numbers in Muslim countries. All reports seem to indicate that atheism is very rare in Africa. Atheists in Latin America are often Marxist, and appear to be few in number. (The World Almanac, according to religioustolerance.org, counts 2.7 million.) As for India, Adherents.com (not always a very reliable site) gives just 2% non-religious OR atheist. Some of those are, of course, communists, which has been fairly strong in India.

Put all that together, and it's clear that in the broad sense of "Marxist," my comment was almost certainly correct. Over the past century, a large majority of atheists in the world have learned their atheism as part of a general Marxist education, and have been "Marxist" at least in that limited sense."

The survey that Marshall refers to, however, apparently measured how many people were religious--not how many were theists. One need not claim to be religious to believe in God (just like one could be a religious atheist). The numbers do not measure what Marshall wants them to, so he is forced to guess that the majority of non-religious people are atheists.

The other point to note, again, is that it's not entirely clear what it means to be a Marxist in Marshall's intended context. The context in his book referred to the "Marxist-Leninist enterprise," but here in his response he generalizes to the broader context of individuals who have received "a general Marxist education" (whatever that might mean).

Is it a virtue for one's book to wallow in such ambiguity?