Group Emotions: Cutting the Gordian Knots Concerning Terms, Levels of Analysis, and Processes

Abstract

Research has established that groups are pervaded by feelings. But group emotion research within organizational science has suffered in recent years from a lack of terminological clarity, from a narrow focus on small groups, and from an overemphasis on micro-processes of emotion transmission. We address those problems by reviewing and systematically integrating relevant work conducted not only in organizational science, but also in psychology and sociology. We offer a definition of group emotions and sort the conceptual space along four dimensions: group emotion responses, recognition, regulation, and reiteration. We provide evidence that group emotions occur at all levels of analysis, including levels beyond small work groups. The accounts of group emotion emergence at higher levels of analysis differ substantially between organizational science, psychology, and sociology. We review these accounts—emergence through inclination, interaction, institutionalization, or identification—and then synthesize them into one parsimonious model. The consequences of different group emotions are reviewed and further constructs (including emotional aperture, group emotional intelligence, emotional culture, and emotional climate) are discussed. We end with a call for future research on several neglected group emotion topics including the study of discrete shared emotions, emotions at multiple levels, the effects of social network patterns, and effects on group functioning.

Acknowledgements

We thank Laurie Weingart and James Detert for a very constructive and encouraging review process, and we thank Alice Kexin Liu and Matthew Samson for their invaluable help with the literature analysis.

Notes

1 The anthropomorphism concern is part of a long-running debate in social science that extends beyond group emotion research and focuses on the question of whether individual attributes such as learning, intelligence, and emotions can be legitimately applied to groups. As scholars have noted (Barsade & Gibson, 1998), on one side of the debate is the argument that individual attributes reside in people's brains and can therefore not characterize groups beyond saying that group members individually exhibit those attributes (cf. Allport, 1962). On the other side is the argument that groups hold attributes that go beyond and exist in addition to the individual attributes of group members, and that therefore a group is more than the sum of its parts (cf. Le Bon, 1896; Sandelands & St. Clair, 1993). At present, there is agreement across behavioral, physiological, and neuroscience research concerning the importance of collectively shared representations of the world (i.e. “group mind”) (for a review, see Van Bavel, Hackel, & Xiao, 2014). With respect to group emotions, both sides of the debate are helpful to understand group emotions (Barsade & Gibson, 1998): just as individuals contribute their own individual feelings to groups and thus shape group emotions, group emotions affect the individuals within a group and infuse them with distinct feelings. Group emotions, as opposed to individual emotions, link individuals to something larger than themselves. Groups provide unique arenas in which emotions transcend the individual to the extent that those emotions are brought about by the group as a whole.

2 We initially searched Google Scholar for “collective emotion”. We went through the first 12 pages of results to gather high-impact publications (independent of where they were published) and we compiled the terms from these papers that were commonly used to refer to group emotions. We then applied these terms in a systematic search for relevant papers in high-impact journals in organizational sciences, psychology, and sociology. The literature search was conducted over the summer of 2014. As search terms, we used the following: group emotion, collective emotion, shared emotion, group affect, collective affect, shared affect, group mood, collective mood, shared mood, affective climate, emotional climate, intergroup emotion, intergroup affect, widespread emotion, emotional contagion, emotional atmosphere, affective atmosphere, affective tone, and emotional energy. For the organizational sciences, we systematically searched the following journals: Academy of Management Journal, Academy of Management Review, Administrative Science Quarterly, Journal of Applied Psychology, Journal of Management, Journal of Management Studies, Journal of Organizational Behavior, Journal of Vocational Behavior, Leadership Quarterly, Organization Science, Organization Studies, Organizational Behavior and Human Decision Processes, Personnel Psychology, and Research in Organizational Behavior. For psychology, we searched the following: American Psychologist, Annual Review of Psychology, Emotion, European Journal of Social Psychology, Group Processes and Intergroup Relations, Journal of Experimental Social Psychology, Journal of Personality and Social Psychology, Personality and Social Psychology Bulletin, Personality and Social Psychology Review, Psychological Bulletin, Psychological Methods, Psychological Review, Psychological Science, and Social Psychological and Personality Science. For sociology, we searched the following: American Journal of Political Science, American Journal of Sociology, American Political Science Review, American Sociological Review, Annual Review of Political Science, Annual Review of Sociology, British Journal of Political Science, European Sociological Review, Journal of European Social Policy, Journal of Politics, Political Analysis, Social Forces, Social Science Research, Sociology, and Theory, Culture & Society.

3 The study included groups as big as 189 employees, but the findings reported are based on work groups with 30 or fewer group members because larger groups were thought to split up into subgroups and would thus not be representative for group-level effects. The group size criterion, together with other criteria concerning the completeness of data, led to a reduction of initially 169 work groups to 97 work groups, on which the analyses were based.

4 There is also some evidence that job satisfaction, a construct that often involves an affective component (Weiss & Cropanzano, 1996), can be aggregated to the organizational level. In a study of 298 schools and 13,808 teachers, it was evident that teachers' levels of job satisfaction within a school, but not between different schools, were relatively similar. Job satisfaction was therefore seen as a construct that can be justifiably conceptualized, aggregated to, and analyzed at the organizational level (Ostroff, 1992; see also Patterson, Warr, & West, 2004).

