Free Team Genius: The New Science of High-Performing Organizations by Rich Karlgaard, Michael S. Malone

Page is right, then the standard (and often government-required) “diversity” practice of hiring graduates from a similar set of top universities while ticking off the boxes for race, gender, ethnicity, and so forth may be a misdirected effort. As externally diverse as these new hires may be, their socialization, training, and education may render them very similar to other new hires in terms of the heuristics, perspectives, interpretations, and predictive models they use to solve problems and achieve their goals. In other words, they aren’t diverse at all—and filling a team with them will likely prove to be suboptimal.
    So just hiring more women as per Woolley works for Page only if those women come from sufficiently unusual backgrounds to think differently from their new teammates. Otherwise, if they are merely cut from the same cloth as the male members of the team, they will have only a minor impact (that is, there will be a comparatively small cognitive difference between the sexes). What matters most are differences in culture, class, and aptitudes.
    To help explain his model, Page has introduced what he calls the diversity prediction theorem :
    The squared error of the collective prediction = (average squared error − predictive diversity)
    Yeah, that’s pretty complicated. But it boils down to this: Teams err when they lack accuracy and diversity. So, when group diversity is large, the error in the team is small .
    Page goes so far as to warn against using traditional stereotyping in selecting for diversity because it may lead to team members’ living down to the expectations imposed on them.
    In one classic study, Asian women were judged differently on their mathematical skills based on whether they were primarily described by their gender or their ethnicity. 16 When the participants were described by their gender, they were rated lower (“womenare bad at math”) than when they were described by their ethnicity (“Asians are good at math”). Even more troubling, the women themselves, when given a math test, performed to those judgments.
    The lesson is that diversity is powerful in teams, but only if it is real diversity.
    That’s just the beginning. Even when we can all agree that diversity is critical to team performance, it’s still a real-world challenge to figure out how to blend all those ingredients into a high-performing team—much less make all those different personalities get along.
    As the research suggests, there’s no point in adding new members to a team just because they fit your diversity requirements. If they cannot influence or improve the team’s collective decision-making, they have little value. Indeed, being placeholders, they may actually reduce the total intellectual capital of the team. 17
    More than forty years of research on diversity has been conducted by psychologists, sociologists, economists, and organizational scholars. A review of this literature in 1999 by Katherine Y. Williams and Charles A. O’Reilly of Stanford Graduate School of Business underscored that diversity is a double-edged sword. 18
    The good news is that group diversity can enhance performance, because group members bring to bear varied ideas, knowledge, and skills to accomplish tasks. However, in a diverse group, members may view each other through a biased lens of stereotypes based on social categories (the same differences in race, gender, and so forth that are supposed to help). This bias invariably reduces the effectiveness of the group’s interaction as group members fail to identify with the group. 19
    Williams and O’Reilly’s key insight is that team members maintain their self-esteem by making comparisons with other team members and then classifying themselves using those same salient characteristics of race, gender, and so forth. And when a particular characteristic allows members to assume a positive self-identity,they then look upon those who lack that characteristic as being out-group