The next president has a unique opportunity to start from scratch in education policy, without the deadweight of a failed, inherited No Child Left Behind (NCLB) law. The new president and Congress can recapture the "small d" democratic mantle by restoring local control of education, while initiating policies for which the federal government is uniquely suited -- providing better achievement data and equalizing the states' fiscal capacity to provide for all children.
This opportunity exists because NCLB is dead. It will not be reauthorized -- not this year, not ever. The coalition that promoted the 2001 bipartisan law has hopelessly splintered, although NCLB's advocates in the administration and the Congress continue to imagine (at least publicly) that tinkering can put it back together.
NCLB, requiring annual reading and math tests in grades 3 through 8 (and one such test in high school), represents an unprecedented federal takeover of education. It punishes schools not making "adequate yearly progress" toward having all students proficient at "challenging" standards by 2014, regardless of students' socioeconomic disadvantages or even of their cognitive disabilities.
Many Republicans supported NCLB out of loyalty to President Bush and because Karl Rove assured them that their vow to improve minority achievement would entice African Americans away from the Democrats. But now, with Democratic congressional majorities and a possible presidency, Republicans have rediscovered their belief in local control of education. Few now support reauthorization.
Many Democrats were equally cynical in supporting NCLB. Some believed a law demanding unrealistic achievement targets would justify big boosts in federal spending when targets proved unattainable. Others, arguing that low minority scores result mainly from poor teaching ("low expectations"), expected that federal demands for higher achievement would whip teachers into shape, even if the mandated goals were fanciful.
What few Democrats understood, however, was that test-based accountability might spur teachers but would also corrupt schooling in ways that overshadowed any possible score increases. Excessive testing is now so unpopular that Congress' newly elected Democrats campaigned in 2006 against NCLB and now won't support reauthorization. Senior Democrats are also hearing from parents, teachers, school boards, and state legislators.
Sen. Edward Kennedy of Massachusetts and Rep. George Miller of California, who sponsored the original legislation, promise colleagues that they can fix NCLB. But no fixes are possible. Weakening rigid testing requirements provokes denunciation from President Bush and Secretary of Education Margaret Spellings, who unabashedly calls the law "99.9 percent pure."
But NCLB was flawed from the start. The 2001–2002 stampede ignored well-established statistical and management theories predicting perverse consequences for test-based accountability.
GOAL DISTORTION
One such consequence is goal distortion, the subject of extensive warnings in the economics and management literature about measuring any institution's performance by quantitative indicators that reflect only some institutional goals. Management expert W. Edwards Deming urged businesses to "eliminate management by numbers, numerical goals" because they encourage short-, not long-term vision. Peter Drucker gave similar advice. Today, management consultants urge "balanced scorecards" using qualitative judgment, as well as financial indicators, to measure corporate success.
Schools have many goals for students: basic math and reading skills but also critical thinking, citizenship, physical- and emotional-health habits, arts appreciation, self-discipline, responsibility, and conflict resolution. Schools threatened with sanctions for failure in only one goal will inevitably divert attention from others. One NCLB consequence has been less social studies, science, art, music, and physical education -- particularly for low-income children, whose math and reading scores are lowest and for whose teachers the consequences of spending time on, say, history, rather than more math drill, are most severe.
Goal distortion has been particularly troubling, as it should be, to conservatives. Two former assistant secretaries of education (under Ronald Reagan and Bush père), Chester Finn and Diane Ravitch, once prominent NCLB advocates, now write:
We should have seen this coming ... more emphasis on some things would inevitably mean less attention to others. ... We were wrong.
They conclude:
[If NCLB continues,] rich kids will study philosophy and art, music and history, while their poor peers fill in bubbles on test sheets. The lucky few will spawn the next generation of tycoons, political leaders, inventors, authors, artists and entrepreneurs. The less lucky masses will see narrower opportunities.
