Richard Kahlenberg and Bernard Wasow build a pleasingly tight case against the usual argument for vouchers: that above all they improve test scores for the individuals involved. But I worry that their argument might also undermine the work of some of the public schools of choice I care about—whose test scores alone may not impress the world. I worry even more about the impact of vouchers on the larger society. They couldn’t have addressed all these issues in the space they had, especially since they took up the vexing but important issue of alternatives that might serve the positive purposes of vouchers without their drawbacks. So a few words on all of the above.

As the authors focus on test schools because proponents do, any argument for or against vouchers based largely on test scores is of course problematic. It’s hard for them to do otherwise, given that test scores are the only game in town. But in fact, tests are poor measures even of academic achievement. Worse, they ignore the other criteria by which most Americans—especially parents—judge the effectiveness of children’s education: moral and character development, reliability, responsibility, competence as writers, speaking ability, problem solving, capacity to get at facts, and qualities of leadership and team membership, to mention just a few. It would take a great deal more long-term study to see how vouchers affect performance in these areas—not to mention whether students stick around long enough to graduate.

It’s even possible that vouchers are strongest in the unmeasured areas. In fact, if the receiving schools were more diverse than the schools the students came from, this should have a positive impact, according to Kahlenberg and Wasow’s own central thesis. In addition, I’d be surprised if the voucher schools weren’t smaller and didn’t have more cohesive missions—two qualities which, studies suggest, produce stronger outcomes. (I sent my own children to local public schools, assured by evidence that this would have no effect on their SAT scores. I later realized that test scores were the measure least sensitive to the impact of schooling. Neither their gains nor their losses could be measured by test scores.)

But even if the facts did suggest that students given vouchers actually do better on both tested and untested skills and aptitudes—which they don’t—the argument would not be over. The long-term impact of removing yet another public space from our lives and the gradual privatizing of all society’s important decisions are more fundamentally dangerous. Democracy has always been a fragile idea. Faced with the obvious, that “the people” may often be wrong, we are sorely tempted to find substitutes. Democracy is under subtle attack by another idea: that the private marketplace serves the same purpose as democracy but with less fuss and bother and no need for messy politics. Each act of privatization of public functions eats away at the idea of the common good, commonly decided. By removing the rationale for public debate over how best to raise the next generation, we relinquish our collective voice in the future.

I like the authors’ alternative version of controlled choice. But alas, given the current political and social climate it is, at least at this time and place, not likely to work any better than the forcible integration of schools did 30 years ago.

And for similar reasons. It is unrealistic to imagine that we can arrange it so that the poor are a minority in a wide enough range of schools to seriously tackle inner-city school failure. I base this on years of personal experience as a parent and professional in Chicago, New York, and now Boston, as well as on studies about the impact of choice on the diversity of schools. Few schools are able to sustain both ends of the socioeconomic curve—unless they resort to ability grouping (tracking). Even this works only sometimes as a means of holding the upper end of the curve. And tracking undermines the likelihood that the benefits will accrue to the less affluent children.

Few middle- and upper-middle-class parents are prepared to send their children to schools in which they keep company with even a healthy minority of poor and academically “slower” learners. Fewer still if special “gifted” programs are not available. There is evidence that, at best, their alternative version of controlled choice can work in early childhood programs, but the numbers required before the tipping point is reached are harder to sustain as students get older. Many factors go into this, including concern that a less competitive and rigorous set of standards will prevail in classes with large numbers of low-income students, and that behavioral problems will accelerate along with diversity. Even early childhood, no longer viewed as a time for play and socialization, is not immune from such fears. (And of course, aside from fears about social class mixing, there is racism.)

Studies suggest that parents and youngsters opt for settings where the school and family “match.” Mixing kids up may be good for society, but what about its impact on “my kid”? Getting the company they keep right is high on the list. For most families that means “like us,” with or without choice.

The gaps in education cannot be seriously addressed if we are not prepared to simultaneously tackle issues of housing, neighborhood life, and the average American family’s working lives. The authors’ perfectly sensible ideas—based on long-known and irrefutable research—only appear irrational because of the constraints we are working under. All kinds of ideas are not even on the table.

So while controlled school choice is a strategy worth pursuing, the racial or socioeconomic mix it offers may not be its only benefit. Equally critical, choice improves the odds that students keep company with interesting adults doing interesting work and witness and identify with the work of older children. The efficiencies of age-grading and strictly differentiated adult roles were largely a natural consequence of enlarged school size and the consequent need for bureaucratic inflexibility to replace human judgment. Schools of choice can more easily stay small, thus allowing adults to know their students truly well, and to connect youngsters to adults beyond the four walls of the school through community service, weekend and summer camps, and so on. Wherever we create small and intimate settings of choice, we have a chance to create a stronger and more cohesive culture of learning for all children and a greater likelihood that family and school will work together toward common ends. Greater diversity in the backgrounds of the families—above all, socioeconomic diversity—would help. But until we tackle the kinds of issues that will make this more feasible, we can better utilize the strengths of each school’s particular adult community. Reducing school size and ensuring choice are both useful tools toward these ends. These measures will not suffice to eliminate the gaps that separate the rich and poor, but it’s not a perfect world, and every step we take in the direction the authors propose will shift the odds a bit in our favor. Someday we’ll have the will again to tackle all the other gaps that divide us.