One of the dangers of high-stakes testing is the opening for fraud, and yet another “miracle” — this time in Atlanta — has been exposed as massive cheating.
High-stakes testing and merit pay assume that the problem in education is that teachers have no incentive to do well. As a result, they give teachers goals to meet “at any cost,” assuming that teachers will just get better at their jobs.
If, however, the teachers are not actually prepared, trained, or gifted enough to achieve the “at any cost” goals, they will meet the goals by other means.
When one looks at the distribution of good teachers and bad teachers over, say, Washington, DC, one sees a concentration of the “highly effective” teachers in rich areas, and a dearth of “highly effective” teachers in poor areas (PDF). Simply changing the incentives will not make the ineffective teachers better.