From the boardroom to the stockroom, corporations place a premium on forming and motivating effective teams. Although the focus in recent years has been on assembling teams with diverse skill sets and points of view, researchers have disagreed about the effects of such diversity on productivity. Now a new stream of research may lead to some answers.
“Networks, Diversity and Productivity: The Social Capital of Corporate R&D Teams,” in the July-August 2001 issue of Organization Science, attempts to reconcile two seemingly contradictory camps: those who feel that diversity leads to better performance because diverse perspectives will add a creative edge and those who feel that diversity impedes performance because diverse teams are less cohesive.
Previous research has focused on the trade-off between those two effects to determine whether diversity is good or bad for productivity overall. But authors Ezra W. Zuckerman, associate professor of strategic management at Stanford University, and Ray Reagans, assistant professor of management at Columbia Business School, suggest that diversity, per se, may not be the issue. Rather, they look at two underlying components of diversity: “network density,” or how much team members communicate with one another, and “network heterogeneity,” or what proportion of that communication is among colleagues of differing backgrounds. Chatting often by the water cooler might drive density, whereas whom you speak with at the water cooler drives heterogeneity.
The study found that diversity itself was not directly linked to productivity (in a sample of 224 R&D teams from 29 companies, some diverse teams were high performers, but others weren't), but both density and heterogeneity were linked. So a very diverse group that doesn't develop strong working relationships won't be productive. “You need a little diversity to drive network heterogeneity, but it won't drive it completely,” says Reagans, because you need social interactions among the diverse cohorts. Those aren't easy levers for managers to get at, the authors admit. “It's a lot easier for managers to manipulate the demography of a group than it is to manipulate the interactions within that group,” adds Zuckerman.
There's more research to be done. The study, and many others preceding it, groups employees by their tenure within the organization, but the authors readily concede that organizational tenure is not interchangeable with demographics like ethnicity, gender, age or residence. “Race and gender might have different effects because [they] are associated with what [researchers] call ‘bad conflict,’ ” says Reagans.