Affects similar to the "relative age effect"[0] are really important. Kids that just so happen to get ahead early (whether by a slight age advantage, or any other) often end up being put into some sort of "gifted" track where they have access to better teachers/resources, which compounds until they are well ahead of the others. Even without any sort of gifted program, the self-confidence that gets built around being "the smart kid" is a powerful motivator. The fight for self-confidence in school classrooms is zero-sum.
What surprises me is that, from what I've read, the IQ literature doesn't seem to take this basic compounding affect into account. So it seems like some skepticism should be held for the idea that IQ is ~50% genetics and ~50% environment. It could be 5% genetics that compounds into 50%. That would seem to be in line with the observation that the measured heritability of intelligence increases with age.
> What surprises me is that, from what I've read, the IQ literature doesn't seem to take this basic compounding affect into account.
According to most of the IQ literature, a person's IQ does not significantly change over time, which means that any compounding effect you might find in academic or professional performance is completely unrelated to the person's IQ actually increasing because of that.
> According to most of the IQ literature, a person's IQ does not significantly change over time
If you're saying that a person's environment (e.g. education) doesn't causally affect abstract reasoning skills (as measured by IQ), then that is not correct according to the currently available evidence.
"We looked at 42 datasets using several different research designs and found that, overall, adding an extra year of schooling in this way improved people's IQ scores by between 1 and 5 points."
First of all, the additional year of schooling has not been shown to compound, i.e. it works for 1 or 2 years, doesn't mean it works for several years (if it doesn't, then as I had claimed, the gains are NOT very significant).
Second, the highest change (5 points per year) was for the cases where "Students who made an age cutoff to begin schooling were compared with students who had not.", which seems to me to imply the difference was of only one year (students who did not make the cutoff had to wait another year to get in), while the other cases, which are what we normally think of when considering years of schooling (i.e. students who were going to get the same number of years of schooling VS the drop-outs) the increase was only 1-2 points over a few years maximum... again, not really very significant (I would consider at least 10 points to be something significant in terms of changing possible outcomes for the person).
Scores on any test can be improved with practice and training. The SATs are a notable example, but similar test preparation techniques will work for any other aptitude test and IQ tests are aptitude tests.
I'd go so far as to say that the rising prevalence of test preparation is critically reducing the signal these tests provide. When one cohort routinely prepares for the aptitude test and another one doesn't, the latter is going to appear disadvantaged, but what's being measured is no longer what the test was designed to measure.
Top tier American universities have long been aware of this effect, which is why they negatively weight certain groups' test scores to balance out the bonus they are getting from test preparation. It's gotten so bad that the UC system is abandoning standardized testing altogether.
That doesn't jive with my experience. When I was growing up in Ontario, there was a program at my high school that put students going through the gifted program (from elementary) together with students who tested as excelling at math & science in grade 8 into the same program. I never felt like the "gifted" students as a cohort excelled in any significant way (overall grades, aptitude for any specific topic, etc). Now the data sample here is obviously small & anectodal, but to my knowledge this is supported by scientific studies on the topic of gifted programs.
What surprises me is that, from what I've read, the IQ literature doesn't seem to take this basic compounding affect into account. So it seems like some skepticism should be held for the idea that IQ is ~50% genetics and ~50% environment. It could be 5% genetics that compounds into 50%. That would seem to be in line with the observation that the measured heritability of intelligence increases with age.
[0] https://en.wikipedia.org/wiki/Relative_age_effect