Dr. Grossi's Blog

Collective Intelligence

Dr. Philip Grossi
Monday, 14 February 2011

In 1904 Charles Spearman published his findings on intelligence which showed that cognitive performance on one test  tended to predict performance on other tests in spite of huge differences in content, methods of administration, and skill tested. He concluded that there was a single general intelligence factor for which he coined the term g  factor.  Tests of intelligence are valuable because they can be administered in an hour or two and have wide utility in predicting school performance, success in occupations, etc.

illustration to collective intelligence blogThe complexity of cross-domain problems we face today require specialized knowledge and expertise in disparate areas and thus their solution requires more team effort than in the past. Therefore, elucidation of the factors influencing group performance is quite important if we want to make predictions similar to those for general intelligence. Now comes Anita Woolley and colleagues who published last year in Science the results of a study of "collective intelligence" which is called the c factor. They found evidence of a collective intelligence factor that explains a group's superior performance on a variety of tasks.  The c factor is correlated with social sensitivity of the group members, the equality of turn-taking in conversations, and the proportion of females in the group. Contrary to expectations, neither the average intelligence of the group nor the smartest individual in the group were the best predictor of the group's performance.  They also found that group cohesion, motivation, and satisfaction did not predict group performance.

These conclusions were drawn from two parts of a study with a sample size of 120 individuals drawn from the Boston area and 579 from the Boston and Pittsburgh areas.  In the first part of the study the 120 people were randomly assigned to teams of three who then worked on a variety of tasks. They found that performance on one task predicted performance on another task.  This was directly analogous to the approach that Spearman took regarding general intelligence over one hundred years ago.  The researchers then enrolled 579 people from Boston and Pittsburgh and assigned them to groups of two to five individuals and found that "social sensitivity", the degree of turn-taking in conversation, and the proportion of women in the group correlated most positively with collective intelligence. (Social sensitivity was measured by showing a photograph of a face cropped to show the two eyes and asking the participant to infer what the individual was feeling).

I think this research is significant and provides a guide to the formation of groups tasked to solve particular problems.