Content provided in partnership with
Business Services Industry
The ripple effect: emotional contagion and its influence on group behavior
Administrative Science Quarterly, Dec, 2002 by Sigal G. Barsade
There were two sets of analyses for each hypothesis, one set conducted with individual-level mood rated by video-coders as the dependent variable and the other with differences in self-reported pre-and post-experiment moods. The analyses were primarily conducted with two-level models. For each group, parameters describing the individual-level phenomena (i.e., means and covariances) were estimated, and group-level differences among these parameters were then analyzed. The basic individual-level (level 1) model was
[y.sub.ij] = [[beta].sub.0j] + [[beta].sub.1j] + [r.sub.ij].
- Most Popular Articles in Business
- Research and Markets : Tesco Plc - SWOT Framework Analysis
- Do Us a Flavor - Ben & Jerry's Issues a Call for Euphoric New Flavors
- eBay made easy: ready to start an eBay business? These 5 simple steps will ...
- Katrina's lawsuit surge: a legal battle to force insurers to pay for flood ...
- Wal-Mart's newest distribution center opened last month near the southwest ...
- More »
In this model, [y.sub.ij] is a measure of individual level mood as rated by video-coders (or self-reported change in mood) for person i in group j; [[beta].sub.0j] is a random coefficient representing emotional contagion (operationalized as the video-coders' ratings of individual-level moods, in the video-coder models, and self-reported change in individual moods, in the self-report models) of people in group j (across the i persons in the group); [[beta].sub.1j]. is also a random coefficient representing the mean of pre-experimental moods of people in group j (across the i persons in the group), important as a covariate controlling for participants' mood before entering the experiment; [r.sub.ij] represents the error associated with the mood measure; and the variance of [r.sub.ij] constitutes the individual-level residual (or error) variance. No additional covariates were found to be significantly related to individual-level mood, or mood change, but had there been, they would have been included at the individual level by including additional terms on the right-hand side of this equation (e.g., [[beta].sub.2j], [[beta].sub.3j], etc.). All covariates should initially be modeled as random effects, and fixed effects should be used only when the random error term cannot be estimated reliably (Nezlek, 2001).
In multilevel modeling, coefficients from one level of analysis are passed on to the next. As such, in the two-level models, group differences in individual-level mood as rated by video-coders (or change in mood for self-report data) were analyzed at a group level (level 2). The group-level model was:
[[beta].sub.0j] = [[gamma].sub.01](C1) + [[gamma].sub.02](C2) + [[gamma].sub.03](C3) + [[gamma].sub.04](C4) + [u.sub.0j].
In this model, experimental conditions were represented by four dummy-coded (0, 1) variables. C1 was coded 1 for high-pleasant/high-energy condition groups and 0 for the other three conditions, C2 was coded 1 for the high-pleasant/low-energy condition, C3 was the high-unpleasant/high-energy condition, and C4 was the high-unpleasant/low-energy condition, and [u.sub.0j] represented the error of [[beta].sub.0j]. Differences among the groups were examined using comparisons of fixed effects (Bryk and Raudenbush, 1992: 49-52). For example, C1 and C2 represented the two pleasant-affect groups, and C3 and C4 represented the two unpleasant-affect groups, and so the "main effect" for valence of affect was examined using a contrast code of 1, 1, -1, -1. These zero-intercept, dummy-coded analyses provided the functional equivalent of the comparisons provided by a traditional ANOVA while retaining the benefits of MRCM.
Find Research Guides for:
Most Popular Articles in Business
- Research and Markets : Tesco Plc - SWOT Framework Analysis
- Do Us a Flavor - Ben & Jerry's Issues a Call for Euphoric New Flavors
- eBay made easy: ready to start an eBay business? These 5 simple steps will ...
- Katrina's lawsuit surge: a legal battle to force insurers to pay for flood ...
- Wal-Mart's newest distribution center opened last month near the southwest ...
- More »