Two kinds of tests, two roles for tests

March 20, 2004

Recently we wrote of the abrupt imposition of top-down, high-stakes tests, such as the standardized bubble-in tests now mandated by the federal No Child Left Behind law, and the new 3rd-grade exit exams.

In the comments section, faithful reader Mike McKeown wrote:

It is my understanding that good programs have diagnostic tests built in at many stages so that student weaknesses and errors can be corrected early on.

It may be abrupt to test at third grade, but if the programs being used were appropriately testing and intervening early, would most students find such a test of third grade level skills too difficult?

Mike hit the nail on the head.

We responded that this "is why the best-performing schools rely on bottom-up (teacher-made or -selected), low-stakes (diagnostic) tests, the very antithesis of this high-stakes testing craze we're in."

'M' writes, "As an engineer who often deals with QA issues, I'm appalled by your post."

Nice to know we're provoking some strong emotions, but unfortunately this is a simple misunderstanding, as we shall explain. M continues:

"Bottom-up (teacher-made or -selected), low-stakes (diagnostic) tests" rely upon an employee to rate the results of her own work. Conscientious teachers will do it honestly and get useful evaluations, although they might not be comparable from teacher to teacher. Lousy teachers will bias the evaluations in their own favor (that is, kids that haven't learned the curriculum will still pass). It's no good without an objective test that can be compared between teachers and between schools.

We have to agree with M's criticism of using teacher-made tests to perform evaluations of schools, but we were speaking of using tests to improve the quality of teaching. This second role of tests is often forgotten, but in the best performing schools, local tests are used to drive instruction.

M's point reminds us of the provisions in No Child Left Behind which allow states to determine their own yardstick for measuring a whole host of attributes, everything from "persistently dangerous schools" to school failure.

This is why the great city of Philadelphia somehow ended up with half of the "dangerous" schools in the entire country! While we've actually taught at several of those schools (and once had to receive emergency-room treatment for a human bite) we can't but help think this is an exaggeration. And some states have no failing schools, while other in other states half the schools are failing? Please.

To return to the point, when we mention teacher-made or teacher-selected tests, we're speaking of schools and teachers that use tests to improve their own effectiveness in the job of educating kids.

If your students are well-grounded when they walk into your classroom, that's great, you just teach them the knowledge and skills in your portion of the curriculum, and you're set. But it is rare that incoming kids will be so well-prepped, so instead of complaining about the absence of the utopian ideal, why not use tests to determine where kids are, and where you need to take them?

Testing--meaning school-based testing, not that at the state level--helps drive instruction at the best schools.

This is a common theme running through one of our favorite books on education reform: Samuel Casey Carter's No Excuses: Lessons from 21 High-Performing High Poverty Schools, which, while out of print, is available as a 125-page pdf file.

Casey Carter says simply, "Diagnosis is not Discrimination."

"The more you test, the better the students do," says David Levin of the KIPP Academy in the Bronx. "Regardless of what teaching style you use, there has to be a constant assessment in place that demonstrates real mastery of what you are teaching."

Angelo Milica, who was then principal of Philadelphia's Stephen Girard school, agrees:

Like so many other high-performing principals, Milicia puts a premium on testing: "Assessment drives instruction. If the kids aren't tested regularly, how are we to know what they need? And standards drive assessment. If we don't know where they're going, how will we get them there?"

Now, what about those standardized tests that have folks in a fit of conniption? We have no problem with these tests, provided they are well-written, reliable, valid, and field-tested standardized tests. We'd also like to share Nancy Ichinaga's perspective, emphasis ours:

"Every profession uses objective measures to determine effectiveness. Educators don't like the results of their tests, so they condemn the measure. But only a poor workman quarrels with his tools."