What is racial profiling, and is it racist? We can think of profiling as using cheap-to-observe characteristics as indicators or proxies for more-costly-to-observe characteristics. A person's physical characteristics, such as race, sex and height, are cheap to observe, and they might be correlated with some other characteristic that's more costly to observe such as disease, strength or ability.
Profiling examples abound. Just knowing that one person is 6 feet 9 inches tall allows one to predict that he's a better basketball player than a 4-feet-9-inch-tall person. That might be called height profiling. While height is not a perfect indicator of basketball proficiency, there is a strong association.
Similarly, just knowing the sex or age of an individual allows one to make predictions about unobserved characteristics such as weightlifting ability, running and reflex speed, and eyesight and hearing acuity because they are correlated with sex and age.
What about using race or ethnicity as proxies for some unobserved characteristic? Some racial and ethnic groups have a higher incidence of mortality from various diseases than the national average. In 1998, mortality rates for cardiovascular diseases were approximately 30 percent higher among black adults than among white adults. Cervical cancer rates were almost five times higher among Vietnamese women in the United States than among white women. The Pima Indians of Arizona have the highest known diabetes rates in the world. Prostate cancer is nearly twice as common among black men as white men.
Would one condemn a medical practitioner for advising greater screening and monitoring of black males for cardiovascular disease and prostate cancer, or greater screening and monitoring for cervical cancer among Vietnamese American females, and the same for diabetes among Pima Indians? It surely would be racial profiling -- using race as an indicator of a higher probability of some other characteristic.
You might say that's different and that using racial profiling as a proxy for potential criminal behavior is indeed racist. Just as race and ethnicity are not perfect indicators of the risk of certain diseases, neither is race a perfect indicator of criminal activity, but they are associations, and people act on those associations.