CHARLOTTE – Having just participated in a panel discussion about diversity in schools and the impact of high concentrations of at-risk students on test scores, I am reminded that North Carolina parents and taxpayers are about to undergo an annual ritual: the official statewide release of results from the state’s testing program.
In today’s debate, hosted by the Greater Raleigh Chamber of Commerce at a board retreat in Charlotte, Wake County school officials used their admittedly high scores on state end-of-grade tests to argue that the school district’s use of diversity criteria to reassign some of its students each year was justified. I responded that while voluntarybusing to encourage educational excellence and diversity was appropriate, forced busing was not. Moreover, I pointed out that it is difficult to prove any case in education policy by citing the state’s tests, which are fundamentally flawed.
You’d have to be living under a rock, or perhaps sequestered with the governor down at his beach hangout, not to have heard about the recent series of snafus in North Carolina’s testing program. Last year, we discovered that several redesigned math tests had “cut scores” so low that students could reach grade-level status by answering only a third or fewer of the test questions correctly. This year, unexpected and sharp drops in writing scores prompted the State Board of Education to conclude that the test was poorly worded, so they threw the results out. Weeks later, it came to light that so few students and schools took seriously their responsibility when asked to field-test new exams that the validity of the process was compromised.
As a major story in the Christian Science Monitor documented nationwide a few days ago, North Carolina is only the most prominent of a number of states with celebrated testing programs that have found themselves on the defensive about them (read the story at http://www.christiansciencemonitor.com/2002/0820/p11s02-lepr.html). While there is no silver bullet that slays all testing flaws, states like North Carolina that commission or design their own tests, rather than using an off-the-shelf test from a national supplier, run a particularly high risk of errors because they have no independent benchmark and must generate their own field tests.
Here’s the main problem with North Carolina’s end-of-grade tests: it’s too easy to pass them. Even after correcting last year’s math tests, we are still administering multiple-choice tests to our children that allow a passing score with fewer than half the questions answered correctly. That means that a student who is well trained to eliminate the obviously wrong answer and the nearly obviously wrong answer, and then to guess between the remaining two possible answers, is virtually guaranteed to get a passing score through random chance. The reason given for such a testing design – that there must be a sufficient number of difficult questions on the test to distinguish between merely a passing student and an exceptional one – is plausible but insufficient. If we need that level of precision, we must lengthen the test so that we don’t sacrifice one goal to try to obtain the other.
North Carolina’s tests are too easy, and seem to have become progressively so over time as students and teachers have learned the tricks needed to get past the cut score. That’s one reason why we have seen the average number of students at grade level rise from just over half in 1992-93 to 72.5 percent in 2000-01. There has likely been some real improvement here, but its real magnitude is obscured by the guesswork-passage problem. Public skepticism is in order – as is serious action by state leaders and lawmakers to address the problem, perhaps by throwing out the state tests altogether in favor of using a nationally standardized test. We’d lose the advantage of having ourtest specifically tailored to fit our curriculum. But the validity, confidence, and comparability to other states we would gain are more valuable.