The Flynn effect
The Flynn effect is the fact that average IQ scores throughout the world rose substantially throughout the 20th century. The scores for both blacks and whites rose but the gap between the two remained essentially the same.
The effect has been something of a puzzle. Why did it happen? There are probably a number of processes causing it -- processes which could be broadly grouped as "modernization". An interesting part of the effect is that scores on subtests that load most highly on 'g' (the general factor) have changed least. This suggests that scores on a perfect test would not have changed at all.
A new researcher has fastened on to that fact and looked at what characterizes high 'g' and low 'g' subtests. He finds that the subtests which have shown the biggest change are tests where a small group of strategies allow you to answer most of the items successfully.
And that ties in with an explanation commonly given for the Flynn effect -- that ever rising number of years spent in the educational system give students more and more practice at using test-answering strategies. And they can use some of those strategies in answering IQ tests too. So education increases scores on the least-central question-types. On items that strategies cannot help you to answer (such as testing how many hard words you know) there has been virtually no change over the years.
So education has now been fairly conclusively identified as the main cause of the rising scores and at the same time the rising scores have been shown as not reflecting a real rise in underlying abilities.
Steve Sailer has the details