[yummyfajitas]

A kids past performance and out of school issues dramaticly impact learning independent of the teacher

A VAM score is (actual score of their students - expected score of statistically similar students). If a kid has low past performance, his expected score will be low, and thus the bar is lowered for his teacher.

VAM tends to hurt teachers in top schools far more than those in the bottom schools due to the ceiling effect - if your students expected score is 97%, there is no room for them to improve.

As for teacher pay, this is a non-problem. Teachers are overpaid when you account for pension, job security and summer vacation - as a result, there is a glut of people attempting to work as teachers, rather than a shortage.

[Retric]

Edit (inner city) Poor children don't have consistently poor performance it gets much worse 8-12th grade than 1-4th. Which is not accounted for by VAM.

Students are not predicted to score 97% school wide due to reversion to the mean.

As to a teacher glut it's not a question of body's it's a question of quality. Plenty of people would be CEO of Ford for far less money that does not mean there over paid.

[yummyfajitas]

Poor children don't have consistently poor performance it gets much worse 8-12th grade than 1-4th. Which is not accounted for by VAM.

This is simply nonsense. How would this fact not be reflected in the mean performance of students grade 8, income in [$0,$15], race=white, grade 7 percentile in [25%,50%]?

I'm beginning to think that most of the critics of VAM don't even understand what it does and are merely repeating critiques of blindly measuring raw test scores.

[Retric]

Poor students in poor areas don't have the same performance as poor students in average areas.

I'm beginning to think that most of the critics of VAM don't even understand what it does and are merely repeating critiques of blindly measuring raw test scores. And your ignoring the huge statistical significant difference having an under preforming peer group has on student performance.

[yummyfajitas]

Clearly you know vastly more about statistics than I do. Could you explain why all statisticians involved in education are unable to include this specific "huge statistical[ly] significant difference" in a predictive model?

I'm also curious - if this effect cannot be included in a model, how can one demonstrate it's existence in a statistically significant manner?

[Retric]

Could you explain why all statisticians involved in education are unable to include this specific "huge statistical[ly] significant difference" in a predictive model?

They can and do. However, when it comes to teacher pay and student performance such things are politically untenable. No Child Left Behind does not mean except for when your peer group is full of truants.

The best evidence for this is actually from tracking randomly assigned edge cases. Often good schools will accept X numbers of students from another area and when they pull randomly from the pool of available students it's not hard to track what's going on and compare crossovers performance with students from each area that stayed in that area.