A corrupt incompetent cronyistic mess is going to be a mess whether or not you use a small set of gamed statistics or nothing at all.
The reporter doesn't get it when they critically remark that the reform model "ignored less quantifiable signs of intellectual development" - no, what would happen if it didn't is just that the people who are willing to make up test scores en masse will also make up all those 'less quantifiable signs' except you won't have any way of knowing about their bullshit.
The NCLB gaming of standardized tests, on the other hand, forces the incompetence out into the open in the form of unmistakable undefensible clear outright fraud.
So, which is better? To have all your metrics gamed and to not know it, or have them gamed and a chance of detecting the fraud?
> The NCLB gaming of standardized tests, on the other hand, forces the incompetence out into the open in the form of unmistakable undefensible clear outright fraud.
The outright fraud isn't the only thing it forces (and, really, its sort of the extreme fringe); the more routine thing it forces is reluctance of schools to promote students based on academic ability since holding advanced students back in grade level improves metrics of students that are function at or above grade level.
(Of course, its an unsustainable optimization, which also increases the rate at which advanced students with involved parents defect from the public school system entirely, but that's an effect that takes longer to materialize than the short-term effect on metrics.)
The reporter doesn't get it when they critically remark that the reform model "ignored less quantifiable signs of intellectual development" - no, what would happen if it didn't is just that the people who are willing to make up test scores en masse will also make up all those 'less quantifiable signs' except you won't have any way of knowing about their bullshit.
The NCLB gaming of standardized tests, on the other hand, forces the incompetence out into the open in the form of unmistakable undefensible clear outright fraud.
So, which is better? To have all your metrics gamed and to not know it, or have them gamed and a chance of detecting the fraud?