Testing: Standardized or Criterion-Referenced?

The testing is the most contested part of this whole program. One of the first things to understand is the difference between the sort of tests your teachers made in high school and standardized tests. When teachers make a test, they survey the facts and skills they expect students to have and design a task that will require those. Standardized tests are designed to spread students out over a normal curve so they can be compared; the items are chosen because they are easy, medium, or hard, not because they represent all the material to be taught. So the first and most obvious criticism is that the tests aren�t actually measuring what the standards say they should. What�s worse, though, is that standardized tests are created so that no matter what, 50% of the kids who take it score below average (that�s what average means after all). But that means by definition, if we set the 50th percentile as proficient, it is impossible to have 100% of students score at the proficient level. This is such an obvious criticism, it is hard to believe a satisfactory response is not part of the policy itself. Yet California uses the SAT-9, a standardized test said to be sufficiently aligned with the standards to measure them, even though it was designed as a norm-referenced achievement test and will continue to fail half the students in the state. What�s needed, clearly, are new tests designed to verify the subject-knowledge outlined by standards, but such tests are very, very expensive, and take three to seven years to develop. NCLB set aside quite a bit of funding for developing such tests, but this year Congress approved less than two thirds the amount the bill called for.