While reading Matt DiCarlo's review of Ohio's new school rating system, I thought I'd note that in RIDE's NCLB waiver request the appendix includes what seems to be a not-very anonymized pdf table (schools and districts have numbers instead of names) of school scores under the new system. It would be pretty tough to figure them out from scratch. You can parse and extract tabular data from pdf's (unless they're images) so if you wanted to do a real analysis of the new scoring system, that would probably be the best approach.
The only thing I have to say about it at this point is that their little systems to convert scores into point values based on a handful of cut scores drives me up the wall. Nobody picking stocks or baseball players would do that (e.g., instead of just looking at batting average directly, assign 1 point if you're under .250, 2 points for over .275, 3 points if you're over .300, etc.). My guess is the motivation for this is making the system look simpler and avoiding the "Look at the crazy long equation used to come up with this meaningless score" issue. Better to just intentionally make it a crude instrument. Either that or the whole thing is just rigged in some subtle way.
That's not really necessary anyway though, since they have a lot of discretion in the end -- they're not bound to the numbers anyhow.
Or maybe it is just to make sure the overall spread of end results looks wider and more significant.
Post a Comment