What's actually being argued about here? I'm confused. Two things seem likely to be uncontroversial:
- age of [self-taught*] reading is negatively correlated with IQ (earlier age, higher IQ, *on average*)
- the absolute value of the correlation coefficient is likely to be quite small.
If there is a correlation, it's possible to use one thing to predict the other; that's what correlation means. If it's quite small, one's predictions using this method will not be very accurate.
So...? This thread is about the size of r, or...?
* Without "self-taught" the correlation will surely still be there - because we have no reason to think the taught readers will correlate in the opposite direction from the self-taught ones, and the latter are still there - but the size of the correlation coefficient will be even smaller. We know it's possible to teach almost any baby to read; Glenn Doman made a fortune out of this observation.
