No Child Left Behind and the Testing Era It Started

On January 8, 2002, a president from Texas stood in a high school gymnasium in Hamilton, Ohio, a manufacturing town north of Cincinnati, and signed a 600-plus-page education bill with a senator from Massachusetts standing close enough to share the photograph. George W. Bush had picked Hamilton because it was the home district of John Boehner, the Republican who chaired the House education committee. Ted Kennedy was there because he had spent most of 2001 negotiating the thing line by line. George Miller, the California Democrat who had been Boehner's opposite number, was there too, and so was Judd Gregg, the New Hampshire Republican. Four months after September 11, the room was full of people who badly wanted to show the country that Washington could still agree on something. What they agreed on was a test, given every year, to almost every child in grades three through eight, and a promise about what those tests would eventually show.

In short: No Child Left Behind made annual reading and math tests a federal requirement and forced schools to report results for poor children, Black and Latino children, English learners and students with disabilities. That reporting exposed gaps many districts had buried in averages. The law's impossible deadline and escalating penalties also narrowed what many schools taught, and its lessons still shape the fight over what a good school is.

A bipartisan bill with a deadline nobody could meet

No Child Left Behind was not a new program so much as a rewrite of an old one. The Elementary and Secondary Education Act of 1965, a piece of Lyndon Johnson's war on poverty, created Title I, the federal money that flows to schools serving poor children. For decades Title I sent dollars with few questions about results. The 2001 rewrite flipped that bargain. Keep taking the money, it said, and prove the children are learning.

The votes were lopsided in a way that now looks like a photograph from another planet: the House passed the final bill 381 to 41 and the Senate 87 to 10. Liberals and conservatives signed on for different reasons. Conservatives liked accountability for public spending and the idea that parents in failing schools should be able to leave. Many civil rights groups liked that a school could no longer hide a failing group of children behind a respectable average. Bush called the old habit "the soft bigotry of low expectations," and plenty of people on the left thought he had a point.

The core requirements were simple to state:

• Test every student in reading and math every year in grades three through eight, and once in high school. Science tests in certain grades came later.
• Report scores not just for the whole school but for subgroups: low-income students, major racial and ethnic groups, students with disabilities and English learners.
• Show "adequate yearly progress" toward a goal that every student would be proficient by the 2013-14 school year.
• Apply escalating consequences to Title I schools that missed the target, from letting families transfer, to paying for outside tutoring, to replacing staff or restructuring the school.

That third bullet is where the trouble lived. The law let each state define "proficient" and pick its own test, but it fixed the destination: one hundred percent. A school with a small group of English learners who missed the mark in a single subject could fail for the year. As the deadline approached, the share of schools labelled as failing climbed, not always because they had gotten worse, but because the target was rising faster than any school system in history had improved.

What the subgroup rule revealed

Before 2002 it was common for a suburban district to report strong averages while its few dozen Black students or its students in special education were being quietly failed. Nobody had to look. The subgroup rule made them look.

This is the part of the law that its harshest critics on the left often skip, and it deserves to be said plainly: disaggregated data is a civil rights tool. It turned "our schools are fine" into a question with a measurable answer. Parent organizers in many cities used those numbers to demand resources for schools that had been starved. Disability advocates used them to argue that students with individualized plans belonged in the same accountability system as everyone else, not off to the side. The habit of asking "fine for whom?" outlived the law that created it, and survives in its successor.

There is also some evidence that achievement rose. Economists Thomas Dee and Brian Jacob, comparing states that already had their own accountability systems with those that did not, found that the federal law was followed by meaningful gains in fourth-grade math, especially for lower-achieving and disadvantaged students, with little sign of gains in reading. Scores on the National Assessment of Educational Progress, the low-stakes national test that no school can teach to, rose in math for younger students during the 2000s, though that rise had begun before the law took effect. Honest readers of the evidence end up in roughly the same place: modest real gains in some subjects and grades, nowhere near the promise in the title.

The day got narrower

Here is what a test can't see: whatever isn't on it.

Within a few years, surveys of school districts by the Center on Education Policy found that many had increased time for reading and math and cut time from social studies, science, art, music and recess, especially in elementary schools. In schools labelled as failing, the pressure was sharpest. Teachers described months of test-preparation booklets. Districts bought interim tests to predict the real test. Some schools focused on "bubble kids," the students just below the proficiency line, because moving them over the line changed the school's rating, while a child far below or comfortably above did not.

Social scientists have a name for this pattern. Donald Campbell, a psychologist, observed decades earlier that the more a single number is used to make decisions, the more it is corrupted and the more it distorts the process it was meant to measure. No Child Left Behind turned that observation into a natural experiment run on tens of millions of children.

At its worst, the distortion became fraud. In Atlanta, a state investigation found widespread erasing and changing of answers on state tests; educators were indicted in 2013, and in 2015 a jury convicted most of those who went to trial, several under the state's racketeering law. The Atlanta case is not proof that testing makes people cheat. It is proof that when careers and school closures ride on one number, the institution will be tempted to manage the number rather than the learning.

