DY: you linked to AZs post that establishes the foundation being in the Avi Chai study. There were two studies, done 3 years apart. The total number of students in say grade 2 day schools in 2005 were x% more than the number of students in grade 2 in 2002. Hence there must have been a growth rate of x/3. Simple.
But the number of students in grade 5 in 2005 sbould equal the number of students in grade 2 in the 2002 study. It does not equal. Not for any grade in the study. Enough said.