The New York Times’s “Bits” blog has a post titled “Google Releases Employee Data, Illustrating Tech’s Diversity Challenge.”  It opens by stating that “Google on Wednesday released statistics on the makeup of its work force, providing numbers that offer a stark glance at how Silicon Valley remains a white man’s world.” Yet just a few sentences later, the post states that “Of its [Google’s] United States employees, 61 percent are white, 2 percent are black and 3 percent are Hispanic. About one-third are Asian — well above the national average — and 4 percent are of two or more races. Of Google’s technical staff, 60 percent are white, 1 percent are black, 2 percent are Hispanic, 34 percent are Asian and 3 percent are of two or more races.” Because the United States workforce overall is 5% Asian according to the Bureau of Labor Statistics, Asian-Americans are over-represented by a factor of six. It is stunning that a company with such high numbers of Asian-American employees could be characterized as “a white man’s world.”

The rest of the post discusses the strategies that Google and other tech firms might use to boost their numbers of African-American and Hispanic employees.  But it ignores the literature on academic mismatch that indicates that racial preferences in university admissions are an important cause of the low numbers of African-Americans and Hispanics in computer science and other technical fields. Because of racial preferences in admissions, African-Americans and Hispanics are disproportionately likely to wind up at colleges and universities at which their entering academic credentials are significantly lower than those of the average student. Science and engineering majors are among the most difficult on many college campuses. A student with credentials well below those of his classmates is therefore likely to have a particularly difficult time in one of those fields and thus to switch something easier. Yet the research shows that many of those same students could succeed at a school where their credentials were closer to those of the median student. To put it in concrete terms: a student with a 1200 out of 1600 on the SAT I is likely to leave science for something easier if she attends a university where the median student’s score is 1450. But if she attends a college where the median student’s score is 1200, her chances of completing a science degree dramatically improve. For a short summary of the relevant studies, see Gail Heriot, “The Sad Irony of Affirmative Action,” National Affairs (Winter 2013); for longer discussions of them, I recommend U.S. Commission on Civil Rights, Encouraging Minority Students in Science Careers and Stuart Taylor and Richard Sander, Mismatch: How Affirmative Action Hurts the Students It’s Intended to Help, and Why Universities Won’t Admit It.

In other words, Google and companies like it might well have more Hispanic and African-American employees if universities didn’t use large racial preferences. It is unfortunate that the NYT did not see fit to mention this.