A few months ago, Richard Phelps attracted attention with an article in The Wilson Quarterly called "Teach to the Test?” Its argument is that "most of the problems with [school] testing have one surprising source: cheating by school administrators and teachers."
Last week an investigative report published in Sunday’s Atlanta Journal-Constitution found indications of standardized test cheating in school systems throughout the US.
Certainly cheating of various types is a big problem in education. But it is not really that surprising. Where else would the highest stakes of evaluation be left up to the individuals or groups being evaluated? But these articles proceed from the unquestioned assumption that state tests are an appropriate way to hold schools accountable for quality. For instance, Richard Phelps wrote, “Without standardized tests, there would be no means for members of the public to reliably gauge learning in their schools.”
The state tests are wrong both for what they leave out and for what they include.
I agree that the purpose of education is to increase academic skills. I agree that tests ought to be used to determine what students have learned. I agree that more learning is better. I do not agree with folks who say that testing is bad and that schools should not give tests because that stifles teacher creativity. I do not agree with the proposition that tests can’t measure what is important in education.
Neither do I agree, however, with the use of state-constructed tests to attempt to hold schools accountable for quality. It has taken me several years to come to this position. I have three main reasons.
First there is the issue of alignment. Whatever the state chooses to put on the test becomes, in essence, the required curriculum of all the schools in the state, even if it is wrong. The state tests are wrong both for what they leave out and for what they include. For example, state tests for elementary age students in reading and math ignore fundamental areas of the curriculum. I refer to accuracy and fluency in decoding the meanings of words, in the statement (memorization) of mathematical facts, in mathematical calculations, and in spelling. State tests simply don’t bother to measure these pillars of an elementary education, even though they are critical to future educational success.
I run six charter schools, which due to our use of a trend-bucking curriculum called Direct Instruction (DI), mostly achieve better test scores than the school districts in which we reside. DI is a specific, scripted, sequential elementary curriculum (grades K through 5) that takes much of the guesswork out of teaching. The lessons are carefully crafted to be easily understood, build only on what has been taught in earlier lessons, and prepare students precisely for what is to come. There are programs for reading, math, spelling, and writing. All but the very lowest special education students can learn from these programs and emerge from elementary school with average or above average skills. DI is hated by the progressive educators at universities, but we love it, and so do our students and parents.
Curricula such as DI that focus on bringing all the fundamental student skills to mastery (including the ones not tested) must do so on top of teaching the things that are measured on the test — while other schools focus all their efforts on the test material. A majority of American elementary schools no longer teach spelling, for example, simply because it is not measured on the state tests. While learning how to spell is an essential skill, the state tests have pushed it out of the curriculum. Not to mention all the other critical content not tested and no longer taught.
Conversely, state tests focus strongly on a number of things that, although they sound good, are not skills to be taught but attributes of intelligence that we desire. These attributes are such things as the ability of bright elementary students to make inferences from unfamiliar texts, to write interesting imaginative stories, and to find creative solutions to unique word problems in mathematics.
These attributes, and their application, are not an emphasis of the very strong DI elementary curriculum. But if schools that use DI, such as my own, taught what is in our curriculum (what kids need) and ignored the less relevant, they would get lower state test scores and be branded as poor schools. Schools ought to be able to use their own tests to measure what their own curriculum plans to teach, and be evaluated on how well the school does what it claims it will do. Parents, of course, could select schools according to the nature of their claims as well as their performance.
Second, people forget important facts about state tests. One is that the results have no consequences for the children. Another is that these are children taking these tests. Children are subject to wide swings in their performance, often depending on testing circumstances. In our schools we have found children who have been well taught but who for years have failed the test. Yet they can reach not only "proficient" but “exceeds proficient” if their teacher sits next to them and makes them read the test aloud and gives them breaks when they get tired. Essentially we are making certain that they actually do their best on what to them is a very long test. This is not cheating. These practices are specifically allowed by the state rules for students who need them; they are called an “accommodation.”And it is an appropriate accommodation. It just shows the best that the student can do. Guess what? Children don’t always do their best. Sometimes they just guess their way through the test to get it over with. If those children go to another school, where no one they know or care about is monitoring their test performance or where they are allowed to do fun stuff when they are “done,” they will probably turn in a failing score the next year.
If we expect teachers and administrators to want to work with populations that are below average in some way, we have to stop proclaiming that those who teach the smarter students are better teachers.
Third is the issue of students' ability. Obviously, the more able students are, the easier it is for them to learn. The less able they are, the harder the teacher and the school must work to teach them. Scores on state tests are as much a measure of how smart the student body is, as they are a measure of how well the teachers teach. It is ridiculously unfair to ignore this fact and proclaim that high test scores mean a school is good and low test scores mean it is bad. That would be true only if the student bodies of the schools were evenly matched in IQ — which is never the case. It is a heavier lift to raise test scores in a school that enrolls many students with low ability, or learning difficulties; and until we begin to measure the weight of the load, we cannot claim to know who is stronger. If we expect teachers and administrators to want to work with populations that are below average in some way, we have to stop proclaiming that those who teach the smarter students are better teachers, just because their students get higher test scores.
We would be far better off if the states stopped giving their tests, instituted more school choice, and left it up to schools to find a way to prove they were doing a good job for the consumers — just as it happens in every other service industry. We could do it easily in our schools, without a state test. If we gave aligned end-of-year final exams for each of our DI programs and shared the results with parents, they would be blown away by what we teach. Few students outside of our schools could match that performance. That’s how you prove quality, not with bogus, we’ll-decide-what’s-important-to-learn, state tests.