TEST RELIABILITY
NCLB relies on an annual test, but single tests can be misleading. Every parent knows children have good and bad days. Every teacher knows particular classes can be talented or difficult. Entire classes can be attentive or distracted. So accurate measurement requires multiple retesting. Most schools are too small for statistical confidence that children's good and bad days will average out on one test. Because a school's subgroups (blacks, Hispanics, or low-income children) are smaller than a full-grade cohort, the margin of error for subgroup achievement is even larger. The more integrated a school, with more subgroups, the more inaccurate accountability becomes.
When the Bush administration and Congress were designing NCLB, two economists (Thomas Kane and Douglas Staiger) demonstrated that many schools would be rewarded or punished solely because of these statistical challenges. Their paper derailed NCLB for six months while administration and congressional experts tried to finesse the problems. They couldn't but enacted NCLB anyway, which engendered remarkable anomalies: schools rewarded one year and punished the next despite identical levels of effectiveness; schools rewarded under a state's system and simultaneously punished under the federal one, or vice versa. Some states dodge these absurdities by reporting large error margins with test scores, but this hides poorly performing as well as misidentified schools, and draws the wrath of accountability enthusiasts.
THE PROFICIENCY MYTH
Even with inordinate attention to math and reading, it is practically and conceptually ludicrous to expect all students to be proficient at challenging levels. Even if we eliminated all disparities based on socioeconomic status, human variability prevents a single standard from challenging all. The normal I.Q. range, 85 to 115, includes about two-thirds of the population. "Challenging" achievement for those at 115 would be impossibly hard for those at 85, and "challenging" achievement for those at 85 would be too easy for those at 115.
The law strongly implies that "challenging" standards are those of the National Assessment of Educational Progress (NAEP), periodic federal tests of national student samples. But while NAEP tests are excellent, their proficiency cut-points have no credibility. Passing scores are arbitrary, fancifully defined by panels of teachers, politicians, and laypeople. Many children in the highest-scoring countries don't achieve them. Taiwan is tops in math, but 40 to 60 percent of Taiwanese students are below proficient by NAEP standards. Swedish students are the best readers in the world, but two-thirds are not NAEP-proficient.
Expecting all Americans to perform at this level can only set them, their teachers, and schools up for failure. (Actually, this charge is slightly exaggerated: NCLB exempts the most severely disabled, requiring only U.S. children with I.Q.s above 65 to be as proficient as the top half of the Taiwanese.)
In a rare bow to local control, NCLB doesn't enact NAEP's proficiency definitions but permits states to invent their own. Not surprisingly, some define proficiency far below "challenging" expectations, and the Department of Education has little choice but to let this pass; if it enforces high standards, the already unacceptably large number of failing schools would be astronomical. But low state passing points are a sore spot for NCLB advocates, who propose to correct this with high national standards. Their demand makes reauthorization even less probable.
THE BUBBLE KIDS
Any single proficiency standard invites sabotaging the goal of teaching all children, because the only ones who matter are those with scores just below passing. Educators call them "bubble kids" a term from poker and basketball, where bubble players or teams are those just on the cusp of elimination. Explicit school policies now demand that teachers ignore already-proficient children to focus only on bubble kids, because inching the bubbles past the standard is all that matters for "adequate yearly progress."
Less obvious are incentives also to ignore children far below proficiency, whom even constant drilling won't pull across the finish line. Because all must eventually (by 2014) pass, ignoring poorer performers should, in the long run, be counterproductive. But NCLB places no premium on the long run. Educators can't worry about possible distant punishment. And since most consider the 2014 goal absurd, they have good reason to expect it to be abandoned, further reducing incentives to worry about the lowest achievers. What's more, the higher the standard, the more children there are who are too far below proficiency to worry about. So the law guarantees that more disadvantaged children will be left further behind, especially in states with higher standards.
For bubble kids, schools have substituted test prep for good instruction. With test development costly, states use similar tests repeatedly, guiding teachers to stress content they suspect will reappear. Teachers impart test-taking skills (like how to guess multiple-choice answers) that don't deepen understanding of math and reading. In the weeks before testing, schools step up drilling; this does little to help children retain what they learned. Thus, student scores on state tests are not duplicated in NAEP, which is less subject to test-prep corruption. Administered only to representative samples of schools and students, with more emphasis on critical thinking, NAEP scores have not shot up along with state test results. NAEP math scores have increased a little, but at about the same rate as before NCLB's adoption -- suggesting that, for all its other problems, NCLB has also been an utter waste of time.
