Individual and group performance compared: when is the group more efficient?

Who, finding themselves on a team assigned to solve a problem, has not thought at least once about how unproductive the group approach can be?

So when does group work pay off, and when should we prefer the individual approach (judging by productivity alone and leaving aside continuity, organizational advantages, and control over people and results)?

Many researchers have worked on this question:

Floyd Henry Allport reached this result in 1920: “the social influence was found to improve the quantity but not the quality of the mental performance.” His study measures how efficiently people perform alone and alongside a colleague or friend, and his conclusions are that working in a group facilitates performance on simpler tasks and hinders it on more complex ones (http://www.brocku.ca/MeadProject/Allport/Allport_1920a.html).

Robert Zajonc (Activation theory, 1965) noted that the presence of other animals of the same species creates arousal that raises performance on well-learned or habitual activities, while lowering it on non-habitual behaviors and reducing the emergence of new responses. The article, titled “Social Facilitation,” appeared in “Science” in July 1965.

Monteil and Huguet (1993) studied the effect of performing a task at the same time as another person who performs it more slowly than the participant, at the same speed, or faster. The results indicate that participants engage in social comparison with the other person and that this comparison creates distraction.

Many other authors have found this effect too, not only in tasks requiring mental work but also where physical effort is involved. The Ringelmann effect, observed by the agricultural engineer Maximilien Ringelmann in rope-pulling experiments (1913), is the tendency for individual members of a group to become less and less productive as the size of the group increases (http://en.wikipedia.org/wiki/Ringelmann_effect). Later research speaks of “social loafing”: a meta-analysis of 78 studies judges it a robust phenomenon that is found across different tasks and populations (Karau and Williams, 1993).

It seems, then, that on simpler, familiar tasks group work can bring an increase in performance, which tends to deteriorate as the activity becomes more complex. A meta-analysis of 241 studies points in the same direction but scales down the size of the effect: the presence of others increases speed on simple tasks and reduces it on complex ones, yet it accounts for only 0.3% to 3% of the variance (Bond and Titus, 1983). More recent studies show, however, that the benefit of the group returns in tasks that are more complex still: in tasks that individuals cannot solve on their own, the participant is motivated by the meaningfulness of the task and by the importance of their contribution to the group. In this last case, though, every participant must be essential and the number must be very small.

In creative activities, however, we tend to expect the group to have a positive impact.

Brainstorming is often regarded as the best way to produce creative answers. But are we sure it is really efficient?

Empirical research comparing brainstorming groups with nominal groups (groups of individuals who generate ideas without interacting with one another) indicates that brainstorming groups are an inefficient means of generating ideas. Individuals working separately and without conferring therefore produce better results in the creative sphere as well. This is confirmed by the meta-analysis of Mullen, Johnson, and Salas (1991): brainstorming groups are significantly less productive than nominal groups, in both the quantity and the quality of ideas, and the loss grows with the size of the group. It is possible, however, to improve the performance of a brainstorming session by using a hybrid technique (electronic brainstorming), in which direct interaction is reduced by having it mediated by computer.

(Alberto Viotto)