In an article about police shootings in last Sunday's New York Times (8/31), Michael Wines disputes the conventional wisdom about a disproportionate number of African-Americans being shot by police, saying there are no data one way or another. But Wines revives the canard about blacks being disproportionately targeted in traffic stops.

There actually is a study for that.

Throughout the 1990s, the nation was fixated on tales of jack-booted New Jersey state troopers who were stopping speeders on the turnpike just because they were black! In a 2000 primary debate, Vice President Al Gore sneered at then-New Jersey Sen. Bill Bradley, saying, "Racial profiling practically began in New Jersey, Senator Bradley."

Attorney General Eric Holder recently paid tribute to the myth, claiming that when he was in college, he had been stopped "driving from New York to Washington." He didn't mention how fast he was going.

The story never made sense. How could the troopers tell the race of drivers in speeding cars? Did they wait until the driver rolled down his window and, if he was white, say, "Oh, sorry—have a nice day!"

But the Clinton administration was slapping consent decrees for racial profiling on police departments across the country, and the N.J. highway patrol was its prime evidence, based on a study that a child wouldn't believe.

As is usually the case with bogus race studies, the pivotal 1993 survey compared speed stops on the New Jersey turnpike to the population of all drivers on the turnpike—not with the population of all speeders on the turnpike.

Such meaningless studies are popular on the left, where it is assumed that people of different races, genders and ethnicities will always behave identically in all respects.

If fewer women pass the physical test to become firefighters, that can only be because of sexism. If fewer blacks pass the written test —that's racism. If fewer whites play professional basketball—no, forget that one. Sports are important. (Unlike arson or vehicular homicide.)

Nonetheless, based on the assumption that blacks speed just as much as whites—because to believe otherwise would be racist!—Temple University's John Lamberth announced that while only 13.5 percent of drivers along a particular stretch of the New Jersey Turnpike were black, 46 percent of those stopped for speeding were black.

Racial profiling, Q.E.D.

The New York Times ran a dozen articles trumpeting the nonsense study, proclaiming it "the most thorough documentation of the contention that the police regularly pulled over black drivers." Lamberth himself praised his research for ruling out the possibility of coincidence—and if you can't trust Lamberth on his own study, who can you trust?

Largely on the basis of that investigation, New Jersey Superior Court Judge Robert E. Francis threw out the contraband evidence seized from 19 African-American men in traffic stops on the turnpike. New Jersey's ninny governor, Christine Todd Whitman, pronounced her state's troopers guilty of racial profiling and agreed to a consent decree with the Department of Justice (DOJ) that basically prohibited the troopers from doing their jobs.

Statisticians, and other people with common sense, tried to explain to liberals that human beings are not identical. Any study purporting to show that too many blacks are stopped for speeding must first determine how many speeders are black.

Being denounced as virtual Klansmen, the state troopers demanded a real study.

Confident that any new study would merely serve to confirm the troopers' racism, the DOJ and the New Jersey attorney general commissioned a statistical investigation from the Public Services Research Institute in Maryland.

The institute's study was a spectacular thing. Using expensive monitors with high-speed cameras and radar detectors, they clocked the speeds of nearly 40,000 drivers on the relevant section of the turnpike. Three researchers then examined the photos to determine the race of the driver—without knowing whether the driver was speeding, which was defined as going more than 80 mph in 65 mph zones.

The result: No racial profiling.

Blacks constituted 25 percent of all speeders and they were 23 percent of drivers stopped for speeding. Controlling for age and gender, blacks sped at about twice the rate of whites. The disparity was even greater for drivers exceeding 90 mph.

Inasmuch as the study was irrefutable, Mark Posner, a lefty Clinton holdover in the Bush Justice Department, tried to block it from being released, continuously demanding more information.

But no matter how statisticians fiddled with the data, the results were identical: Blacks were twice as likely to speed as whites—and at much higher speeds. The troopers were completely vindicated.

When the study finally leaked—over Posner's objections—he informed the press it wasn't "valid" without articulating any actual problems with it. The attorney general of New Jersey, David Samson, nonsensically said the results didn't matter because New Jersey had already admitted its troopers were engaging in racial profiling.

Perhaps the NYT is right and there is no comprehensive study of police shootings by race. But it's also possible that there is one, it didn't come out as planned, so it has never seen the light of day.

Ann Coulter is the legal correspondent for Human Events and writes a popular syndicated column for Universal Press Syndicate. She is the author of TEN New York Times bestsellers—collect them here.

Her most recent book is Never Trust a Liberal Over Three-Especially a Republican.