The study, conducted by the university’s Haas School of Business, found that contrary to previous findings in analyzing race and gender in the workplace, "diversity does not consistently predict high turnover."

Published by Berkeley professors Jonathan Leonard and David Levine in the July issue of the journal Industrial and Labor Relations Review, the study was intended in part to gauge whether there was evidence supporting claims by industry consultants that diverse workplaces require special training to mitigate the risk of high turnover.

"We think a reasonable interpretation of [the findings] is employees don’t care too much about the race and gender of their co-employees," Leonard says.

The professors analyzed turnover rates during three years among 70,000 employees at 800 workplaces of an American retailer to see if workers left at a higher rate at stores that were more diverse. The retailer asked not to be identified, Leonard says.

Diversity training has become standard at many companies, according to the Society for Human Resource Management. Seventy-eight percent of employers who implement diversity practices say they do so to lower the cost

of turnover, absenteeism and low productivity.

Diversity consultant Terry Simmons, managing partner of Simmons Associates in New Hope, Pennsylvania, disagreed with the study’s hypothesis: Turnover is lower in workplaces with little diversity.

"They’re saying diversity is a negative condition you have to overcome," Simmons says. "I’m saying people are different and that’s great; how do we use that to increase productivity? It’s a different starting point."

The study noted some exceptions when turnover increased. The authors write that "isolation"—when an individual is a numerical minority within the work group—is a useful predictor of retention. Among most racial groups, turnover is lower when they are the numerical majority. Blacks, followed by Hispanics, were most sensitive to being a numerical minority within the group.

Among seven categories—age, men, women, whites, blacks, Asians and Hispanics—women had higher turnover rates when working in a mixed group than in a highly male or highly female setting. By a small margin, women in the study had lower turnover rates when the majority of their co-workers were male.

As a matter of policy, the authors called the results "encouraging," since they hinted that workers were indifferent to issues of race and gender when choosing to leave a job. The study’s conclusions, however, may not hold in sectors such as insurance and banking, where people are more invested in their jobs and stay longer, and where teamwork is more important, Leonard says.

—Jeremy Smerd