5 Inclination, interaction, institutionalization, and identification are moderators that shape how emotion elicitors affect group emotions, but they can also act as mediators that bring about group emotions. For example, interaction is a moderator, because the emergence of the group-shared emotion depends on the social interaction among people: if there is no social interaction, then there is no group-shared emotion. And if there is social interaction, then the extent and type of interaction may affect the extent to which group-shared emotions converge in response to the event. But social interaction can also be a mediator. An event might attract a group of people to interact, and this interaction among people might in turn create a group-shared emotion. Here, the effect of the event on group-shared emotions is channeled through interaction. Thus, social interaction as a process of group emotion emergence can underpin both mediation and moderation, and the same applies to the other processes of group emotion convergence. The studies from which we extract the four I processes feature both mediation and moderation designs. Scholars have highlighted that mediation and moderation are theoretically intertwined concepts (even though statistically mediation and moderation rely on different procedures). For more on the commonalities and the distinction between mediation and moderation, please refer to Baron and Kenny (1986), and to Spencer, Zanna, and Fong (2005).

6 Scholars have acknowledged that Aristotle's work presaged much of contemporary emotion research. Indeed, Aristotle argued for the adaptive role of emotions in social life and decision-making, foreshadowing the notion of emotional intelligence (Mayer & Salovey, 1997), as well as for the unity of mind and body in the generation of emotions anticipating current thinking about the embodiment of emotion (Barrett & Lindquist, 2008). He also suggested that events trigger emotions (Weiss & Cropanzano, 1996) and that it is the appraisal of the event that shapes the specific emotional experience, along the same lines as the predominant appraisal theories of emotions of the twenty-first century (e.g. Lazarus, 1991; Roseman & Smith, 2001; Scherer, 1988). An aspect of Aristotle's theory of emotion that is often overlooked, however, is that it addresses group emotions in particular. Aristotle presented most of his emotion theory as part of his treatise on Rhetoric which deals with how speakers can exert influence on an audience (i.e. a group of people). Among other means of persuasion (i.e. logos and ethos), a speaker can employ pathos—the appeal to the listeners' group emotions. Aristotle outlines, for a variety of emotions, those events that necessarily elicit appraisals leading to a relevant discrete emotion (e.g. anger is elicited by barring people from a desired outcome). He added that dispositions influence the ease with which specific emotions can be stirred (e.g. people are more likely to respond with anger if they tend to be angry—today, we would say: if they have a neurotic, disagreeable personality). Aristotle suggested that an audience will experience, and converge in, the intended emotion to the extent that it is composed of people with similar dispositions and that the speaker is able to present relevant events in ways that arouse those emotions. Furthermore, he added, foreshadowing the in-group–out-group dynamics of intergroup emotion research in psychology (Smith et al., 2007), that to the extent that the speaker is able to attribute events to a distinct source (e.g. an opponent or out-group), the listeners' collective emotions will be directed toward that source.

7 For further evidence that group-based emotions are actually experienced, not just perceived, see the work by Rydell et al. (2008).

8 In Figure 1, we depict the four I processes as moderators that affect how the different stages of group emotion emergence are linked. Note, however, that this moderation-based logic does not preclude the possibility that the processes mediate the flow of the group emotion process. Accordingly, group membership is likely to facilitate identification with the group, and identification, in turn, is likely to inculcate group-based emotions. Group membership may also trigger the institutionalized regulation of the emotional response, which, in turn, leads to group-norm conforming expressions of group emotions. Group membership, as well as group-based emotional experiences and group-conforming emotional expressions, may be the start of an interaction with other group members that, in turn, produces group-shared emotions. The experience of the group-shared emotions with others may prompt people to consider their dispositional inclination for such experiences and may thus lead them to start, continue or end their group membership. Therefore, moderation concerns whether and to what extent the processes set in (as described in the main text), whereas mediation tells us how the processes link the different stages of group emotion emergence once the processes have set in.

9 The question of which group members' emotional reaction or interpretation of the event becomes the institutionally accepted one, has not been answered. Furthermore, there may be more communication, more sharing, under the condition of uncertainty rather than certainty, but the evidence for such an imbalance is limited.

10 Note that this blend of emotion and solidarity is missing in cases that feature only synchrony in emotional experience and the presence of others, but neither group membership nor interaction. For example, in a movie theatre people also tend to simultaneously experience similar emotions, but these emotions lack the collective quality of group emotions, because people are not joined by a common identification and the interactive sharing of the emotion is prevented by the set-up and dimming of the room.

11 There may be rare conditions when even in mid-size and large groups the interaction part of the group emotion emergence process is enacted by all members simultaneously. For example, above we mentioned Microsoft's annual employee meetings in Safeco Field Stadium in Seattle. These meetings brought tens of thousands of employees together in one place as well as several thousand others who joined in via webcast. The meetings thus involved a group of considerable size—and yet, the meetings were reportedly infused with intense group-shared emotions, based on a common group membership and ritualized interaction.

12 Some see emotional climate and emotional culture as distinct constructs (Ashkanasy, 2003; Ashkanasy, Wilderom, & Peterson, 2011). Our reading of extant research is that emotional culture tends to be defined more in prescriptive terms, specifying how group members ought to feel and which emotions they are implicitly expected to express or suppress, whereas emotional climate appears to be defined as a more descriptive construct, capturing explicit perceptions of how group members have been feeling over a defined period of time. Otherwise, however, there is so much variance, and so much overlap, in explanations of emotional culture and emotional climate that it is difficult to clearly distinguish between the two constructs. In part, this is because the general constructs of organizational culture and organizational climate are so difficult to discern (Denison, 1996; Schneider, Ehrhart, & Macey, 2013). We cannot resolve this debate here. We hold that both emotional culture and emotional climate are in principle useful constructs and we recommend that researchers, when examining these constructs, define clearly how they understand emotional culture and emotional climate, respectively.

13 Retrieved January 6, 2015, from https://www.apple.com/jobs/us/.

14 Retrieved January 6, 2015, from http://www.swatchgroup.com/group_profile/nicolas_g_hayek_message_from_the_founder.