One way to think of this is that, instead of the physical birth rate of people with the trait, what is the rate with which the trait manages to reproduce itself? A trait of "being Mormon" with an average of 6 kids, of whom 2 become Mormon, leads to a Mormon population of stable size, forever having 6 kids per couple.
If the Mormons manage an average of 2.5 Mormon kids per couple, then their population will grow without bound, though they may never be an actual majority of the population. However history shows that culture is malleable. They may grow that fast now, but a rapid social change can destroy the dynamic. (Witness Italy and Ireland, both strongly Catholic, with declining populations.)
Now contrast this with, say, a genetic trait that manifests as a strong desire to actually have children. Eventually you get a population of people who really want children, most of whose children really want children. This type of fundamental desire is more stable than a social construct.
I believe that there are such traits out there in the population. They are much more strongly favored today than in the past. Eventually those traits, whatever they are, will become more widespread, and evolution will have happened.
For some bizarre reason, Irish people love having babies.
Full disclosure: I am a childless Irish male in my older 30's, I know many people I grew up who have had kids since the economy went into the crapper.
I personally suspect that the long oppression of the Catholic Church under English rule lead to a cementing of how important it is to the Irish people. I further suspect that, in time, Ireland will become as secular as other countries. But that time is not yet here.
Great-great-grandparent:
> Secular culture has managed to peel away many of their kids in recent decades but [...] that process is about to change
However I, personally, see no sign of that process changing. Popular culture does not have to be good to pull away the children of religious people. It merely has to be attractive. And popular culture is constantly, relentlessly, optimizing for shallow attractiveness.
If population A (e.g. mormons, people of low conscientiousness) has a birthrate of X, and B (everyone else) has a birthrate of Y, with X > Y, then population A will eventually become the entire population. The eventual birthrate will be X.
Even if both birthrates are reduced by some exogenous factor K (e.g. urbanization, reducing from X -> K x X, Y -> K x Y), this fact doesn't change.