Alex is right on one point, that Bryan's book is very important and everyone should read it. Otherwise, his discusson of Yana indicates that both genes and upbringing matter, which is my point — that both matter. He mistakenly cites that example against me, rather than against Bryan, who claimed upbringing and inculcation do not matter for language.
Orthodox Jews are clearly a case where parenting matters for the religion and religiosity of the kids, not just the abstract fact of having the parent or the possible genetic transmission of general religious fervor. Orthodox Jewish parents are effective, in part, because they inculcate the religion in their children. Or look at the data on Korean Protestants. I could spend a whole day finding credible studies on this question and they are not restricted to a few extreme points (though they also don't cover everyone, and probably strict inculcation of Unitarianism won't work). There is overwhelming evidence that parenting often does transmit religious observance and if twin adoption studies do not show that it is a sign of their limitations.
In the comments, Cournot (not Augustin) nails it:
Tyler is not rejecting the studies. He's rejecting Caplan's interpretation.
Imagine this were a drug and the trials showed that the drug clearly reduced fever in a stat sig group.
However a number of non-statistically significant side effects were observed. Some patients died, got rashes, or became crippled. We also have anecdotal evidence and even some good theory that there may be weird interactions with some subpopulations, but to date no good studies showing these effects have been conducted.
Should I assume that it's safe to take these drugs? I think not.
Indeed, doctors often warn us of side effects even though most of the side effects don't show up as statistically significant in clinical trials.
Tyler is just saying some of the side effects of parenting on kids may well be significant and until clearly proven otherwise, it is unsound to ignore non-statistical anecdotal evidence [TC: there is plenty of statistical evidence also] just because the major studies don't show any effects.
And here is Richard A.:
The Body-Mass Index of Twins Who Have Been Reared Apart
We conclude that genetic influences on body-mass index are substantial, whereas the childhood environment has little or no influence. These findings corroborate and extend the results of earlier studies of twins and adoptees. (N Engl J Med 1990; 322:1483–7.)
IOWs, the reason why white kids of today are much fatter than white kids of the 50s and 60s is due to genetic influences and environment has little or no influence
This shows that the twin studies are flawed.
If that doesn't convince you, ponder the Zoroastrians, an extreme minority in most of the places they have lived. Parents inculcated religious observance for centuries, yet now that pattern of replication seems to be falling apart and the religion and way of life is in danger. In other words, now we are viewing the time series evidence (as does Richard A. in the case of body weight). Parental inculcation can matter a lot, or not so much at all, depending on circumstances. The very size of that variation suggests two points: a) cross-sectional evidence alone won't pin down the proper genetic component, for instance studying this group today, rather than across time, would overestimate the genetic influence, and b) if the transmissibility can change so quickly we are pushed away from genetic explanations, though of course without abandoning them altogether. Genes are still very, very important.
In contrast, what do you expect for the heritability of IQ over the generations? Lead poisoning and the like aside, I expect estimates of that heritability don't change much over time at all. That is more closely tied to genetics, and the time series evidence confirms it.