Archive for December, 2006

“The only way to be sure of whether or not every child is learning is to test regularly…”

– George W. Bush

I respect the intentions of the No Child Left Behind act. In no way should a school, through lack of organization, lack of will or lack of resources, ignore the needs of any of its students. All students should have an equal opportunity to succeed.

However, success is a very subjective idea – my definition of success for myself is probably different from your personal criteria.

That said, the vast majority of us are average. Yup. Ho-hum, Joe Lunchbucket average. We are mostly average athletes, average earners, average-looking and of average intelligence. Most of us are private material.

Private material? Ok, I’ll explain. The first standardized test, or at least the one that would look most familiar to us, was schemed up during World War I, when the military was looking for a quick and efficient way of finding potential officers among the sea of privates. (My grandmother knows about this. When we were discussing this very topic, she said, “Oh, your great-grandfather took that test. He had just finished his officer training when they sent him home. The war had ended.” This is clearly a gratuitous aside, but hey, it’s my blog. I can be proud of my officer material great-grandfather.) It was a test of intellect, of inherent cognitive ability, not of academic achievement. It was specifically designed to rank people, much like the military’s use of a hierarchical ranking system as its basic structure. It seems to work for them. The problem, though, is that educating children is different from developing a well-trained, robust military.

Here is the astonishing thing (I say that only ever so sarcastically), given the structure and stated purpose of modern education: standardized achievement tests are still used in the same way – to rank people. “Achievement” is a misnomer. The tests are structurally, by intent and design, used to rank students into the familiar bell curve, with the vast majority of students falling within the realm of “average.”

Now, I’m no statistician, but there seems to be a problem here. Public schools need to state certain goals (primarily for federal funding purposes), stating what percentage of their students should be at X percentile by year X. For example, ABC Elementary has to have a plan like, “By the year 2008, 95 percent of our students will score at or above the 55th percentile on the Acme Test of Student Achievement.”

Here’s the problem: it is statistically impossible for 100 percent, or even 95 percent of all students, to score at or above the 55th percentile. The nature of the tests simply does not allow that outcome. 50 percent of students within the testing population will fall at or below the 50th percentile and 50 percent will fall at or above the 50th percentile. That’s just how it works.

This is one way that happens: Standardized tests are examined and reviewed question-by-question each year. If, say, 60 percent of students got question X correct, the review panel will generally toss the question in favor of a “harder” one, because question X was “too easy.” But wait! If 60 percent or more of students know the correct answer to the question, doesn’t that mean that TEACHERS ARE DOING THEIR JOBS BY EFFECTIVELY TEACHING THAT CONCEPT!? You’d think. In practice, though, if the kids know it, it’s too easy. Since the test is designed to rank students, a question that is too easy must be scrapped in favor of a harder one.

Hence the opening quote – whenever somebody says, “the only way to achieve X goal is to…,” I get suspicious. Nothing gets under my skin more than black-and-white thinking. That said, sure, I’ll accept Bush’s judgment that we have to have some sort of method for evaluating student learning. Let’s call this method “testing.” But let’s design a test that truly does measure learning, not one that is inherently designed to show that, unlike in Lake Wobegone, your kid is probably…average.

Read Full Post »