Statistical malpractice

A record-breaking new class-action lawsuit against Wal-Mart claims the retail chain discriminates against women, for which of course vast millions of dollars are demanded. The New York Times aptly summarized the case — “about 65 percent of the company’s hourly paid workers are women, but only 33 percent of its managers are.” The grand fallacy of our times is various groups would be equally represented in institutions and occupations if it were not for discrimination. This preconception has undermined, if not destroyed, the crucial centuries-old legal principle that the burden of proof is on the accuser. Wal-Mart is only the latest in a long series of employers who have been hit with charges of discrimination on the basis of statistical differences among members of their work force — differences between women and men in this case. Back during the 1980s a similar charge was brought against Sears, though no one could find a single woman in the hundreds of Sears stores who had been discriminated against — just numbers different as between women and men. When you broke down the numbers, it turned out women were not equally represented among clerks who sold automotive equipment or construction materials. It also turned out many women had no interest in selling automotive equipment or construction materials, and had turned down chances to do so. In many other situations, women have avoided jobs that demand such long hours of work, or so much travel, that it would caring for their children virtually impossible. The biggest difference in income is between married women and everyone else. Women who never married have long held their own economically. The most blatant fact about male-female differences is often ignored by seekers of discrimination: Women have babies. That usually means interruptions in careers and different choices of careers before, because some occupations tolerate interruptions better than others. It is hardly surprising women work part-time more often than men, drop out of the labor force more often than men, specialize in a different mix of jobs, and major in a different mix of subjects in college and postgraduate education. Seldom are the data detailed enough to permit comparisons of women and men who are the same on all variables that matter. But the more detailed the data, the higher is a woman’s income relative to that of a comparable man, sometimes surpassing it. Male-female differences in incomes and occupations rose or fell throughout the 20th century as women’s age of marriage and childbearing rose and fell. But such mundane facts matter little to lawyers or social crusaders on the hunt for discrimination. Once a lawsuit is under way, the pressure is on the accused employer to settle, rather than risk bad publicity that could hurt profits. And, once they settle, that is taken as proof of guilt, no matter what anybody says. People without the slightest knowledge of economics or the slightest experience running a business will boldly assert women are paid only 75 percent — or some other percent — of what men make for doing exactly the same work. Think about it. If an employer could hire four women for the price of hiring three men, why would he hire men at all? Even if the employer was the world’s biggest sexist, he could still not survive in business if his competitors were getting one-third more output from their employees for the same money. Sheer dogmatic repetition has pounded into our minds the notion all groups have similar capabilities, when they are not necessarily even interested in developing the same capabilities. Potential may be the same but developed capabilities depend on much more, including interest and circumstances. Yet those who start with the preconception of equal capabilities are quick to seize upon numbers showing group differences in results as proof someone else did something wrong. That is the grand fallacy of our time. Thomas Sowell is a nationally syndicated columnist.