Originally Posted by Bostonian
Approximately 2/3 of standardized normally distributed samples are between -1 and 1, with 1/6 below -1 and 1/6 above 1, so your suggestion would result in the middle group (-1 to 1) having 4 times as many students as the bottom group or the top group (since 2/3 = 4*1/6).

Thanks for putting numbers with it. My questions is really about what prevents teachers from meeting students at their level effectively. Is it the number of students, or the discrepancies in ability? A combination? My guess it's a combination, but what is the proper weighting of those factors?

The class limits could also work in tandem:
1) No class shall be larger than 25 kids.
2) No class shall include students more than 2 SDs apart.