Diversity and Problem-Solving

There are so many reasons to despair about human diversity. There's Iraq, Kenya, the immigration debate, the research of Robert Putnam. It seems that, in tragic example after tragic example, humans react to diversity by splintering into tribalisms, regressing to an Us vs. Them mentality.

So that's why The Difference, a new science book by Scott Page, is so uplifting. The basic premise of the book is simple: when it comes to group achievement, diversity often trumps ability. To prove his point, Page draws on a variety of data, from the anecdotal to the experimental. But much of the book is actually about computer models of diversity. (Page's other book is called "Complex Adaptive Systems: An Introduction to Computational Models of Social Life".) The mathematical details of these simulations are numbingly complex, but their basic protocol is rather simple. Page typically begins with some sort of difficult abstract problem. (For the purposes of the simulation, the exact nature of the problem is irrelevant.) Page then populates his simulation with different kinds of "problem-solvers". Sometimes, he fills the simulation with a select sample of the top individual performers, or what Page refers to "the best and the brightest". (These performers are all talented, but they are all talented in the same way. Think of a group composed entirely of people with high math SAT scores.) Other simulations are filled with a randomly selected group of individuals, each of whom possesses a slightly different set of problem solving "tools". Such a random selection process ensures diversity, even if the diversity comes at the expense of talent.

When Page first started running his problem-solving model, he assumed that individual ability would trump group heterogeneity. But Page soon discovered that, at least in his simulation, the more diverse group almost always performed better than the "smarter" group. The best and the brightest would often get stuck at the same place, so that the problem became impossible. The randomly selected group, however, had a diverse cognitive tool kit, which allowed it to overcome such apparent obstacles. "Progress depends as much on our collective differences as it does on our individual IQ scores," Page writes.

Of course, this is just a model, a hypothetical abstraction running on a mainframe. And yet, it's still a nice affirmation of a civic ideal.

More like this

History has had no shortage of outstanding female mathematicians, from Hypatia of Alexandria to Ada Lovelace, and yet no woman has ever won the Fields medal - the Nobel prize of the maths world. The fact that men outnumber women in the highest echelons of mathematics (as in science, technology and…
In my post yesterday, I briefly mentioned the problem with simulations as a replacement for animal testing. But I've gotten a couple of self-righteous emails from people criticizing that: they've all argued that given the quantity of computational resources available to us today, of course we can…
Having spent a lot of time solving equations related to sticky tape models, including trying to work solutions in my head while driving to Grandma and Grandpa's with the kids, and making some measurements of real tapes, there was only one thing left to do: try simulating this problem in VPython.…
The Times has an interesting interview with Scott Page, a professor of complex systems, political science and economics at the University of Michigan: Q. In your book you posit that organizations made up of different types of people are more productive than homogenous ones. Why do you say that? A.…

Are differences a cultural necessity or a cultural tool? When I view the problems you listed, they don't strike me as diversity issues so much as geographic issues. People divide themselves into communities and cultures, which inherently requires definition, exclusion and a "protect ones own" mentality... but most of these divisions strike me as, historically, originating in geographic divides that served political purposes. It seems there /are/ communities where people of different races, beliefs or abilities do coexist when they have no reason not to. I think being Jewish in a modern American society is a good example: 80 years ago, I might have faced great stigma, but today, I generally don't. That's not to say that all divisions always disappear, but I do believe that diversity bears positive connotations as time progresses when one assumes no mitigating political (geographically based) forces. I think it is the use of diversity as a cultural/political tool which has brought about many of its negative connotations.

The phrase "Complex Adaptive Systems" really brought a smile to my face. I took a class by that name in undergrad. As the title implied, the focus of the course was a discussion of how complexity can arise from chaos. It was taught by two professors. One: the epitomy of disorganization with an office stacked flour to ceiling with papers. His office was an actual maze of information, and you had to navigate narrow paths to get to his desk. The other: a seeming a-type, organized, carefully, exacting. The two of them were very good friends and the flipsides to complexity theory's coin. They refused to write a syllabus, insisting that the structure of the course would arise out of necessity, and they also occasionally bickered like an old married couple. But I'm just being nostalgic : )

What may be the biggest factor here is that diversity of people means diversity of experience, knowledge, and skills.

As an analogy, look at your personal library. Put you together with a large number of people with identical libraries and the resources barely increase at all. Nobody brings in anything new to the group.

Compare this to the same number of people whose libraries are so unlike each other that seldom does any book appear more than once: sharing disparate libraries makes everybody smarter.

Or consider a similar analogy with software libraries, where we compare people working with the identical set of algorithms, versus the same number working with widely varied sets. For the second group, more parts of the problem at hand may already be solved in advance.

Diversity in experience, personality and so on is probably a strength, but when it comes to cultural or ethnic diversity I'm not so sure. Merging conflicting value systems will create rather than solve problems.

There is also the fact that if all division should dissappear then diversity would do so too.

So diversity of thought improves performance. This makes far more sense than the hypothesis that diversity of skin color, gender, religion, national origin or sexual orientation would improve performance.

The challenge of course rests in whether we structure ourselves in a fashion that empowers diversity in tackling problems. Most organizational environments are hierarchically structured and are therefore susceptible to following the flaws of decision-makers and experts. At the same time, completely unstructured and egalitarian arrangements, where every player and every idea is given equal weight are spectacularly inefficient and fail-prone. Distributed, multi-nodal, and semi-open networks seems to be an effective means, as the computer models clearly favor.

James Surowiecki's "Wisdom of the Crowds" looks at this very notion in much more detail, and in more human/social terms...

Diversity of thought is silly! That is everyone! True diversity is different backgrounds, cultures etc.