1. Standards are often too strict where they should be more flexible, and often too flexible where they should be more strict. For starters, there's a complete lack of coordination on standards among states. For some odd reason NCLB makes schools accountable to a series of standards without saying what those standards actually are. We have no idea what counts as "proficiency," there's no indication of what sorts of tests should be used. It's utterly bizarre. You could argue, for example, that standards are inappropriate because every school and neighborhood is different. Fine. But if you're going to have broadly accepted standards at all, they might as well be nationwide—it's not like kids in New York need different types of math from kids in Alabama. Meanwhile, schools shouldn't be allowed to fiddle with their own timelines, making it so that all the big achievement gains need to come years in the future.Now some of these things require more money, but not all of them do. What we really need is a calm, good-faith discussion about what sorts of tweaks the law needs, where schools are struggling, what resources states do and don't need, etc. At the moment, though, we have hysterical Democrats on the left, and dishonest Republicans on the right who capriciously slash funding, monkey with the regulations, and pull stunts like Armstrong Williams. Truly a pox on both houses. But notice: As soon as a Democrat who believes in NCLB comes to power, this can all end, and we've got a wonderfully workable, progressive law in the making.
Then there are places where more flexibility would help. Larger and more diverse schools are at greater risk of failing, simply because they have many more demographic subgroups. NCLB doesn't distinguish between a school with one failing subgroup and a school with 20 failing subgroups. Disadvantaged schools should probably be rewarded for making progress, rather than simply for meeting some absolute standard. "Value added" metrics are more appropriate for a law like NCLB, I think. And the law shouldn't punish good schools that take on a lot of weak students. Then there are various problems with how testing cycles are scheduled. But you get the idea. It's all highly tweakable stuff—nothing radical.
2. Too many of the remedy provisions are misguided. If a school fails for two years in a row, the district must allow that school's students some form of public school choice. But a districts with one failing school often has many failing schools, or all failing schools. The choice isn't always there. Wealthy suburban schools may not want your poor, tired, huddled classes. A failing school in North Dakota can't send its students anywhere.
Meanwhile, the incentives for improvement are all wrong. The theory behind NCLB's sanctions and punishments was that schools would feel the heat and start on the path to self-betterment. Not so. A school that starts losing students to its neighbors may welcome the change—and it may even see test scores rise through sheer numbers.
Again, though, this can be easily fixed. On the "lack of choice" problem, Andrew Rotherham has suggested simply increasing the supply of charter schools (though holding them to NCLB's standards). And Marc Tucker and Thomas Toch have suggested that NCLB contain provisions for "assistance teams" of experts to intervene and improve failing schools. In other words, more bureaucrats.