With its beguilingly simple logic, President Bush’s plan to test every third- through eighth-grader in reading and math every school year is attracting backers on both sides of the congressional aisle. Legislation including the Bush testing plan recently breezed though a Senate committee by a vote of 20-0.
It’s hard to argue with Mr. Bush’s notion that “states, school districts and schools must be accountable for ensuring that all students … meet high academic standards.” But the Bush testing plan could be a disaster for students and advocates of accountability alike—a disaster if states end up merely doing more of the sort of testing that many of them are already doing.
Mr. Bush says that more testing would increase accountability by casting a bright light on student performance. But nearly half the states that already test students every year in reading and math use tests that don’t measure students’ performance against standards. They only gauge how students’ knowledge of subjects stacks up against that of a national sample of students.
The national sample could know a lot of math or hardly any. The tests, including the Iowa Test of Basic Skills and the Stanford 9 Achievement Test, don’t tell us one way or the other. They don’t deliver on the basic pledge of the president’s testing plan.
States that link their exams to standards tend to test the lowest-level skills reflected in their standards, say accountability advocates such as Robert Schwartz, president of Cambridge, Mass.-based Achieve Inc., a school-reform organization set up by corporate leaders and governors.
The result: Instead of pushing teachers to teach to higher standards, many of the accountability-driven tests used today lower the level of instruction in many classrooms.
Rather than reading literature, students are increasingly reading short, disconnected passages of the sort included on many state reading tests; rather than writing essays, they spend classroom time finding mistakes in short writing samples. The teaching of concepts and problem solving gives way to rote memorization of information that’s often quickly forgotten.
University of Colorado Professor Lorrie Shepard retested students several weeks after they had taken today’s widely used standardized tests, only to find that students’ scores declined sharply when she asked questions slightly differently on the second test. Why? Students hadn’t learned the concepts being tested.
What’s more, valuable subjects that aren’t tested, such as art and music, are routinely pushed out of the school day altogether in favor of low-level test preparation.
Ironically, such dumbing-down of instruction in the name of accountability is most prevalent in schools with large populations of poor and minority students—exactly the students that accountability systems built on high quality assessments would help the most.
Stronger testing systems are emerging. Maryland is among a handful of states that have crafted tests that are tightly linked to state standards and that include writing tasks and other “performance” measures that push students beyond low-level skills.
But such tests are more expensive to create and to score. Maryland spends upward of $30 per student a year on its performance-based tests, while the Stanford 9 and other off-the-shelf exams that rely heavily on multiple-choice questions cost several states about $6 per student.
Yet $30 per student is only a fraction of the roughly $6,500 spent annually to educate each American public school student.
Given Mr. Bush’s plan to make standardized testing a priority of every public school district, states should invest far more in high-quality exams.