ASU expert decries high-stakes testing of students

The debate over high-stakes testing for students was highlighted last month when President Barack Obama said that schools should have fewer, but better quality exams.

In Arizona, starting this year, students are no longer required to pass a standardized test to graduate from high school. But the state’s public schools still are required to test third-graders to prove reading proficiency, and a portion of teachers’ evaluations are tied to test results.

Audrey Amrein-Beardsley, an expert on testing and an associate professor in Mary Lou Fulton Teachers College at Arizona State University, weighs in on the debate about the effectiveness of testing students. She writes a blog about educational policy and testing called VAMboozled!

Question: Do you believe students are being tested too much? Or is the amount of testing appropriate, but the emphasis — “high stakes” — out of whack?

Answer: I definitely believe students are being tested too much, especially given that students across the nation spend over four solid days on tests alone. These estimates do not include the time it takes to prepare for tests, taking away from valuable instruction time. In addition, “high-stakes” testing is not a good way to measure the quality of education in our schools. Research over the past 30 years shows that attaching test scores to academic progress never works, even though conceptually doing so makes sense. A good rule of thumb is that the farther the test gets away from the classroom-level (such as tests designed by entire statewide systems), the less useful the test becomes for its intended informative purposes.

Q: What level of testing do you believe is appropriate for younger students? For high schoolers?

A: I do not believe large-scale standardized tests are appropriate for young children. There are many methodological issues that come into play when testing young children related to their ages and levels of maturity, the complexities that come along with defining and measuring young children’s development, learning and academic productivity, and other ethical concerns. As for high school students, research indicates it is OK to test these students. However, the most effective tests are developed near the school level instead of on a larger scale. These are the tests that are the most closely aligned with what has been taught and are capable of yielding data and information that is actionable. Large-scale standardized tests yield little to no timely or actionable data, especially at the high school level, given that most tests are not developed to measure specialized subject areas.

Q: Testing has increased as taxpayers have demanded more accountability — requiring teacher pay to be tied to test scores, for example. And requiring a third-grade test to determine reading proficiency or the student is held back. If testing is cut back, how will schools provide accountability to taxpayers?

A: Despite policymakers' increased focus on accountability, few taxpayers demand it. National survey data shows that the average taxpayer does not demand “more accountability.” Politicians and policymakers are largely the ones demanding accountability. In fact, there is a widening disparity between policymakers and taxpayers. The growth of the “opt-out” movement, where parents of hundreds of thousands of public school students choose to not take high-stakes standardized testing, highlights this disparity. While parents are concerned with frequent testing, policymakers responded with serious consequences and penalties to force parents and students to take these tests. The question to me is: How will the federal and state governments ensure that the accountability measures and models they put into place, to hold schools accountable to the taxpayers and for which taxpayers also pay millions, are actually working?

Q: The National Assessment of Education Progress scores were recently released. These are often considered to be among the “gold standard” of standardized test scores because of the rigorous and consistent testing methods used. Yet budget reductions have reduced the scope of these tests. Do you think these large-scale, nationwide tests are important? Do you think it’s important, or even relevant, to compare states to each other?

A: Testing experts across the nation respect the NAEP and admire it for its ability to monitor states’ educational programs. The NAEP is also respected because “high-stakes” have not been attached to the test and have not distorted the results. I also think it is important to compare states to one another. However, we do know a lot about these states without such tests and all tests in general, when we know other correlated factors (such as states’ political demographics, teacher-student ratios, average class size, teacher salaries, per-pupil funding). For anyone who thinks that any test-based measure is picking up or assessing only “student learning,” they are seriously wrong as test scores are so highly correlated with these and other variables. While this is unfortunate, it is true.

Q: Is there any research proving any negative effects of testing on students? Does it address the quantity or quality of the tests?

A: There is a lot of research on the negative effects of testing on students. In fact, it’s more worrisome that our nation continues to perpetuate such test-based accountability approaches to educational reform. Tests, while still extremely expensive, are still the cheapest educational reform measure to adopt and implement.

Q: Is testing actually a better reflection on a student’s socioeconomic status than their subject proficiency?

A: Unfortunately, this is also true. With a handful of variables about any student population, we can predict with about 80 percent accuracy what students’ test scores would be without students actually taking the tests.

Q: Some states, such as New York, require end-of-course tests for high school graduation. What are your thoughts on those tests? How about SAT and ACT scores? Some universities are moving away from requiring them. Is that a good idea?

A: End-of-course tests are OK because they are more instructionally sensitive and better assess a student’s learning outcomes. However, if we add end-of-course tests, other tests need to give way. In addition, the SAT and the ACT do not predict student success in college as well as most people think. Common stats indicate that approximately 50 percent of the students who perform well on either test will perform well in college and the other 50 percent will not. I think that universities should get rid of the current college entrance exams and replace them with more quality — albeit likely less efficient — indicators for making college entrance decisions.