It is a tenet of the diversity movement that variation of individuals within a group makes the group better. Every team, committee, or board must be formed with diversity in mind. The magic of diversity will lead to better decisions, avoid groupthink, and save us all. The trouble is that most often the wrong kind of diversity is chosen, and the magic uses up all its manna without producing a single skittle or rainbow.
When we speak of diversity, what is it that we mean? There are several types of diversity: biological, random, natural, racial or type, or geographical. There is skill diversity; educational, opinion, viewpoint diversity, and so on. It is a mistake to conflate any of these types of diversity, or to assume that one kind necessarily implies another.
As Justice Clarence Thomas said in a BusinessWeek interview:
I've thought a lot about these things, and I've spent the bulk of my life, beating my head against a wall, trying to get people to see that they can have their grand theories but, in the end, you can't impose them on other people's kids. How many kids do you have? They're different, aren't they? If your kids are different—and they're all yours—what about just some kids who happen to be different shades of black, different degrees of Negro? They're all from different family settings—some two parents, some no parents, some raised by grandparents. Come on. How can you just all of a sudden treat them as all the same?
Suppose you were creating a team of ten people, and your goal (hypothetically) were to have three people with dark skin, three with blond hair, three females, four above a certain height, and three in each decade of age ranging from 20 to 50. To get this diversity, you could look at their resumes and have them submit by email answers to questions about nuclear physics, Central Asian history, and field hockey. That would let you select a diverse set of opinions and knowledge, which would then yield a chance of successfully getting the physically diverse group you are after.
What's that? Silly, you say? The obvious way to make sure to get the exact mix of physical attributes you needed would be to ask the people for their physical attributes, and ignore altogether their knowledge of desert flora or India's glorious past, since that would have nothing to do with your needs.
But we see the reverse employed: to achieve a goal of certain knowledge, skills and viewpoints on a team, members are chosen for their physical or even educational diversity, in the hopes that somehow they will have the needed abilities. It would be far better to pick people with the needed skills and experience levels, and not worry about their irrelevant characteristics. They may not look different, but they will be as different as they need to be, in the way that really counts.
And even if the goal is opinion diversity itself, and not any specific set of skills, knowledge, or beliefs, using racial or other physical diversity to achieve opinion diversity relies on stereotyping. As Justice Thomas notes, a physically homogeneous group of people probably will not think all alike. Conversely, it is quite possible to find people of different physical descriptions who will have very similar views, especially if they are all members drawn from a certain set of candidates.
The key failing of using physical diversity as a proxy for viewpoint diversity is that it introduces unrelated variables, and we don't know what those variable are. There is tremendous, unknowable variation within what we call "races", and the difference in viewpoint between any two people of a given race might be greater than the difference between either of them and someone of another race.
So are diverse teams really better? A committee, team, or other group is tasked collectively with performing a project no one of them alone could do. A team requires therefore a range of skills. The skills of the team members must mesh together to accomplish the team goal. Choosing people for "diversity" yields random skill sets, and unknown skill levels. To believe that diversity implies meshing requires magical thinking: it will just somehow work.
Diversity should be a natural byproduct of choosing the right people for the job. Insisting on doing it some other way belies the belief that diversity aids the team, showing it to be an end in itself. Diversity in that case is not the source of any strength, but an exercise in bureaucratic phoniness.
(This is a post I wrote at my old blog in 2007. It is posted as a sidebar for an upcoming one on immigration, which I think is what Barack Obama intends as occupation for our autumn.)