Given this comment from Ruf below that blog post I'd be reluctant to believe anything she says about the Flynn effect!
The same is true, by the way, for kids who need support services at the low end of the ability spectrum, e.g., the scores drift away from average at that end, too. This means that [when the identification test is new] a child with learning difficulties, a slower learner, will score HIGHER than the cut-off score needed for him to qualify for special education help in school. Flynn himself uses the example of death row cases where prisoners whose IQ's are too low are considered incapable of understanding that what they did was wrong; they're spared the death penalty. But when tests are new, more of them miss that low-end cut-off and are put to death.
(She has it backwards.)
