The growth over the years of Advanced Placement testing allowing high school students to earn college credit has been one of the better things to happen to American education.

Unfortunately, as with most things involving schooling, the subject of AP is beset with confusion. The most common problem is the recurrent confusion of taking AP classes with passing AP tests. The conventional wisdom assumes that the former more or less equates to the latter, so if we just get enough poor and minority students to take classes called "AP" then our problems our solved. But that's clearly not true.

Huge numbers of kids take courses in high school each year labeled "AP" and then bomb the national AP test in May. In contrast, other kids pass AP tests without taking AP classes, or even any classes at all in the subject. (My kid, for example, received 9 hours of college credit, more than half a semester, through AP testing for subjects he never even took in high school: World History and Comparative Government. He simply piggybacked off courses he did take, European History and US History and U.S. Government, with some home reading to fill in the gaps, such as memorizing the Chinese dynasties.)

A new book of social science research sounds like it can help clear up some of the confusion. (Of course, that assumes that people want to become less confused, which is a very risky assumption in anything having to do with American schooling.)

Education Week reports:

At a time of mushrooming interest in Advanced Placement tests, a new book assembles studies on how capable the program is of meeting the increasingly diverse expectations held up for it.

”The AP program is not going to solve all the problems of American education,” said Philip M. Sadler, one of the editors of AP: A Critical Examination of the Advanced Placement Program, being published this month by Harvard Education Press. ”I think AP is really good for some things and not very good for other things,” said Mr. Sadler, a senior lecturer in astronomy at Harvard University, ”and people should be aware of what its strengths and weaknesses are.”

Despite quibbles with some of the studies in the book, officials from the College Board, the New York City-based group that runs the program, called the new volume a ”landmark collection” in an area where scholarship is badly needed.

”The College Board has long lamented the fact that, except for the grants our organization distributes to fund AP-focused research, there has seemed to be little interest among independent researchers in AP,” Trevor Packer, the board’s vice president for Advanced Placement programs, wrote in an e-mail message.

Growing out of a symposium held at Harvard in 2007, the book focuses on AP science courses in particular and offers evidence on whether they give students an academic edge in college or persuade them to earn degrees in science-related fields.

One big example of the confusion between the value of taking AP classes and passing AP tests is that college admissions people tend to treat AP totally backward. At least in public discussions of how they weigh applications, they give more weight to you taking classes designated by your high school to be "AP" than to you getting good scores on the national AP tests.

In calculating high school GPA, they give an extra GPA point to any high class that claims to be AP (thus an "A" in Advanced Placement Psychology is worth a 5 instead of a 4), which is why, say, Berkeley students enter college with high school GPAs that sound hallucinatory (e.g., 4.47!) to older generations who didn't benefit from this gimmick.

But, admissions offices are reluctant to publicly admit that they give much weight to actually passing the national AP exams, even though those are very good predictors of whether you have the smarts and self-discipline to do well in college. Unlike high school GPA, AP tests are both nationally standard and they are more challenging than the vast majority of high school classes. Unlike the SAT, you can't be a smart slack-off and ace them. So, they function well as an acid test that this kid is college material.

In their chapter, for instance, researchers Chrys Dougherty and Lynne T. Mellor of the National Center for Educational Achievement, in Austin, Texas, draw on five years of data collected when Texas’ efforts to expand access to college-preparation courses were getting under way. They found that, once differences in students’ backgrounds were accounted for, AP students were no more likely to graduate from college than non-AP students. But the opposite was true for AP students who both took and passed AP exams.

College admissions offices, however, tend to argue that directly weighing scores on AP tests would discriminate against students at schools that don't offer AP classes. That's true, but less true than giving an extra GPA point to kids in classes labeled AP.

This book offers a new rule of thumb:

Based on that calculation, he figures that students who take honors courses ought to receive an extra half-point on a grade-point-average scale of 1 to 4, while AP courses ought to be worth an extra point, and an extra 2 points if students pass the exam.”I think it’s the only article I’ve seen that provides evidence for how to calculate high school GPAs,” Mr. Sadler said of his study. ”Everywhere else, they just use a rule of thumb.”

I recall, however, a University of California study that recommended reducing the the AP GPA bonus from 1.0 to 0.5 points.

The study also found that exam failure rates were disproportionately high among African-American, Hispanic, and low-income students, the disadvantaged groups the policy aimed to help, and that many such students had to take remedial courses in college.

States and districts have tried to increase access to AP by providing subsidies to cover students’ exam fees. But Ms. Klopfenstein argues that, given her findings on the amount of time it takes AP students to earn a degree, such policies are not cost-effective.

I've looked at a ton of AP data, and it's really obvious who is taking fewer than the optimal number of AP tests: Red State white kids. But nobody is supposed to be looking out publicly for the welfare of white kids, so you only hear about how more NAMs need to take the AP tests, even though their returns on taking them in terms of passing are much more diminishing.

A couple of things are going on in why Red State white kids tend to be oblivious to AP: First, AP is administered by the College Board, which is affiliated with ETS in Princeton, NJ, where the SAT comes from, so AP is bigger in SAT states than in ACT states. The ACT college aptitude test is developed in Iowa and is most popular in the middle of the country, so Red State kids tend to take the ACT and Blue State kids the SAT, and there is a spillover effect in terms of marketing.

Second, the people who are crazy about the AP are the Asians, who take about three times more per capita than whites, so whites who live near a lot of Asians tend to pick up AP Fever from them.