A sample of 50 is not statistically significant at all. There is way too much chance for variability in that sample.
They also "applied corrections" which really begs the question.
A better explanation would be that they as a group were exposed to the same curriculum in later years so they evened out. And that that curriculum did not challenge each kid so the early readers slowed down.
He then conducted two studies based on research in New Zealand only. The first compared the reading ability of 54 children who had attended Rudolf Steiner Schools (who begin learning reading from age 7) with another 50 children who had attended primary schools. Children were tested at the age of 12, at state-run intermediate schools in Dunedin, Christchurch and Hastings.
