I think that polarbear is partially right, at least from my observations locally.
There are the status-conscious parents who want the LABEL without the real challenge (which their kids
can't actually meet). For that group, we probably need to have "Gifted" classrooms and "Gifted*" ones. That way the teachers have some idea what they
should be offering.
While this is a good idea, it would never work. Somebody would rat out the administrator who sensibly tried that.
Back to polarbear's observations-- I think it IS true that if you have the full spectrum of ability in a classroom, peeling off one tail
or the other shifts the mean (where teaching is aimed in that classroom) downward or upward. This is naturally not good news for the (now) top/bottom of the remaining distribution, since teaching is shifted disproportionately toward the TAIL opposite where the bulk of the learners in that room lie. Teachers naturally aim at the middle of the classroom, or slightly under it. So yes, removing the top 10% has a pretty profound (and negative) impact on the students at the 75th-90th percentiles, who remain in the room. I'm less convinced that removing the top 2% does that, because there just aren't that many of those children. In four classrooms of 25 students each, after all, that's just 2 students. Not even one per class.
Peeling both top 5-10% AND bottom 5-10% away would probably work better-- and I strongly suspect that this is why historical data shows that tracking/grouping
works and works WELL, while contemporary data is sometimes more mixed.
However, LRE and mainstreaming make it a legal imperative to
not peel away the left side of the distribution-- only the right.
It's all about the statistics. Interesting to think about in the abstract, but a fairly knotty problem pragmatically when one considers inclusive practices for those with significant disability as the reason for being in that lower group.
