>IQ in children is a lot more reflective of environment / happenstance

Yes, this study goes about measuring it instead of just shrugging their shoulders like your statement is.

It is correlating three variables (predictors) against IQ (the outcome variable). That is, they find that genetics, schooling, and SES all independently influence a child's IQ test score. But that school does not make the gap caused by genetics or SES better or worse.

I remember reading about this idea over a decade ago, and it comes to mind whenever I observe that I am becoming more like my father, despite living in a different country and culture, and less like other people who had an impact on me. The effects of education slowly give way to genetics in the absence of continued training.

It's also a self-reinforcing cycle.

Children that told they are smart from tests tend to have higher confidence and that has positive effects during a time where most children are unsure about themselves and their role in their world.

Confidence also has secondary effects in other aspects of life from dating to seeking out quality mentors and opportunities.