"Group intelligence quotient"

A news article in Science magazine, based on a research report in the same issue, suggests that there may be a generalized ability of small nominal groups to solve problems analogous to the general intelligence attributed to individuals by "intelligence tests". Interestingly, this group intelligence seems only weakly predicted by the intelligence of the smartest member of the group and/of the average intelligence of group members, but that the social ability of the group members to work together is also predictive of group problem solving ability.

It occurs to me that the research was conducted using nominal groups -- that is putting people together for the purpose of the experiments in problem solving, which were rather short term. That situation might be different than the situation in a formal organizations such as a research laboratory or corporate executive suite in which people work for extended periods on a limited class of problems with feedback based on success or failure of past approaches. Certainly my own experience in small group professional decision making situations suggests that such groups are far more effective with increased experience, and in that sense do "learn" to solve problems better.

Thus, less socially adept groups might still find ways to work together effectively over time, increasing the importance of individual intelligence in group problem solving. Still, it might well be the case that the average research group would do well to spend some time on the social aspects of planning and evaluating its research agenda.