The Illusion of Accountability
Myth-busting the most common arguments for high-stakes testing
Most people, even some educators, don’t really understand high-stakes standardized testing. This lack of understanding continues to fuel our national obsession with testing, even after knowing the harms it inflicts.
The good news is that you don’t have to take a graduate-level statistics course to get a real idea how high-stakes testing works. It just takes a little myth-busting. Here are some of the most common arguments for high-stakes testing.
Growth in test scores over time will tell us if a school is improving.
If the test is norm-referenced (that is, scores are compared to a national average), growth may not indicate improvement. A school’s rating over the years, for example, could mean that schools across the nation improved and therefore, even if the school did indeed improve, they are still average. Normed test scores that show lack of statistical growth could hide the fact that improvement really was occurring.
Changes in test performance could also reflect changes in the population tested. If a school’s population changed so now 50% of all students tested are English language learners, its scores would go down even though the instructional quality may be good. The test scores do not have a direct relationship to how well the teachers teach.
My daughter gets high scores on her schools’ standardized test. That means she is learning.
Not necessarily. It just means your daughter performed better than most of a national sample of students in her grade. Standardized tests typically only test grade-level content knowledge. If your daughter is capable of 11th-grade math in 6th grade, but she’s only taught 6th-grade math, a high score means she’s not being challenged. And this isn’t necessarily a good thing.
But our school gives a test that will adapt the questions based on how a student responds. This way, I know that the school will identify and teach to my son’s abilities.
There is a limit as to how far beyond grade level these tests will go. But, let’s say there isn’t, and the test determined that your son in 2nd grade was capable of reading at a 7th-grade level. Unless there is a particular advanced program he can attend, chances are his instruction will not be based on 7th-grade reading material. A teacher with a large class of students with highly diverse abilities is not modifying every lesson. (And if they are, I’ll show you a teacher that doesn’t sleep, eat, or have anything that looks like a weekend.) The test score is only as good as a school’s ability to change curriculum and instruction. And this kind of change takes more than a mandated test.
The only way we can keep schools and teachers accountable for results is by testing students.
The best way to keep schools and teachers accountable is through strong community relationships. If the local families and community members are involved in transparent school policies, they will hold schools responsible. High-stakes testing takes decisions away from local communities and imposes top-down regulations, thereby creating distance and mistrust between locals and neighborhood schools.
High-stakes testing improves services to special education and English language learners.
Let’s be honest. A test itself doesn’t DO anything. For educational equity to occur, things need to change in the school. And for things to change, you need money. Highly qualified special education and ELL teachers are hard to find in many districts. If we pay teachers more, we can get more qualified teachers. If we reduce class sizes, students with significant challenges will get more individualized attention. And if we build more schools, we can reduce class sizes. Requiring SPED and ELL students to take standardized tests doesn’t automatically make their schooling more equitable.
But without testing data, how can teachers know what they need to teach?
Teachers collect data continuously. They give quizzes, tests, homework assignments, grades, and rubric scores. They assess student conferences, science fairs, class performances, student portfolios, and group projects. They observe, reflect, ask questions, and take notes on student performance and behavior. They discuss student achievement with parents, other teachers, administrators, and students themselves. They do not need a high stakes standardized test to tell them how their students are doing and what they need.
But what about bias? Isn’t a teacher’s opinion subjective?
That’s why students take benchmark exams that measure the content taught in a particular class. Benchmarks may be standardized (in the sense that they are a common test given to all students across the district in that class.) But unlike high-stakes national standardized exams, they are specifically tied to the content covered versus a limited sampling of Common Core material. Benchmark exams allow teachers and students to see the problems they got wrong and use these results to improve learning and instruction.
How can we compare states if we don’t have high-stakes testing?
First of all, education shouldn’t be a competition. By forcing states to compete for federal resources, you ensure that some will lose. Do we really want to be in a country that allows some students’ schools not to be funded because they lost the “race to the top”? Before all the high-stakes frenzy, the US was tracking student achievement on a national scale with NAEP (the National Assessment of Educational Progress.) Educational statisticians will tell you this is the most reliable measure of education over time. One big reason is that it is NOT high stakes; schools cannot prep students for it, and there are no funds tied to it.
High-stakes testing has worked. My child’s school adopted a program designed to raise test scores, and it did.
By focusing instruction on test score growth specifically, the school made the test scores invalid. The result is a score that is more a reflection of test-taking skills than a students’ content knowledge. Let’s say you were going to take a test on Spanish vocabulary. And in order to raise your test score, I drilled you on a selection of words most likely to be on the test using a replication of the type of test you would take. You’d probably get a higher score than if we didn’t practice the test. Unfortunately, after the test, you promptly forget the Spanish I taught you. The high test score doesn’t mean you learned Spanish; it just means you temporarily memorized some words. It also means that by familiarizing you with the type of questions asked, I increased your chances of getting more items correct. But this has nothing to do with your ability to speak Spanish.
If I genuinely wanted to make you a successful Spanish speaker, I would have you engage in authentic, meaningful activities. These types of lessons would not only produce real growth, they would also help you develop a love of Spanish and a desire to continue with it long after my class. Test prep programs do none of these things.
High-stakes testing ensures our kids are prepared for the future world of work.
Over and over, the most innovative organizations have told us that the qualities they look for in outstanding employees are: creativity, self-motivation, the ability to work with diverse groups, and the ability to solve real-world problems. High-stakes tests cover none of these things. In fact, a hyper-focus on test scores have reduced the very things in schools that lead to career-ready students: music, free play, art, foreign languages, field trips, long-term research projects, shop classes, and community internships.
Although information can help combat misunderstandings, we also must address the real concerns people have about education. We must offer feasible alternatives to high-stakes testing; measures that account for student growth, teacher effectiveness, and school quality. Such practices are possible and can lead to gains far beyond anything achieved with our current system of accountability.