My family recently traveled through a small airport. I breezed through the security checkpoint, but my wife and our 15-month-old baby were pulled over for a more thorough search. As the guard passed the metal detector over the baby's body, another traveler turned to me and said, "I guess they fear that al-Qaeda's recruiting them young these days."

To understand why security screeners are inspecting babies as closely as men, turn to the pages of Heather Mac Donald's new book about racial profiling: Are Cops Racist? newly published by Ivan R. Dee.

A dozen years ago, some 2,000 New Yorkers lost their lives to criminal homicide in a typical year--the equivalent of a 9/11 every 18 months. By the time he left office, Mayor Rudy Giuliani had reduced that number by nearly two-thirds. Giuliani's anti-crime achievements sparked imitation across the country. And how did the nation's leading newspapers and its governing administration interpret this incredible success? How else but as an outburst of systematic racism?

To listen to the American Civil Liberties Union, the New York Times, and other such voices, American police forces were deliberately and unfairly singling out young minority males on the streets and roads. "Racial profiling," they called it. Law-abiding people were being harassed and even arrested for no offense worse than "driving while black."

The accusation, Mac Donald shows, is a hoax--a hoax made up of two elements, one a complete falsehood, the other a half-truth.

Police in America do not stop motorists or pedestrians for no reason--they almost always have a cause, usually an infraction of some sort. As Mac Donald shows, these stops actually line up pretty exactly with the proportion of infractions committed by minorities. Blacks and Hispanics are much more likely than whites to drive significantly faster than the speed limit and to drive cars with broken tail-lights and other equipment violations. In fact, she claims, there's no evidence at all to support the proposition that a law-abiding black motorist is more likely to be pulled over than a law-abiding white one.

Where the differences do begin to show up is after the stop. Once police have halted a car, they will typically check its passengers for signs of suspicious behavior--whether it is intoxication or carrying drugs or guns. Mac Donald, who holds a degree in English from Yale, is a careful reader of people as well as texts--and she has spent hundreds of hours listening to police. They tell her that race alone means nothing, but race in conjunction with other factors can tell a lot.

"A group of young blacks with North Carolina plates traveling south out of Manhattan's Lincoln Tunnel into New Jersey? Good chance they're carrying weapons and drugs, having just made a big buy in the city."

And all those factors put together, police say, can justify a search. In the mid 1990s, 53 percent of the people searched on the southern end of the New Jersey Turnpike were black; only 21 percent were white. Discriminatory? Well, consider this: Searches of those black drivers were just as likely to find drugs as searches of the whites. In other words, the intuitions of the police matched up with reality.

Understand what Mac Donald is saying: not that 50 percent of black drivers get searched for drugs--but that 50 percent of those searched for drugs are black. And this proportion, she argues, reflects actual drug usage in America.

How does she know? Again, consider: Sixty percent of the people arrested in New Jersey for drugs and weapons offenses are black. So are about 60 percent of the killers and victims in drug-induced fatal brawls. In that context, the 53 percent search rate on the New Jersey Turnpike looks like a reasonable response to hard facts.

Mac Donald charges that the anti-profiling crusade is already hampering police work. New Jersey, Maryland, and California are imposing new rules that will discourage police from using the stop-and-search techniques that did so much to curb crime in the 1990s. And police are keeping score of the number of minorities they search in any given month--and when they reach their quota, they simply cease their law-enforcement activities.

Most ominously, though, she warns that anti-profiling may impair the war on terror--by, for example, wasting time and resources on searches like the one conducted on my daughter.

Mac Donald's work is a powerful corrective to the anti-police myth-making of recent years. Let me add just one caveat. Silly as it is to search babies, I myself am not so sure that the practice of random searches at airports is mistaken. In Europe and even in this country, radical Islamist groups and left-wing extremists have succeeded in winning thousands of perfectly respectable-looking people, not just to their antiwar protests, but to their campaigns of "direct action." It's not so incredible to me that one of those wholesome Midwestern college kids recruited to blockade the entrance of a Boeing plant could be persuaded to carry some device aboard a plane. It may be improbable, but wasn't 9/11 improbable? Our enemies in the Terror War are united by their ideas, not the complexion of their skins.

Where Mac Donald is absolutely right is in her insistence that police work should be governed by facts, not wishes. And a good way to start responding to things as they are would be to lay a copy of Heather Mac Donald's important book on the desk of every editorial writer, every academic, and every opportunistic state attorney general who has joined the anti-police chorus. They might learn something--even if it is only to treat with a modicum of respect the officers who risk their lives to protect everyone else's.

David Frum is a resident fellow at AEI.