(FORTUNE Magazine) – The presidential candidates have seized on education in the hope of finding an issue that will ignite voter excitement. Bush's and Gore's plans differ in nuance, but both men advocate policies that would end up pouring more money into the school systems in order to make them more accountable for their students' performance. The candidates' motives are good, and on the surface their timing seems perfect: Government coffers are flush with cash. But does spending more money on education actually make any difference?

The hard data indicate it doesn't. While real spending per pupil increased 80% between 1970 and 1995, average test scores as measured by the National Assessment of Education Progress remained flat. The states that spent the most on education haven't produced the best-performing students (see chart). "There's a consistent story from the evidence," argues Eric Hanusek, of Stanford's Hoover Institute. "We shouldn't expect any gains from extra funds, given the current situation." Hanusek's survey of major education research projects, first published in 1981 and updated five times since, concludes that there's no discernible relationship between government spending and education. To date, it has been the most influential paper on the matter.

But now the study is under attack by economists, who say the truth about the effects of education spending is more complicated. Sure, test scores on the whole have stagnated, but those of African-American and Hispanic students have actually improved substantially in the past two decades of higher spending. Moreover, a higher percentage of school funding is earmarked for social programs, such as AIDS education; the rate of return on these programs can't be measured by a standardized test score.

These trends have led a number of economists to reexamine Hanusek's study. Their findings are provocative: While middle-class white students may not be helped by more education spending, disadvantaged social and economic groups are. Says David Grissmer, senior management scientist at Rand and lead author of a newly released study on the subject: "The real payoff is for disadvantaged kids. We have put a lot of money in higher-income kids already and may be near the saturation point."

The leader of this new wave of thinking is Alan Krueger, an economics professor at Princeton. Krueger recently finished a study of his own that reworks Hanusek's. Instead of using Hanusek's method, which weighted the results from the 59 papers differently based on how important he judged them, Krueger weighs all the results equally. His findings are antithetical to Hanusek's: Money does improve academic performance. "It shouldn't be surprising," says Krueger. "After all, schooling is an investment, and it takes money to make any investment." Krueger says he was surprised that the average rate of return on that investment in the form of test scores doesn't appear to be high, but when the money is targeted, "there's a much, much higher payoff."

So how can Bush and Gore effectively target spending on education? Economists have found three factors that matter most: class size, teacher quality, and school size. There's a lot of evidence that demonstrates class size is particularly crucial in a student's early years. In the late 1980s, Tennessee conducted a now famous random trial in which administrators shrank class sizes from kindergarten through third grade in 79 schools. They found that low-income and minority students in the smaller classes performed much better than kids from the same social and economic groups in larger classes. Later the state followed up on these same 7,000 students to see how they were performing in high school. The students from the trial were more likely to take the SATs, and fewer were held back a grade. After evaluating the results, states like California (which ranks dead last in student test scores) are trying to limit kindergarten through second-grade classes to 20 students or fewer. Research has also shown that a teacher's verbal skills are important in early grades, while in-depth knowledge of their subject matter is more crucial in high school. The last factor, school size, seems to be more significant in inner cities, as studies have found students at smaller, more intimate schools tend to have better test scores than those at large high schools.

Spending money on some kids and not on others is a hard sell politically, which is why neither Bush nor Gore is advocating it. Their solutions are more far-reaching: school vouchers in Bush's case, better-paid teachers in Gore's. With the work force more productive than it has been in years and the economy in the midst of a record expansion, it's hard to argue that the country is in the midst of a serious educational crisis. The trouble may be in the longer term. It's unclear whether our educational system can work well enough to keep up with the demands of this tech-driven, knowledge-based economy. That's why the candidates should think about where they're allocating money, rather than simply writing a check.