ON THE SOCIAL INELUENCE OF EMOTIONS IN GROUPS 273 d (). (square-root transformed to for ity)was 1994 swho ha ype of ed the instruction to show the sweeosniedectdopornatrdC -2.14: -1.41,or to show no emotion, -152,95%C .171.Sim participants who h cted other d (each cod ℃B--0.79. majority happy(and 5%C-12 d only on the focal participants in the remaining ed to d idnotrepothaN ing been me tion Acceptance/rejection.To test the hypothesis that focal par for rationality:5.08 <M<5.75. 1.41 n the majority happy condition,we regressed felt rejection tion conditions using the ed to sho sin the than 95%C1-1.92:-0.91.or ocal participants in th nal ( 06 SD 2.08.SD=0.62)an 10 :0501 Pa .88,9 the latter twc ad Relative influenc To test whether focal 28.95%C-073:020 had less influence in their groups thar sful in the maic ideas dis ed earlier.Because all betv en-g happy were less clear-cu a thes andom I ed to ow happir eir peers d山 o percei ants in the ss no em accord ng to on as the focal participants in the maio group task than the ence of 056.Sp oecially those int unfamiliar ople.the default -0.36.1=-151. =068 is to be nice and friendly and to smile (e.g Hess S Eresed Emations Accordine to Sel Renor and Pe Condition:Maiority angry Maiority happy Instruction:Angcr Instruction:No cmotion Instruction:Happiness Instruction:No emotior Mesure of anger 2888 18 留8 1H883 8阁 Peer repon 48 4(075) 5a9 40Manipulation checks. First, the self-report manipulation checks were examined (for means and standard deviations of expressed emotion, see Table 2). When self-reported anger (square-root transformed to correct for nonnormality) was re￾gressed on the type of instruction, the results showed that partic￾ipants who had received the instruction to show anger during the group task reported having shown more anger than participants who had been instructed to show happiness,
1.66, 95% CI [2.14; 1.24], or to show no emotion,
1.52, 95% CI [1.96; 1.17]. Similarly, participants who had been instructed to show happiness reported having shown more happiness than par￾ticipants who had been instructed to show anger,
0.79, 95% CI [1.36; 0.22], or no emotion,
0.80, 95% CI: [1.25; 0.34]. Participants who had been instructed to show no emotion did not report having been more rational than participants who had been instructed to show anger,
0.26, 95% CI [0.81; 0.24], or happiness,
0.47, 95% CI [0.93; 0.03]; all Ms and SDs for rationality: 5.08  M  5.75, 0.81  SD  1.41. The peer-reported manipulation checks showed the same pattern for anger but not for happiness (Ms and SDs in Table 2). Partici￾pants who had been instructed to show anger were reported to have shown more anger than participants who had been instructed to show happiness,
1.41, 95% CI [1.92; 0.91], or no emotion,
1.22, 95% CI [1.64; 0.75]. Participants who had been instructed to show happiness were reported to have shown more happiness than participants who had been instructed to show anger,
0.79, 95% CI [1.34; 0.19], but not compared with those who had been instructed to show no emotion,
0.28, 95% CI: [0.73; 0.20]. Together, the manipulation checks indicate that the manipula￾tion was successful in the majority angry condition, but the results regarding the majority happy condition were less clear-cut. Al￾though the self-report measures indicate that participants who had been instructed to show happiness indeed showed more happiness than all other participants, their peers did not perceive them to be showing significantly more happiness than the participants who had been instructed to express no emotions. Thus, according to the peer reports, participants in the nonemotional condition faced the same situation as the focal participants in the majority happy condition. The absence of a difference may be explained by implicit social etiquette or display rules. In most social interac￾tions, especially those involving unfamiliar people, the default behavior is to be nice and friendly and to smile (e.g., Hess & Bourgeois, 2010; Hinsz & Tomhave, 1991). Although we told the control participants to show no emotion, they might have smiled in order to affiliate rather than to show happiness (see Fridlund, 1994). Analytical strategy. To reflect the finding that, in terms of expressed emotions, the majority happy and nonemotional condi￾tions were so similar, we decided to adapt our analytical strategy accordingly and used specific contrasts to test our hypotheses. We first describe the effects in the majority angry (coded as 1) con￾dition versus the other two conditions combined (each coded as 0.5), and then report the (orthogonal) contrast between the majority happy (1) and nonemotional condition (1). Further￾more, we focused only on the focal participants in the remaining analyses. Acceptance/rejection. To test the hypothesis that focal par￾ticipants would feel more rejected in the majority angry condition than in the other conditions and less rejected (i.e., more accepted) in the majority happy condition, we regressed felt rejection (square-root transformed to correct for nonnormality) on the emo￾tion conditions using the contrast coding described previously. The results showed that the focal participants in the majority angry condition (M
3.12, SD
1.06) had felt more rejected than the focal participants in the nonemotional (M
2.08, SD
0.62) and majority happy (M
1.96, SD
0.71) conditions,
0.88, 95% CI [lower limit: 0.50]. Participants in the latter two conditions did not differ in terms of felt rejection,
0.09, 95% CI [0.17; 0.35]. Relative influence. To test whether focal participants in the majority angry condition had less influence in their groups than focal participants in the other conditions, we examined the ratio of contributed ideas discussed earlier. Because all between-group variance is accounted for by these ratios, the random intercept for group had no remaining variance (i.e., the variance equaled 0) and was therefore dropped from the analysis. Thus, a regular regres￾sion analysis was conducted. The means indicate that, as expected, focal participants in the majority angry condition (M
0.44, SD
0.24) had less influence than their fellow group members, as indicated by a mean lower than 0.5. Furthermore, they had less influence in their group task than the focal participants in the majority happy (M
0.56, SD
0.35) and nonemotional (M
0.53, SD
0.24) conditions. This differ￾ence was marginally significant,
0.36, t
1.51, p
.068, one-tailed. The difference between the majority angry and majority Table 2 Means (and Standard Deviations) for Expressed Emotions According to Self-Report and Peer-Report (Study 4) Variable Condition: Majority angry Condition: Majority nonemotional Majority happy Instruction: Anger Instruction: No emotion Instruction: No emotion Instruction: Happiness Instruction: No emotion Measure of anger Self-report 3.08
(1.56) 1.65 (0.88) 1.27 (0.50) 1.18 (0.38) 1.21 (0.33) Peer report 2.51
(1.58) 1.23 (0.32) 1.27 (0.37) 1.17 (0.36) 1.15 (0.19) Measure of happiness Self-report 4.15
(1.49) 3.30 (1.42) 4.38 (1.19) 5.24
(1.08) 4.25 (1.41) Peer report 4.17 (1.14) 4.22 (1.04) 4.77 (0.75) 4.94 (0.95) 4.81 (0.92) Note. Within each row, bold-faced means should be the highest for a successful manipulation.
Mean is higher than other means in the same row, based on a 95% confidence interval. This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. ON THE SOCIAL INFLUENCE OF EMOTIONS IN GROUPS 273