garthsundem (1702946) writes "I wrote a post that got Slashdotted last week, suggesting three somewhat half-baked metrics that parents can use to pick and elementary school (the first two were solid!). Due to the ensuing comment war, I thought it best to call around and get the REAL take on education statistics. Wired just pubbed the result, and here's the scoop from the Journal of Ed Stats and the Dept of Ed Stats.
First, be wary of "high performing" schools — the best schools will show a nice, linear gain in test scores from year to year. A 10-20 percent jump should make you question scores' validity.
Second, classrooms increasingly include a mix of student abilities — a concerned parent might do well to look for a classroom with a more restricted range. (Warning: massively controversial.)
Finally and most importantly, my Ed Stats sources point to a Harvard/Columbia study of teachers' value added. A good teacher is HUGELY important, but currently no metrics exist to *predict* teacher quality — you can only see it in hindsight.
The best thing a parent can do: until an objective measure of teacher quality crawls from the Petri dish, gather *subjective* information about your child's potential teachers."
Link to Original Source