The implication here is that American students, on average, should have gotten better at reading over the last 20 years.
But might their performance in fact be asymptotic to some maximum possible achievement level, as constrained by general intelligence?