The exits that mostly went unused

The law's designers did not imagine that labels alone would fix anything. They built in two escape hatches for families, and both are worth a look because they foreshadowed arguments that came later.

The first was public school choice. Once a Title I school missed its targets two years running, the district had to offer its students a transfer to a better-rated public school, with transportation. On paper this was a big deal. In practice, take-up was small. In many districts the "better" schools were full, far away or labelled as failing themselves by the next year. Parents who liked their neighborhood school, its teachers and the friends their children had made there, did not want to leave. A rating is an abstraction; the crossing guard who knows your kid by name is not.

The second was "supplemental educational services," free tutoring paid for out of the district's Title I money and often delivered by private companies. Here, too, enrollment ran well below eligibility, and studies of the tutoring's effects were mixed. Some providers were good. Others signed up children with gift cards and delivered little. States struggled to monitor quality.

The lesson is not that choice or tutoring are bad ideas. Intensive tutoring done well, in school, by trained people, is one of the better-supported interventions in education research. The lesson is that an exit offered to families without capacity, information or quality control does not do much. That finding is worth remembering whenever vouchers and tax credits are sold, as they often are now, as the cure for a failing school.

The strongest case for the testing era

It would be easy to write the history of this law as a cautionary tale and stop. The other side's case is better than that, and it should be heard in full.

Supporters argue that before 2002, accountability for public schools was mostly a story adults told themselves. Graduation rates were calculated in ways that flattered districts. Report cards were vague. Poor children and children of color were sorted into lower tracks and nobody was responsible. The law forced every state to build a data system, publish results and admit which schools were not working. The narrowing of the curriculum, they say, was a choice made by districts, and the cure is better tests and smarter accountability, not no accountability at all.

Some of that is right. The failure was less the decision to measure and more the decision to use one blunt measure to punish, on a timetable that ignored how slowly institutions change, while the money promised to help schools improve never matched the demands. Many Democrats who voted for the law later said the funding authorized for Title I in the bill was not fully appropriated in the years that followed. Measurement without investment is an audit, not a reform.

And progressives should own a hard truth of their own. For years the left's critique of testing ran together two different things: the overuse of high-stakes tests, which was a real problem, and the reporting of results by race and income, which was a real gain. Parents in the neighborhoods the law was meant to help were often the ones who most wanted the numbers kept. When the opt-out movement of the mid-2010s grew strongest in comfortable suburbs, some civil rights groups warned that throwing out the tests would also throw out the evidence of who was being left behind. Both things were true at once.

Waivers, and the law that replaced it

By 2011 the arithmetic had caught up with the statute. With the 2014 deadline approaching and Congress unable to agree on a rewrite, the Obama administration's education secretary, Arne Duncan, began granting states waivers from the law's core requirements in exchange for adopting the administration's priorities, including college- and career-ready standards and teacher evaluations tied partly to student test results. Most states took the deal.

That fix created its own backlash. Teacher evaluation systems built on test-score growth angered unions, and the federal push behind shared academic standards gave conservatives a target. When Congress finally replaced the law in December 2015 with the Every Student Succeeds Act, the new statute kept the annual tests and the subgroup reporting but handed the design of accountability back to the states and barred the federal government from dictating standards or teacher evaluations. The testing calendar Bush signed in Hamilton is still in place. The national deadline and the federal penalties are gone.

What the testing era should teach this generation

The question the law asked was a fair one: how will we know whether the children with the least are being served? Its answer, a single annual score attached to punishments, was too thin for the job.

A better answer looks less like a scoreboard and more like a check-up. Keep the annual data broken out by group, because the alternative is going back to not knowing. Pair it with measures a school can be held to without gutting the day: whether students have a science lab and an art teacher, whether chronic absence is falling, whether ninth graders are on track, whether the school has a counselor. Spend money where the numbers show need. Treat a low score as a reason to send help, not a reason to fire the staff of a school that serves the poorest neighborhood in town.

None of that is a partisan program. Republican governors and Democratic legislators have both built state accountability systems along these lines since 2015, and some of the most thoughtful critiques of test-driven schooling have come from conservative education writers worried about civics and history disappearing from the elementary day. The Hamilton gymnasium photograph is worth keeping in mind for that reason. The coalition that wrote the law was real, and so is the possibility of a coalition that repairs what it got wrong.

What you can do

The testing system your child sits through each spring is mostly decided in your state capitol now, which means it can be changed by people who show up there.

• Ask your state legislator what the state's accountability system measures besides test scores, and whether chronic absence or access to arts and science counts. You can find who represents you, with phone numbers and their latest votes, at the representatives lookup.
• Read your school's state report card with an eye on the subgroup tables, not just the headline grade. If a group of students is being missed, bring that page to a school board meeting.
• Follow what is happening to federal school funding and the agencies that collect this data on the education topic page, where reporting from across the web is gathered in one place.
• Look for local parent and education groups working on school funding in your state through the organizations directory, which lists nonprofits by cause in every state.
• School board seats are on the ballot in many places in odd-numbered years. Check your registration and your state's deadlines on the voting guide before the next one comes around.