By Jay Mathews

David Levin worked hard to improve as a teacher. By his second year at Bastian Elementary School in Houston he had made progress. His fourth graders were learning quickly, fortified by a rich collection of games, chants and songs Levin borrowed from Harriett Ball, a magical veteran teacher who worked across the hall.

Levin, a tall, curly-haired 24-year-old New Yorker, did so well he was, with encouragement from Ball, elected as Teacher of the Year 1994 by the Bastian faculty. That might have left him with a warm glow, except that shortly after his triumph he was fired, an experience he can still recount in great detail 15 years later, the emotions coming back.

This year, Levin and his friend Mike Feinberg are close to becoming the most famous teachers in the country. They have founded the nation’s most successful network of public charter schools, the Knowledge Is Power Program (KIPP). Hundreds of groups have asked them to speak. My book about them, Work Hard. Be Nice: How Two Inspired Teachers Created the Most Promising Schools in America, (Algonquin, 2009) has just come out, and awards for them are pouring in.

Yet getting their schools, and their students, to the highest possible levels of learning remains as difficult now as it was when they started. As I researched how KIPP began, it became clear that the day Levin was fired was not an anomaly, not an exception to their great success, but a typical moment in the evolution of powerful educational ideas. Creating great schools in inner city and rural America, raising the achievement of long-neglected students, can be done, Levin and Feinberg have proven that, but not without regular setbacks. Many traditional educators still believe that KIPP is just another anti-public school plot by troublemakers who don’t understand their students’ fragilities, and don’t see that pushing kids too hard can be a terrible thing.

The events that led a Houston principal to fire his school’s teacher of the year unfolded gradually. Levin did not expect that his answers to the principal’s questions about 11 low-performing students in his class would create such indignation.

In Texas at that time the rule was that if fewer than 75 percent of students in each ethnic group passed the state achievement test in any school, that school would be rated unsatisfactory. The system posed a problem for Levin’s principal because the young teacher’s class had relatively few white and Latino students. The African-American children who were the bulk of his students were doing fine, and would pass the 75 percent mark easily. But there was little margin for error if all of the non-black students did not pass. Levin was told that eight of the ten Latino students, plus the white student, had to pass or the school-and the principal-would look bad.

There was a way out, the principal said. If the students’ teacher or parents signed a statement saying their children’s language skills or learning disabilities meant they were not ready for the test, they did not have to take it. Levin, however, refused to sign. He thought the students were improving, and could pass. When the principal asked the parents to sign, they also refused. That explained that that nice Mr. Levin, who was so polite to stop by their homes and keep them in touch with their children’s progress, had told them not to sign.

Levin proved to be right. All but one of the students passed the math section of the test and all but two passed the reading section. The school would not suffer in the rating. But the test results did not come back until weeks after school closed for the summer. Even if the principal had known the outcome, his rage at the gall of his teacher might not have abated. A week before summer vacation he stopped by Levin’s room and handed him a letter saying he was being terminated. The document did not say why, but Levin checked with headquarters and was told the reason: insubordination.

Fortunately, he had already agreed to transfer from Bastian to the elementary school where his friend and KIPP co-founder Feinberg taught. That school had agreed to let the two young men start their program with 45 children in two classrooms, which has now grown into what this summer will be 18,000 students in 84 schools in 20 states and the District of Columbia.

Writing the book, chronicling repeated battles between the two teachers and assorted administrators refusing to help them, and in some cases trying to stop them, I realized that the story of KIPP was one endless case of insubordination.

What they were doing was nothing else than unleashing the power of great teaching. They had high expectations for all of their students in the fifth through eighth grade public charter middle schools that comprise the bulk of the KIPP network. All of those students, they announced, were going to college. They provided much more time for learning-a 9 hour school day, 4 hours every other Saturday and a required three-week summer school. They took tests seriously. They needed to be sure their students were learning. They made sure they-and the principals they selected to start the schools that followed theirs-had the power to hire the best instructors, and fire them quickly if they did not reach KIPP’s high expectations.

Both Feinberg’s first school in Houston and Levin’s first school in the south Bronx became the highest-performing schools in their areas. The KIPP schools that followed them also produced results at or near the top of their local rankings. Experts skeptical of the test score results were moved by the quality and imagination of the teaching when they visited the schools to check.

The overall KIPP national results can be summed up in a statistic that I too, having covered urban education for 27 years, had difficulty believing. The first 1,800 students to have completed four years at a KIPP middle school have, on average, risen from the 31st to the 58th percentile in reading and from the 41st to the 80th percentile in math. Gains that great for that many low-income children in one program have never happened before. There will be more studies of KIPP, the most studied charter school network in the country, but we have enough data to be confident that those results are real.

And yet many educators, like Levin’s first principal, are not that happy about it, and I don’t blame them. Levin and Feinberg, and the several other teachers leading other organizations whose results approach theirs, have shown that the children in our poorest neighborhoods have the potential to do as well as the children in our richest neighborhoods, if given the extra time and encouragement, and the good teaching, they need to learn.

It is a hard lesson for some people who think gaming the system, excusing students from the toughest exams, is the best way to operate. But for the rest of us, it is welcome news, a sign that we should be hiring, not firing, the people can do that.

Mathews is the Washington Post education columnist. His New York Times bestselling book is Work Hard. Be Nice: How Two Inspired Teachers Created the Most Promising Schools in America (Algonquin, 2009).