National Education Testing

Those with a good memory may recall that during his first term, President Clinton gave some support to the idea of the Department of Education applying a nationwide standardized test to all public schools. Supposedly, testing every student in the country by a single standard would allow every parent to know both where their child stood and how well their school was performing. Clinton wanted to test students in the 4th, 8th, and llth grades. Other than that, the details were unclear, and the idea never went very far.

Those with a good memory may also recall that when the Republicans took Congress in 94, they wanted to eliminate the Education Department entirely. Well, times sure have changed. With vouchers now dead in the water, President Bush is now turning to focus on the other major campaign promise that he made regarding education, namely a new national testing system that would give the Education Department more money and power than Clinton ever dreamed about.

Bush wants to test every student in 3rd through 8th grade, every year. He hasn’t specified exactly what the new test would be like, but has suggested that it would be modeled after the high stakes exams now used by almost every state. So it seems reasonable to say that we should take a look at these statewide tests to determine whether creating a national model based on them is a workable idea.

First of all, have the high stakes tests adopted at the state level improved education overall? Common sense says no. Certainly, the mere act of testing itself cannot raise student performance. There is also statistical evidence that high stakes testing does not lead to better academic performance. Teachers, students, and outside researchers have all identified a common problem; namely, that teachers tend to “teach to the test”, or focus narrowly on topics which they know will appear on the test while failing to deliver a broader and more general education. Considering that their salaries or even their jobs can depend on test results, this is hardly surprising. And because Bush’s test would be used to determine funding for schools in some cases, it would be likely to create similar problems.

Second, we have the issue of accuracy. A report in my local newspaper looked at this issue a few days ago after it was discovered that fourth graders in North Carolina had accidentaly been given a test that was almost impossible to fail, because the organization that prepared the exam had failed to analyze it correctly. There a four major companies that create, distribute, and grade high stakes tests. All four of them have been racked by major mechanical problems in the system. These weren’t minor problems either, but major defects that affected thousands of scores and were extremely difficult and costly to fix. A single printer that put numbers in the wrong column incorrectly sent thousands of Minnesota children to summer school, while another one informed hundreds of Kentucky teachers that they would be fired if they didn’t improve their performance. However, even when the machinery runs correctly, accurracy can still be called into question. For example, many of the testing companies have trouble predicting how students will react to the test. There are cases like the one mentioned above in N. C., and then there are the exact opposite cases, such as high school tests in Arizona and Virginia that flunked more than 90% of students. The testing companies insist that these problems happen because of the sheer volume of tests that they are forced to handle in a very small time, and that they will eventually get these problems ironed out. However, if a national plan goes into effect, it will put a gigantic burden on one or more of these companies, so more mistakes will probably occur frequently at first.

In addition to accurracy, there is the issue of fairness. Most of the high stakes sytems that states used are based fully or partially on essays, rather than multiple choice. But the grading of essays will always be at least partially subjective no matter what guidelines are supposed to be put into use. In my home state of Kentucky, an independent analysis of the grading system found that quite frequently, two different graders will give the same answer to different grades, yet during the official process, most of the essays are reviewed by only one grader, introducing an element of randomness into the evaluation process.

And finally, there is the issue of time. The nationwide sytem would presumably employ essays and measure student performance in several different areas, so it would probably take a week a more for students to complete, and that’s one less week per year spent in the classroom. Also, teachers would doubtlessy spend some time preparing their students for it.

So the great debate is: what should be do with national testing? Do we implement it as is, insist on major changes or just throw it out? As you can probably tell from my analysis, I’m in the throw it out category.

Oh, and by the way, America’s public education system works much better than most people think it does.