We must rethink testing to encourage real learning.
The “No Child Left Behind” Act of 2001, an enormously important federal law intended to improve America’s public schools, may actually produce the opposite effect. Such an unforeseen consequence stems from the law’s requirement that substantially larger numbers of students must be judged proficient each year according to their performances on statewide achievement tests. If a school’s teachers don’t get sufficient numbers of students to meet each year’s “adequate yearly progress” toward test-determined proficiency, the school is considered to be low-performing. And if a school remains low-performing for more than a year, a series of heavy-duty penalties are dished out, such penalties leading to serious staff shake-ups or, ultimately, to outright closure of the school.
Most states, based on existing test data, anticipate that the bulk of their schools will fail to meet the new federally decreed proficiency targets. Accordingly, some states have simply lowered their definitions of proficiency in an effort to avoid the prospect that most of their educators will soon be seen as ineffectual. (“States Revise the Meaning of ‘Proficient,’” Oct. 9, 2002.) But the lowering of educational expectations for our nation’s students was surely not what the architects of the No Child Left Behind Act intended.
On first reading, this federal legislation appears to leave many key decisions to the states. For example, although the law requires that statewide achievement tests must assess students’ mastery of a state’s “challenging” curricular aims, each state can decide which achievement tests it will use. And states are even allowed to set up their own year-by-year schedule of schools’ adequate-yearly- progress targets. However, the law also stipulates that, by 2014, all of a state’s students must be proficient or better.
Moreover, whatever schedule of annual proficiency targets a state’s officials establish, that schedule’s annual increases toward the 12-year, total-proficiency goal must, by federal law, be equal. Thus, given the current performance levels of students in many states, and the size of the annual progress increases to be required, the number of test-determined low-performing schools will surely be staggering.
Unfortunately, the statewide achievement tests currently used in the United States do not accurately measure students’ mastery of most states’ frequently sprawling curricular aims. Even worse, most of today’s state-adopted achievement tests are instructionally insensitive; that is, they really can’t detect even first-rate instruction on the part of teachers. Statewide tests are constructed according to a traditional measurement model aimed chiefly at providing comparisons among students, rather than at producing evidence regarding school quality. Such traditionally constructed tests, because they are highly related to students’ socioeconomic status, dominantly measure what students brought to school, not what they were taught there.
As long as traditionally constructed achievement tests are the major measure of educators’ instructional success, state policymakers will be faced with an unsavory choice between (a) seeing the bulk of their state’s educators labeled as incompetent, and (b) lowering state curricular expectations so that most schools appear to be making adequate yearly progress. Neither of these options will benefit children.
There is, however, a test-focused solution strategy that can allow the federal law to attain its intended educational improvements. A state can base its proficiency targets on students’ assessed mastery of a modest number of the state’s most important curricular aims. Achievement tests can be built to measure students’ mastery of such truly pivotal skills as a child’s ability to: (1) read varied kinds of written materials and comprehend their meaning, (2) solve diverse kinds of age- appropriate mathematical problems, or (3) compose a compelling persuasive essay.
The statewide achievement tests currently used in the United States do not accurately measure students' mastery of most states' frequently sprawling curricular aims.
We now know how to build statewide achievement tests so they provide accurate and credible evidence of how many students at a school are actually proficient with respect to their mastery of such powerful curricular aims. But those very same achievement tests, if constructed with their instructional implications clearly in mind, can also be catalysts for improving classroom instruction. If the appropriate kinds of statewide achievement tests are installed, there can be genuine progress toward worthwhile curricular goals, rather than public-relations progress toward trivialized levels of student accomplishment.
If states install instructionally supportive tests to satisfy the accountability requirements of the No Child Left Behind Act, there will be no need to lower a state’s definition of proficiency. However, if state officials rely on the “same-old, same-old” sorts of traditionally constructed achievement tests—tests that give teachers and students no real chance to succeed—we should prepare ourselves for an avalanche of test results reflecting debased student “proficiency.”
W. James Popham is an emeritus professor in the graduate school of education and information studies at the University of California, Los Angeles. In 2001, he chaired a national Commission on Instructionally Supportive Assessment. He can be reached by e-mail at firstname.lastname@example.org.