The Misuse of Standardized Tests in America

There seems to be a national focus on using standardized achievement tests to not only compare students' knowledge and/or skills with those of other students across the country, but to judge the quality of education in this country and to label schools as high-performing or low-performing. This is a mistake.

Standardized achievement tests are not accurate measures of educational quality for several reasons. First, there is a rather significant amount of diversity in curricula across the country, and this is based on varying ideas as to what is most important to know. Whether or not this is a good thing is debatable. Nevertheless, the American school system is designed to maximize state and local control of curricula, as opposed to there being a national standard. The problem arises when a single one-size-fits-all test is administered to students (say, 8th graders) nationwide. A single standardized test cannot possibly align with only those topics taught in the classroom, in each classroom in America, as the list of topics varies. (Perhaps it does not vary a great deal, but it certainly varies.) One may argue that there is a core set of ideas that ALL students should know, and a standardized test could test just these ideas. I would agree, but again, there are no national standards as to what students should know, and who is to say that Company X or Company Y is to make the decision as to what everyone should know and when they should know it? In any case, at present (and likely into the future) different schools have slightly different objectives and a single standardized test administered nationally is not going to measure if students have met the objectives decided upon by their local or state administrators. This makes such a test invalid for judging quality of education in specific schools or districts.

Second, standardized achievement tests are designed such that test items (or questions) are answered correctly by about half of test-takers. This is done for statistical reasons; primarily, it spreads out students' test scores, making it easier to rank the students. If a question is answered correctly by most students, the question will likely be dropped from the test, as it doesn't help differentiate between the students. But ... the questions answered correctly by most students nationally generally cover the most important topics, i.e. topics that were stressed by teachers. The company that produces and markets the standardized test has an incentive to use questions concerning less-important concepts. Does such a test truly measure educational quality?

Finally, such tests often, perhaps inadvertently, measure things that students do not learn in school. Questions often test a student's innate intelligence and out-of-school learning. Now, I'm not really comfortable asserting that some people are inherently smarter than others, but it does seem reasonable that not all people are born with the exact same capacity for math, or languages, or art. Should a school be penalized for failing to teach students something that, by definition, cannot be taught? Regarding out-of-school learning, students are born into different socioeconomic classes and are raised by different parents, both of which lead to different life experiences. If a kid has never been taken to the beach before, perhaps for financial reasons, and a test question asks something about ocean waves, he or she may be at a disadvantage compared to other students that have been to the beach. Such questions do exist on standardized tests, and they invalidate the test as a measure of educational quality. Why are they put on the test? Quite likely because it is known that not all students will be able to draw on the same experiences, and this can be exploited to find those ideal test questions that are answered correctly by only one-half of all students.

I don't think we should do away with all standardized testing, but we need to understand what it is that we are testing and not misuse test results. The current system is not working. Test results are being misused. Teachers are feeling pressured to "teach to the test." Schools are forced to hyper-focus on a single, arbitrary measurement of student knowledge for political and financial reasons. Schools shouldn't live or die based on these kinds of test results. This focus on standardized testing is not making American students any smarter. We need to either improve the tests (which cannot be left up to the companies that currently produce them) or change the way we use their results.

(BTW, I found a lot of useful information for this essay in an article written by W. James Popham for the March 1999 issue of Educational Leadership.)