It isn’t the total IQ of the team, and whether they’re working face to face doesn’t matter.
The factors discovered were that the members make fairly equal contributions to discussions, level of emotional perceptiveness, and number of women, though part of the effect of number of women is partially explained by women tending to be emotionally perceptive.
On the one hand, I’ve learned to be skeptical of social science research—and I add some extra skepticism for experiments that are simulations of the real world. In this case, the teams were working on toy problems.
On the other hand, this study appeals strongly to my prejudice in favor of niceness. I found the presence of women to be a surprising factor, since I haven’t noticed women as being easier to work with.
A notion: the fairly equal contribution part may be, not exactly that everyone contributes more, but that if the conversation is dominated by a few voices, those voices tend to repeat themselves a lot, and therefore contribute little compared to the time they take up.
I wonder how all female groups compare to groups with just one male, and how all male groups compare to groups with just one female. It seems to me like it’s harder for any one person to dominate whenever people feel the need to signal egalitarian values like a preference for gender or racial equality. I don’t know anything about statistics yet, so maybe this is implausible, but I think part of the reason that diversity was an insignificant predictor was that poor theory of mind caused by (?) ingroup favoritism dominates the effect as diversity increases and it drowns out the effect of the need to signal egalitarian values, so I think it would be cool to see how the collective intelligence changes when you go from ‘completely’ homogenous to ‘almost’ homogenous in experimental groups composed of subjects from cultures that value egalitarianism highly. I would like to see this replicated by subjects from less egalitarian cultures as well, but that’s hard sometimes.
My guess: People in the team need to communicate. This can be essentially achieved by two ways:
1) All team members voice their opinions openly.
2) Some team members don’t voice their opinions, but other members are good at reading emotions, so the latter recognize when the former believe they know something relevant.
If this model is true, we would see that equal contribution (no one is silent) or emotional perceptiveness (other people recognize when the silent person wants to say something) increase the team output.
What makes teams more effective
It isn’t the total IQ of the team, and whether they’re working face to face doesn’t matter.
The factors discovered were that the members make fairly equal contributions to discussions, level of emotional perceptiveness, and number of women, though part of the effect of number of women is partially explained by women tending to be emotionally perceptive.
On the one hand, I’ve learned to be skeptical of social science research—and I add some extra skepticism for experiments that are simulations of the real world. In this case, the teams were working on toy problems.
On the other hand, this study appeals strongly to my prejudice in favor of niceness. I found the presence of women to be a surprising factor, since I haven’t noticed women as being easier to work with.
A notion: the fairly equal contribution part may be, not exactly that everyone contributes more, but that if the conversation is dominated by a few voices, those voices tend to repeat themselves a lot, and therefore contribute little compared to the time they take up.
Here are the papers:
Woolley, Chabris, Pentland, Hashmi, & Malone, 2010
Engel, Woolley, Jing, Chabris, & Malone, 2014
I wonder how all female groups compare to groups with just one male, and how all male groups compare to groups with just one female. It seems to me like it’s harder for any one person to dominate whenever people feel the need to signal egalitarian values like a preference for gender or racial equality. I don’t know anything about statistics yet, so maybe this is implausible, but I think part of the reason that diversity was an insignificant predictor was that poor theory of mind caused by (?) ingroup favoritism dominates the effect as diversity increases and it drowns out the effect of the need to signal egalitarian values, so I think it would be cool to see how the collective intelligence changes when you go from ‘completely’ homogenous to ‘almost’ homogenous in experimental groups composed of subjects from cultures that value egalitarianism highly. I would like to see this replicated by subjects from less egalitarian cultures as well, but that’s hard sometimes.
My guess: People in the team need to communicate. This can be essentially achieved by two ways:
1) All team members voice their opinions openly.
2) Some team members don’t voice their opinions, but other members are good at reading emotions, so the latter recognize when the former believe they know something relevant.
If this model is true, we would see that equal contribution (no one is silent) or emotional perceptiveness (other people recognize when the silent person wants to say something) increase the team output.