It was mostly a numbers thing. If probability of any given kid being a good match is p, I wanted a good chance of getting 3-4 good matches i.e. wanted (1-p)^(N-4) to be small, while not losing too much quality due to overpopulated class. I thought a cohort of 20 in the class year was too small. We've also had a cautionary experience with an earlier grade where too many classmates were aggressive little shits who were no fun to deal with (though with enough good eggs to counterbalance).
ok, i do agree that prior bad experience does shape ones expectations, and i could not say that i would not allow my self to be influenced by such an experience. that said, from an outside perspective, i don't think the odds are stacked like that. from my personal experience, a large class size doesn't make it more likely for any one child to make friends. on the contrary. in my class of 25-30 kids i had no friends at all. i believe that a smaller class of say 15 kids would have increased the opportunities to make friends because there would be less opportunities for others to exclude me from their activities.
even if your child makes friends easily, large classes allow the class to split into multiple subgroups, cliques that stick together. the smaller the class, the less likely this should happen. at least that is what my intuition suggests. i would put the limit for that to 10-15 kids though. any more than that is an invitation to form subgroups.
but we also must not forget the montessori aspect here, which has a strong influence on the group dynamics and individual childrens behavior.
for one, i believe that the montessori approach is driving and motivating children in a way that they simply don't show as much negative behavior as they would exhibit in a traditional class.