SCHOOLS AND SOCIAL POLICY
In one respect, NCLB betrays core Democratic principles, denying the importance of all social policy but school reform. Inadequate schools are only one reason disadvantaged children perform poorly. They come to school under stress from high-crime neighborhoods and economically insecure households. Their low-cost day-care tends to park them before televisions, rather than provide opportunities for developmentally appropriate play. They switch schools more often because of inadequate housing and rents rising faster than parents' wages. They have greater health problems, some (like lead poisoning or iron-deficiency anemia) directly depressing cognitive ability, and some causing more absenteeism or inattentiveness. Their households include fewer college-educated adults to provide rich intellectual environments, and their parents are less likely to expect academic success. Nearly 15 percent of the black-white test-score gap can be traced to differences in housing mobility, and 25 percent to differences in child- and maternal-health.
Yet NCLB insists that school improvement alone can raise all children to high proficiency. The law anticipates that with higher expectations, better teachers, improved curriculum, and more testing, all youths will attain full academic competence, poised for college and professional success. Natural human variability would still distinguish children, but these distinctions would have nothing to do with family disadvantage. Then there really would be no reason for progressive housing or health and economic policies. The nation's social and economic problems would take care of themselves, by the next generation.
Teachers of children who come to school hungry, scared, abused, or ill, consider this absurd. But NCLB's aura intimidates educators from acknowledging the obvious. Teachers are expected to repeat the mantra "all children can learn," a truth carrying the mendacious implication that the level to which children learn has nothing to do with their starting points. Teachers are warned that any mention of children's socioeconomic disadvantages only "makes excuses" for teachers' own poor performance.
Of course, there are better and worse schools and better and worse teachers. Of course, some disadvantaged children excel more than others. But NCLB has turned these obvious truths into the fantasy that teachers can wipe out socioeconomic differences among children simply by trying harder.
Denouncing schools as the chief cause of American inequality -- in academic achievement, thus in the labor market, and thus in life generally -- stimulates cynicism among teachers who are expected to act on a theory they know to be false. Many dedicated and talented teachers are abandoning education; they may have achieved exceptional results with disadvantaged children, but with NCLB's bar set so impossibly high, even these are labeled failures.
The continuation of NCLB's rhetoric will also erode support for public education. Educators publicly vow they can eliminate achievement gaps, but they will inevitably fall short. The reasonable conclusion can only be that public education is hopelessly incompetent.
"FIXING" NCLB
Few policy-makers have publicly acknowledged NCLB's demise. Instead, they talk of fixing it. Some want to credit schools for student growth from year to year, rather than for reaching arbitrary proficiency levels. Clearly, adequate progress from different starting points leads to different ending points, but growth-model advocates can't bring themselves to drop the universal-proficiency goal. Doing so would imply lower expectations, on average, for disadvantaged children -- too much for unsophisticated policy discussion to swallow. Consequently, the "fix" is incoherent.
Growth models have even larger error margins than single-year test results because they rely on two unreliable scores (last year's and this year's), not one. And accountability for math and reading growth retains the incentives to abandon non-tested subjects and skills. So some NCLB loyalists now propose accountability for "multiple measures," such as graduation rates. But presently quantifiable skills are too few to minimize goal distortion -- the federal government is unprepared to monitor, for instance, whether students express good citizenship. Further, any mention of diluting a math and reading focus elicits the wrath of "basics" fundamentalists, such as the president and his secretary of education.
Although NCLB will not be reauthorized, the underlying Elementary and Secondary Education Act (ESEA), with funding for schools serving low-income children, will continue. NCLB will remain on the books, increasingly ignored. Virtually every school with minority, low-income, or immigrant children will be labeled a failure; the federal government will be hard-pressed to punish all. Eventually, under a new administration, ESEA will be renewed, perhaps including vague incantations that states establish their own accountability policies, once Washington abandons the field.
States will do so. Some, not having learned NCLB's lessons, will retain the distortions and corruption that NCLB established. Others, more creative, will use qualitative as well as quantitative standards, relying on school inspections as well as test scores.
