President George W. Bush has won plaudits for the diversity of his cabinet officials, most notably when he promoted Condoleezza Rice, an African-American woman, to secretary of State in his second term. Is this the kind of diversity that makes the White House team better equipped to come up with creative solutions to world problems? It is, according to the conventional wisdom in management circles, which for years has held that a diverse team is a creative one.

The idea is simple. When people of different points of view get together, they question each other's way of doing things and habits of thought. From this collision of different ways of looking at the world come new ideas--creativity. But what kind of diversity is most important for ensuring creativity? Researchers at Northwestern University set out to answer that question. Their results, published last week in the journal Science, suggest that although diversity is essential, ethnicity, race and gender may not be the most critical factors in fostering creativity.

The scientists sought to compare successful teams with less successful ones, and figure out exactly how they differed in composition. They focused on a few academic fields, defining successful teams as those whose papers were most frequently cited. They also looked at the annals of Broadway musicals over the past century and studied how the mix of composers, lyricists, directors and so forth correlated with the success of the production, measured in terms of the number of weeks a production ran before closing.

What they found was that the most successful teams did two things right. First, they attracted a mixture of experienced people and those who were newcomers to whichever field they were in. That's not surprising--the need for fresh blood has long been recognized as an important ingredient in success. The second criterion, though, was far less obvious. What successful teams had in common was at least a few experienced members who had never collaborated with each other. "People have a tendency to want to work with their friends--people they've worked with before," says Luis Amaral, a physicist at Northwestern and a coauthor. "That's exactly the wrong thing to do."

The most illuminating examples come from the Broadway musicals. "West Side Story" was considered to be a breakthrough for many reasons, not least the innovative use of dance and music to tell the story of the Sharks and the Jets. Among the handful of creators, though, were newcomers to Broadway (lyricist Steven Sondheim) as well as experienced people who hadn't collaborated before (composer Leonard Bernstein had never worked with choreographer Peter Gennaro). Broadway's worst decade, by contrast, came when it had a surfeit of big names, including Cole Porter, Gilbert and Sullivan, and Rodgers and Hammerstein. The problem, from the standpoint of creativity, is that these mavens--or "incumbents," as the authors call them--fell into the habit of collaborating with the same people repeatedly. As a result, between 1920 and 1930, 87 percent of Broadway musicals flopped, says coauthor Brian Uzzi, a business professor at Northwestern's Kellogg School of Management. (On average, 75 percent of musicals are flops.) "The Broadway musical industry had some of the biggest names ever," says Uzzi, "but the shows were too full of incumbents and repeat relationships."

One of the interesting things about this study is that it helps makes sense of the macro notion that relationships among innovators in an industry are an important ingredient in success. Studies have shown that part of the reason Silicon Valley is so good at spawning new companies is that there's such an extended network of people who worked together earlier in their careers. In Boston's Route 128 beltway, by contrast, networks are more "clumpy"--people tend to work with the same people, who in turn tend to be isolated from their counterparts at other companies. If more teams were formed with an eye to bringing together veterans who hadn't previously worked together, it would, over time, lead to a more Silicon Valley-like distribution of talent.

The study also suggests a role for technology in bringing seasoned people together. Tacit Knowledge Systems, a start-up in Palo Alto, California, is marketing a computer system that links people with similar professional interests. The system monitors e-mail in a corporation or other large organization and keeps tabs on what employees are interested in. If a worker is looking for somebody to collaborate with, he or she can query the system to find somebody appropriate. Tacit is developing a new version that actively forges connections by prompting employees when it finds people who, on the basis of shared interests, might make a good team. Finding a way to maximize creative potential is one of the most pressing problems in corporations. Knowing what makes one team more creative than another is an important first step.

Join the Discussion