Barsade (2002) examined themes of that kind in a laboratory experiment. In leaderless group discussions, students talked about supposed candidates for a merit bonus. In each group a research confederate, who role was not known to participants, initiated the presentations. The confederate had been trained to act in either a pleasant or unpleasant manner, for example, being either warm, happy, and optimistic, or hostile, pessimistic, and impatient. In addition to self-reports of affect from group members, observer rating were also obtained from video-recordings of each discussion. Barsade (2002) found that the mood of participants exposed to the pleasant confederate became more positive over time, whereas other participants’ mood became less positive in interaction with a confederate who behaved unpleasantly. (The group discussion lasted for 30 minutes.) An impact of the experimental manipulation was also confirmed through the video coders’ assessments. Barsade concluded that “group members experience moods at work, those moods ripple out and, in the process, influence … other group members’ emotions” (p. 670).