Diane Ravitch shares a timely lesson: We don't (yet) know how to save schools.

What is new is the thoroughness with which Ravitch debunks heralded gains of student achievement under NCLB. For example, in New York City, where Mayor Michael Bloomberg and schools chancellor Joel Klein have trumpeted substantial rises in student scores, Ravitch reveals the state simply lowered the score required to be judged "proficient" in reading and math. (She notes that scores in New York City went up even as scores on the NAEP began to stagnate.) And in Arne Duncan's Chicago, where he was superintendent of schools before becoming Obama's secretary of education, big jumps in students' scores again proved illusory when compared with the NAEP. Nor did they persist into high school. This has been the pattern throughout the country. In 2007, Texas reported that 85.1 percent of its fourth-graders were proficient in reading, but the NAEP found that only 28.6 were. Tennessee reported 90 percent of its students were proficient, but the NAEP found that only 26.2 percent were. And so on.

The data, as Ravitch says, disappoints on other fronts, too—not least in failing to confirm high hopes for charter schools, whose freedom from union rules was supposed to make them success stories. To the shock of many (including Ravitch), they haven't been. And this isn't just according to researchers sympathetic to labor. A 2003 national study by the Department of Education (under George W. Bush) found that charter schools performed, on average, no better than traditional public schools. (The study was initially suppressed because it hadn't reached the desired conclusions.) Another study by two Stanford economists, the Walton Family and the Michael and Susan Dell Foundation (staunch charter supporters), involved an enormous sample, 70 percent of all charter students. * It found that an astonishing 83 percent of charter schools were either no better or actually worse than traditional public schools serving similar populations. Indeed, the authors concluded that bad charter schools outnumber good ones by a ratio of roughly 2 to 1.

Obviously, some high-visibility success stories exist, such as the chain run by the Knowledge Is Power Program, or KIPP, which I've previously discussed here. But these are the decided exceptions, not the rule. And there's no evidence that a majority of eligible families are taking advantage of charters, good or bad. "While advocates of choice"—again, Ravitch included—"were certain that most families wanted only the chance to escape their neighborhood school, the first five years of NCLB demonstrated the opposite," she writes. In California, for example, less than 1 percent of students in failing schools actually sought a transfer. In Colorado, less than 2 percent did. If all this seems a little counterintuitive, Ravitch would be the first to agree. That's why she supported charters in the first place. But the evidence in their favor, she insists, simply hasn't materialized.

Ravitch isn't arguing for pessimism. She's arguing for humility—the kind she's had to learn the hard way—and a healthy skepticism about silver-bullet notions of reform. The dirty dark secret of NCLB is that we may know how to identify the worst performing schools, but no one (yet) knows how to turn them around in any consistent and reliable way. And I mean no one. Not the Gates Foundation to date. Not most charter programs. No one.

As one study Ravitch cites concludes: "The only guaranteed strategy [for improving schools] is to change the student population, replacing low-performing students with higher-performing students." And this is, in fact, what the rare success stories—like KIPP—typically do: skim off the best and most motivated students from disadvantaged neighborhoods. These best students deserve better options, but this approach doesn't address the larger problem of how to fix chronically failing schools.

Ravitch's realism should be a bracing force in the dogged reform efforts—not crusade—we will be shouldering for a long time to come. Still, Obama, and other new voices in the debate, are surely right that teacher quality and recruitment are important ingredients of comprehensive reform. Near the book's end, she conjures up the ghost of her own beloved high school English teacher, Ms. Ratcliff, and Ravitch might have gone on to reflect on the main difference between then and now. In the 1950s, smart women, except for truly determined trailblazers, had few professional options beyond teaching. Ditto for blacks and other minorities. If you had a particularly smart and ambitious daughter, people would say, "I bet she grows up to be a teacher!" While many things have happened to public schools over the last 50 years, one of the most important is that this low-cost captive labor pool of extremely talented men and women has evaporated completely—and along with it the respect that was once automatically accorded to those who entered the profession. Today, with so many more (and better-paying) careers to choose from, it's unclear Ms. Ratcliff would be a teacher at all.

Ravitch is absolutely right to caution that trying to evaluate teachers solely on the basis of crude standardized tests is too reductive and will likely only alienate the kind of creative, dynamic people the profession hopes to attract in greater numbers. But the question remains: How do we lure more, talented people to the profession and give them—and the many superb teachers who already exist—the support and respect they deserve? For all its faults, if NCLB has done nothing else, it's helped to clarify this challenge.

Correction, March 19, 2010: This article originally misspelled Eric Hanushek's name. Return to the corrected sentence. (Return  to the corrected sentence.)

Correction, March 26, 2010: This article originally stated that the Eli and Edythe Broad foundation helped fund a study by two Stanford economists. Actually, it was the Michael and Susan Dell Foundation. Return to the corrected sentence.

Become a fan of Slate on Facebook. Follow us on Twitter.