Will the religious out-breed the atheists?

Tom Rees has an interesting post about the effect of religiosity on fertility, Why do atheists have fewer children? Here's my comment to his post:

And modern people have retained a capacity for atheism, which suggests some competing reproductive benefit.

I would be very skeptical of the idea that a 'capacity for atheism' must have a reproductive benefit. You basically saying that just because there are atheists around, this capacity must be selected for. That does not need to be so. Besides, if the numbers are really true, and assuming religious belief is heritable, then it is hard to argue that a capacity for atheism increases fitness.

However, I would also doubt that religiosity is heritable enough that it can overcome 'cultural evolution,' meaning that even though religious people have more children, atheism might still grow when the children are influenced by factors that cause them to lose their faith (education, in the broadest sense).

That post let me to another on the same subject, The Reproductive Advantages of Religiosity, where I posted this comment:

For religiosity to be adaptive, it is not enough to show that religious people have more children. Rates at which people change their religiosity must be incorporated. Do you have these data?

For example, assuming that all children born of religious parents become religious themselves, it is very hard to explain that there are any atheists around at all.

But from an evolutionary as well as philosophic perspective, it may seem rather odd to try to defeat nature with naturalistic arguments."

Why? Nature can inform us, but that does not mean we should bow to it. This is the same thing with many other aspects of human life. It is "natural" to get diseases and die before the age of thirty, and yet we fight it.

Lastly, I gather that when you say that there is no consensus on the meaning of the biological terms you mention (adaptation, etc.) that you mean among social scientists like yourself? Evolutionary biologists have strict definitions of the terms, of course.

Basically, I think people are over-eager and extrapolate wildly from the data they have, which is not enough to draw the conclusions that they would like to make. Crucially, religiosity is a meme, not a gene. How many children you have might be influenced by how often you go to church/mosque/temple/shrine, but as any Christian missionary knows very well, conversion* happens on a daily basis, and so the heritability of religiosity is not perfect (H²<1).

* Both ways, of course. And I suspect that in the developed world the trend is away from religiosity - in the US there is reason to conclude that atheism is "on the rise."