Pages: 1 2
We live in a world of imperfect and costly information, and people seek to economize on information costs in a variety of ways. If we don’t take that fact into account, we risk misidentifying and confusing one type of human behavior with another. Let’s look at it.
Pima Indians of Arizona have the world’s highest diabetes rates. With knowledge that his patient is a Pima Indian, it would probably be a best practice for a physician to order more thorough blood glucose tests to screen for diabetes. Prostate cancer is nearly twice as common among black men as white men. It would also be a best practice for a physician to be attentive to — even risk false positive PSAs — prostate cancer among his black patients. What about physicians who order routine mammograms for their 40-year and older female patients but not their male patients? The American Cancer Society predicts that about 400 men will die of breast cancer this year.
Because of a correlation between race, sex and disease, the physician is using a cheap-to-observe characteristic, such as race or sex, as an estimate for a more costly-to-observe characteristic, the presence of a disease. The physician is practicing both race and sex profiling. Does that make the physician a racist or sexist? Should he be brought up on charges of racial discrimination because he’s guessing that his black patients are more likely to suffer from prostate cancer? Should sex discrimination or malpractice suits be brought against physicians who prescribe routine mammograms for their female patients but not their male patients? You say, “Williams, that would be lunacy!”
Is an individual’s race or sex useful for guessing about other unseen characteristics? Suppose gambling becomes legal for an Olympic event such as the 100-meter sprint. I wouldn’t place a bet on an Asian or white runner. Why? Blacks who trace their ancestry to West Africa, including black Americans, hold more than 95 percent of the top times in sprinting.
Pages: 1 2