The ripple effect: emotional contagion and its influence on group behavior

Administrative Science Quarterly,  Dec, 2002  by Sigal G. Barsade

• E-mail
• Print
• Link

RESULTS

Manipulation check. As shown in table 1, analyses of participants' perceptions of the confederate showed that the confederate successfully enacted the affective behavior required for each experimental condition. Participants who were with the pleasant confederate perceived the confederate as more pleasant than participants who were with the unpleasant confederate [overall means = 6.59 versus 3.89; [chi square](1) = 42.67, p < .001]. Moreover, there was no significant effect of the confederate's energy level on ratings of his pleasantness, nor was there an interaction of energy and pleasantness on these ratings. With regard to energy, participants who were with the high-energy confederate perceived the confederate as more energetic than participants who were with the low-energy confederate [means = 7.68 versus 3.27; [chi square](1) = 152.52, p < .001]. Although there was no main effect for pleasantness in the analysis of confederate energy, there was an interaction of pleasantness and energy such that both differences were significant: the difference between the high-and low-energy confederate was greater when the confederate was unpleasant (8.38 vs. 2.65) than when the confederate was pleasant (6.98 vs. 4.14). Because this difference in perceptions of the confederate's energy was unexpected, I controlled for it in all of the hypothesis-testing analyses by including it as a covariate at the individual level, and it did not change the results. Table 2 reports the means and standard deviations of each of the variables and their correlations.

Emotional contagion. (3) Hypothesis 1 examined whether emotional contagion would occur in the groups at the individual level and at the group level. Video-coder ratings of participants' Time 2 mood and participants' self-reported change in mood were used to operationalize participants' emotional contagion.

Hypothesis 1 was first tested by comparing video-coders' ratings of participants' Time 2 pleasant mood across the experimental conditions. Participants were not video-taped before beginning the experiment (Time 1), so change scores could not be analyzed; nonetheless, because participants were randomly assigned to experimental conditions, differences in Time 2 video-coders' ratings of participants' pleasant mood can be inferred to represent differences due to the experimental manipulations. To be more conservative, preexisting differences in Time 1 self-reported pleasant mood was included as a covariate at the individual level (none of the demographic or task control variables were significant covariates). Supporting hypothesis 1, this analysis showed a main effect of confederate pleasantness on ratings of participants' pleasant mood, as shown in table 3. Video-coders rated the mood of participants who were with a pleasant confederate as more positive than the mood of participants who were with a negative confederate (means = 2.75 versus 2.33). Neither confederate energy level nor the interaction of confederate pleasantness and energy significantly influenced participants' displayed pleasant mood.