Renouncing federal micromanagement will require liberals to abandon a cherished myth: that only the federal government can protect disadvantaged minorities from Southern states' indifference. The myth is rooted in an isolated fact: In the two decades following Brown v. Board of Education, the federal government forced states to respect rights not only of African Americans but of disabled and immigrant children.
But at other times, the federal government has been no defender of the oppressed. In the early 20th century, state governments enacted minimum-wage, health, and safety laws, only to see them struck down by the Supreme Court. Today, Southern states' attempts to improve education are often impeded by federal policy. Only last year, school integration efforts of Louisville, Kentucky, were prohibited by federal courts, while federal administrative agencies block efforts at integration and affirmative action. In recent decades, states like North Carolina and Texas have been innovators in school improvement. North and South Carolina and Arkansas have had nationally known "education governors" (Jim Hunt, Richard Riley, and Bill Clinton). The greatest potential for greater education improvement in the South lies in boosting African American voting participation, not more federal mandates.
WHAT THE NEXT PRESIDENT CAN DO
With the federal government proven incapable of micromanaging the nation's 100,000 schools, what education roles remain for a new administration? There are two.
One is to provide information about student performance, not for accountability but to guide state policy. NAEP should be improved. Now given regularly at the state level only in math and reading, such coverage should expand to include history, civics, and the sciences, as well as art, music, and physical education. For example, NAEP could provide state-by-state data on physical education by sampling students' body mass index numbers and upper body strength, characteristics for which standardized tools are available.
When NAEP was first designed in the 1960s, it included important elements that were soon abandoned under cost pressures. While employing paper-and-pencil tests, early NAEP also dispatched field assessors to observe, for example, how young children solved problems in cooperative groups. NAEP assessed representative samples of adolescents, whether in or out of school, as well as of young adults in their mid-20s. Assessors tracked down 17-year-olds and young adults, administering tests to determine if their schooling had lasting impact.
A dramatic expansion of NAEP, covering multiple skills and out-of-school samples, with state-level reporting, would be expensive, multiplying by several times the current NAEP budget of $90 million. But this would only slightly increase the roughly $45 billion in federal funds now supplementing state and local school spending. Provision of state-by-state data on a balanced set of outcomes should be a federal responsibility.
The other new federal role should be fiscal equalization. New Jersey now spends about $14,000 per pupil, more than twice what Mississippi spends. Adjusting for the dollar's purchasing power still leaves New Jersey spending 65 percent more than Mississippi.
This cannot be attributed to New Jersey caring more about children than Mississippi. New Jersey's fiscal capacity, its per capita personal income, is over 70 percent greater than Mississippi's. And Mississippi's needs are greater: 10 percent of New Jersey's children live in poverty, compared to Mississippi's 29 percent. Again, after adjusting for the value of the dollar, Mississippi still faces greater educational challenges, with less ability to meet them.
Washington now exacerbates these inequalities. Federal school aid -- ESEA aid to districts serving poor children -- is proportional to states' own spending. So New Jersey, which needs less aid, gets more aid per poor pupil than Mississippi, which needs more.
It will be politically tough for a Democratic Congress and administration to fix this, because sensible redistribution, with aid given to states in proportion to need, and in inverse proportion to capacity, will take tax revenues from states like New Jersey (which sends liberal Democrats to Congress), and direct them to states like Mississippi (which sends conservative Republicans). Funding equalization requires political courage not typically found in either Washington party. There's a role here for presidential leadership.
Narrowing huge fiscal disparities will take time. Whether the next Democratic Congress and administration -- if they are Democratic -- take the first steps will test whether the party is truly committed to leaving no child behind.
Abandoning federal micromanagement of education has a hidden benefit: helping to reinvigorate American democracy in an age of increasingly anomic and media-driven politics. Local school boards in the nation's nearly 15,000 school districts (but not in the biggest cities) can still provide an opportunity for meaningful citizen participation. Debating and deciding the goals of education for a community's children is a unique American privilege and responsibility. Restoring it is a mission worthy of a new administration.