The basic question is whether the performance of a team can be predicted from a 15-minute observation, a “thin slice” of the team resolving a conflict. The answer appears to be “yes.” A follow-on study, again based on a short observation, teased out particularly harmful individual behaviors. The approach is based on a highly effective predictor of whether a couple will stay married [1].
The research used video recordings of an engineering design team working through a significant design disagreement. After the discussion/conflict, each team member was shown the video and asked to record her/his emotional state, positive or negative, at particular moments of the discussion. The predictor is the arithmetic difference between the amount of time the state was positive and the state was negative; 35 percent of team performance, measured 2.5 months after the experiment, was predicted by this difference. (Team performance was evaluated through a self-report survey.) In a second study, the videos were analyzed looking at facial expressions, tone of voice, verbal content, and body posture. Hostile behavior, gleaned from the video analyses, was found to be highly correlated (negatively) with measures of team performance.
This paper is recommended because the approach is novel and promising, although it is unnecessarily long. Also, the word “affect” is used with a particular meaning that needs clarification.