Gary Rubinstein, the brilliant math teacher at Stuyvesant High School in New York City, has done it again.

He has dissected the Gates MET study–the one that says test scores are better at determining teacher quality than observations–and he says that the data in the study don’t make the point that has been widely reported.

Gene Glass, Research Professor at the University of Colorado, takes apart the MET study, and like Gary, says that the $50 million was a waste.

I wonder when Gates will abandon his mission to find the perfect metric to measure teacher quality. It isn’t working anywhere; it has perverse incentives; it is inaccurate and unreliable.  How long will he stick with this failed idea?

Just think how many musical instruments that $50 million would have bought, how many librarians could have been rehired, how many after-school programs might have been funded.