In other words they base this on mathematical models and not reality. Silliness. Joe Carter had a very good point about it all:
The team then applied their nonlinear dynamics model, adjusting parameters for the relative social and utilitarian merits of membership of the "non-religious" category.
They found, in a study published online, that those parameters were similar across all the countries studied, suggesting that similar behaviour drives the mathematics in all of them.
And in all the countries, the indications were that religion was headed toward extinction.
What is most interesting is that no one seems to have asked the obvious question: If religion goes extinct in societies where non-religious affiliation is more socially useful than religious affiliation, wouldn’t it also follow that religion would reach a saturation point in societies where religious affiliation is more socially useful than non-religious affiliation?
Can you imagine the panic that would result from a conclusive mathematical model that predicted America would soon be 100% religious?