The structure of group task performance—A second look at “collective intelligence”: Comment on Woolley et al. (2010).

Abstract

Collective intelligence has been described as a general factor that "explains a group's performance on a wide variety of tasks" (Woolley, Chabris, Pentland, Hashmi, & Malone, 2010, p. 686), much like the general intelligence factor explains individuals' performance on cognitive ability tasks. This construct has received widespread attention in both the media and academic community. In this article we reexamine the data from 6 previously published samples that have been used to examine the existence of the collective intelligence construct and show that the empirical support for the construct is generally weak. Specifically, we show that the general factor explains only little variance in the performance on many group tasks. We also highlight how 2 statistical artifacts-the apparent presence of low effort responding and the nested nature of the data-may also have inflated the little covariation that exists between group performance on different tasks. These findings suggest that there is insufficient support for the existence of a collective intelligence construct. (PsycINFO Database Record

Towards Stable Principles of Collective Intelligence under an Environment-Dependent Framework

“…Skill diversity (i.e., variance in group members' ability) has been shown to both enhance [98] and handicap collective performance [18,67,60]. Similar inconsistencies arise for cognitive style diversity [118,150,3], social perceptiveness [217,69,72,120], and even the relative performance of teams versus individuals [48,190,219].…”

“…As organizations have moved inexorably to more team-based structures, the problem of improving team performance through judicious selection of team members has preoccupied management scientists and managers alike [115,2,101,197,145]. Previous research has found a variety of intriguing results regarding the impact of skill diversity [98], cognitive style diversity [68], and social perceptiveness [110,217] is little consensus on whether teams always outperform independent individuals (i.e., the relative performance of teams versus individuals) [48,219,190].…”

“…task (see Sections 2.2.1, 2.2.1), we determine how, or if, the relative importance of different attributes changes with task difficulty (e.g., does being in a team, or the level of social perceptiveness, or skill diversity, matter more for the most difficult tasks than for easy tasks? ).2.The separation of the experiment into two "phases" eliminates confounding between individual and team measures of performance[48]. In phase one we measure all relevant attributes for individual workers (see Section 2.2.2); then in phase two we use this information to construct teams with desired combinations of individual attributes (e.g.…”

Neural Correlates of Group Versus Individual Problem Solving Revealed by fMRI

“…A seminal study by Woolley et al (2010) suggested that small groups may be treated as collective subjects with their own level of collective intelligence, and that this level does not correlate with either the average or the best individual intelligence of the group members, but may be predicted by gender, average social sensitivity, and in-group discussion style. As the replicability of this phenomenon is still debated (Bates and Gupta, 2017;Credé and Howardson, 2017) the question arises of what conditions are necessary for the collective intelligence to emerge, and whether it can be revealed in computer-directed communication or via face-to-face interaction only (Engel, 2015; Barlow and Dennis, 2016).…”

Adolescent’s Collective Intelligence: Empirical Evidence in Real and Online Classmates Groups

“…Although there is a lot of empirical evidence about the CI general factor, some scholars have tried to resize its effects on groups' problem-solving performance [7,8]. In particular, a recent re-analysis of the four main empirical studies in the field of CI [4,[9][10][11] does not support the hypothesis of a general factor able to explain the performance variation across a wide variety of group-based tasks [8]. Studies carried out in the online environment show that it manifests differently depending on context [9].…”