Check out commencement photos from elite private universities like Harvard or Stanford, and you'll see a considerable number of black and Latino graduates. Race-conscious admissions policies are the main reason.

A recent study finds that, at such schools, black and Hispanic students receive the equivalent of a 200 point edge in SAT scores. (Several states, including California, prohibit public universities from taking race into account, which is why just 3 percent of Berkeley students are African American - the same percentage as 30 years ago, when Berkeley started keeping racial statistics.)

Conservatives cry foul. Not only is affirmative action unfair to better-qualified white students, they argue, it's also unfair to those "under-qualified" minority students. The "minority mismatch" argument goes like this: Nonwhites who wouldn't otherwise be admitted to selective universities find themselves competing with better-prepared white students, and that's a prescription for failure. They'd be better off at a school a rung or two down the prestige ladder.

But the research disproves this mismatch claim. Give minorities the chance to attend top-flight universities, the evidence shows, and they'll actually do better. That's not only good for them, it's also good for the country.

A case in point is a recent study of Texas' "10 percent" admissions policy, which guarantees high school students who graduate in the top tenth of their class a place at one of the state's flagship schools, the University of Texas and Texas A&M. Their SAT scores don't matter; and whether they come from a leafy Dallas suburb or a Brownsville barrio doesn't matter, either.

Because many Texas high schools are overwhelmingly Hispanic or black, the 10 percent rule has boosted minority enrollment. And since those students typically have lower SAT test scores, their experience is a good test of the mismatch theory. If the theory is right, these black and Latino students should do worse than if they had attended a less selective university. In fact, those students are 21 percent more likely to earn their bachelor's degree than are students with similar credentials who enroll in a second-tier school. The losers are minority students whose class rank isn't quite in the top 10 percent but who, because they graduated from top high schools, previously would have been admitted to one of the top-tier state universities - their graduation rates decline. Better schools = higher expectations = better outcomes.

Berkeley's experience with community college transfers bolsters these findings. A quarter of Berkeley's undergraduates are transfers, most of whom wouldn't have been admitted as freshmen on the basis of their high school grades or SAT's. But they do as well as the "true freshmen" - they major in the same subjects, have the same grades and graduate sooner.

The conclusion that students' academic careers are molded by what's expected of them might surprise the anti-affirmative action crowd, but it's old news to any college professor. At Harvard or Stanford, flunking out is a rarity. Students on the verge of failing get kid-gloves treatment, the expectation being that anyone who is good enough to be admitted is good enough to graduate. By contrast, at Big U, the mass universities where almost all applicants are accepted, students largely are left to their own devices and it's assumed that many will drop out or flunk out.

Big U could learn a lot from studying the "Accelerated Schools" model widely used in public schools. The program's premise is that at-risk students can thrive with the kind of idea-filled instruction typically reserved for gifted children instead of having to suffer through skill-and-drill teaching supposedly geared to their abilities. Evaluations have largely borne out this prediction: set the bar higher and students will respond in kind.

Another line of research, focusing on what psychologists call "stereotype threat," suggests the same conclusion: Expectations can spell success or failure. Minority students are especially vulnerable: When they are exposed to negative stereotypes, their test scores plummet. "Conditions designed to make black subjects stereotype-vulnerable," like telling them that the Graduate Record Exam is a test of their ability, "depress their performance relative to white subjects," says Stanford psychologist Claude Steele, a pioneer in the field. The good news is that when this stereotyping is eliminated - when undergraduates are told they're taking a problem-solving test, not an ability test - black students do dramatically better.

All the research tells the same story: Expectations are crucial. Advise minority students with comparatively weak paper records that they have a decent shot at succeeding, enroll them in highly demanding schools, and they're more likely to thrive than if they're fed an intellectual diet of pabulum.

Goodbye ideology, hello evidence? While it's hard to persuade ideologues to confront inconvenient truths, this research should embolden defenders of affirmative action to challenge what then-President George W. Bush rightly denounced as the "soft bigotry of low expectations." Race-conscious admissions is not, as the critics insist, a sop to political correctness. It's smart human capital policy.

David L. Kirp, James D. Marver professor of public policy at UC-Berkeley, is the author of the upcoming "Improbable Scholars: The Rebirth of a Great American School System and a Strategy for America's Schools." This article was originally posted on SF Gate.