A recent paper by Goldenberg, Weisz, Sweeny, Cikara, and Gross (2021) suggests that we might be not so good at estimating the emotion in crowds. Their research suggests that we are focusing mainly on the expressions with the highest intensity in a crowd and this might lead to overestimating how a crowd might react.
This is quite interesting - and particularly at a time when, at the time of writing, we might frequently be confronted with a situation where we are talking, while looking at grids of faces looking back at us in a Zoom meeting, or an online class. So the question is how good we are taking the affective temperature in a physical or virtual room.
However, Goldenberg and colleagues studied this question using arrays of a static face (the same face) with different expressions that were presented to participants in three different experiments.
While I believe this is interesting and timely research, I worry a bit that it is difficult to translate these findings into the real world. As I have stated elsewhere, I look at interaction in real-time a bit as if it were a dance, where the dynamic entrainment of motions is an important element of understanding how a crowd reacts - whether there is brow movement when I say something controversial or difficult, whether there is chuckling when I tell a joke, or whether there is no response to a controversial statement. In this context a good speaker is setting up an expectation as to how a crowd reacts - this is why a script for a speech sometimes contains markers to make a pause to enable applause. In that sense, the expectation becomes a feed-forward to a perceptual process that is then sensitive to deviations from that expectations - this is when, if possible, there is time for the speaker to modulate the presentation - take more time, rephrase, emphasize. It would be fun to create such environments and test then experimentally, also with an eye-tracker, as in the third study of Goldenberg et al whether there is evidence not only for the effect of very intense reactions, but also of low expressivity in the context of expecting a response that does not happen.
Goldenberg, A., Weisz, E., Sweeny, T. D., Cikara, M., & Gross, J. J. (2021). The Crowd-Emotion-Amplification Effect. Psychological Science, 32(3), 437–450. https://doi.org/10.1177/0956797620970561
A couple of my papers dealing with the dynamics of interaction
Kappas, A. (2013). Social regulation of emotion: Messy layers. Frontiers in Psychology, 4, Article 51. (link)
Kappas, A., & Descôteaux, J. (2003). Of butterflies and roaring thunder: Nonverbal communication in interaction and regulation of emotion. In Philippot, P., Coats, E.J., & Feldman, R.S. (Eds.) Nonverbal behavior in clinical settings (pp. 45-74). New York: Oxford University Press.
Arvid Kappas is Professor of Psychology at Jacobs University Bremen. He has been conducting research on emotions for over three decades in the US, Canada, and in several European countries.