Sigal G. Barsade (2002), Administrative Science Quarterly 47(4):644–675 — DOI-10.2307/3094912
Key takeaways
- Mood spreads through a group, and the spread is causal. With a confederate randomly assigned to show pleasant or unpleasant mood, members’ self-reported mood rose by .41 or fell by .26 over a 30-minute negotiation. Blind video-coders, who never saw the confederate, rated members’ mood higher in the pleasant conditions (2.75 against 2.33).
- Contagion shapes what the group then does. The more pleasant mood a member caught, the more cooperative he or she was rated by self and by peers. At group level, coder-rated contagion correlated with coder-rated cooperativeness (r = .44) and negatively with conflict (r = −.42), and groups with more contagion split the bonus money more evenly. In Study 2, without a confederate, group contagion correlated with members’ ratings of cooperation (r = .49) and conflict (r = −.49).
- It goes unnoticed. Members related how pleasantly they behaved to how others acted, but showed “absolutely no awareness” that their mood bore on how effective they judged themselves. They attributed their effectiveness to the money won and their own abilities.
- What did not hold up. Unpleasant mood did not spread more than pleasant mood. Energy cut both ways, marginally: high-energy unpleasantness spread more than low-energy, but low-energy pleasantness spread more than high-energy. The low-energy unpleasant (“depressive”) condition showed no significant contagion.
Standing in the record
Filed on 2026-10-04 at the owner’s request, with four works on emotion in groups and organisations. No anthology topic holds it, and it carries no instruction for machine-learning practice.
It is the batch’s empirical anchor. Von Scheve & Ismer (LIT-696) define collective emotion as convergence in affective responding and treat contagion as one route to it among several (shared appraisals, identity, norms). This paper tests that one route causally in small groups. Study 2’s measure of group contagion, the share of variance in mood due to group membership, rising from .00 to .24, is that definition put into numbers. The two do not compete. Barsade’s question is whether the mechanism exists and what it does. Theirs is how it combines with the others, and whether the result is a “We-mode” emotion. Barsade’s groups were ad hoc and competitive, so they test the I-mode end only.
Sanchez-Burks & Huy (LIT-699), unread, take the other side of the same phenomenon. Contagion is how a group’s emotions come to be shared; emotional aperture is how an observer reads the resulting composition. Barsade’s suggestion that leaders be “emotionally attuned to and influenced by their followers” is the practical point at which the two meet. Her paper cites Ashforth & Humphrey (LIT-703) for treating emotion as a broad label for subjective feeling. It cites Frijda’s “The laws of emotion” (LIT-704) for mood as information about how the group is doing. Both are filed unread in the same batch.
For the record’s emotion theories. The mechanism Barsade relies on, mimicry and then afferent feedback (“emotions can also be produced from the outside in”), is Hatfield, Cacioppo and Rapson’s. It is closer to the bodily-feedback strand of Damasio (LIT-388) than to appraisal; that comparison is the record’s, not Barsade’s. It does not need an appraisal of the event to produce the mood. Where she does appeal to appraisal, contagion is said to transmit “a type of group appraisal of events”. On social ontology, Study 2 is offered as evidence that contagion is “a group phenomenon and not due to the dyadic nature” of the confederate design. That is a claim about level of analysis, not about group subjects, and it is the empirical floor any account of group emotion (Huebner, List) has to stand on.
No relation is declared. It cites neither von Scheve & Ismer, who came later, nor any account in the record that it tests or corrects.