Measuring Emotional Responses to Negative Commercials

• In Event: Political Advertising

Abstract

We are interested in examining people’s emotional responses to negative advertising. We provide two potential advances to understanding how people react emotionally to negative campaign messages. First, we introduce a theoretical framework based on the notion that people hold predispositions regarding how well they tolerate negative attacks in politics. We have shown in prior work that people do indeed have different levels of tolerance to negative information and this variability affects how citizens react to candidates and political message (Fridkin and Kenney, 2019). We expect, then, that people’s tolerance for negativity will influence their emotional reactions to negative advertising. Second, we introduce a multimethod research design that includes a new methodology to test our theoretical expectations. We believe the combination of new theory with a strong research design will nudge the literature forward and enlighten our understanding of the growing importance of how emotions are related to politics.

The research design involved the collection of three types of data. First, we created a module for the 2018 CCES study where a national representative sample of respondents viewed two negative advertisements airing during the campaign. These advertisements were chosen because they did not reference a specific candidate; therefore, they were relevant for respondents witnessing all senate races in 2018. Respondents were asked a series of questions about the advertisements, including their emotional reaction to each advertisement (e.g., the advertisement made me feel very angry). Second, we commissioned an internet survey of Arizona residents where people viewed these same commercials as well as several advertisements airing in the 2018 Arizona Senate race. The Arizona respondents also answered questions about the advertisements, including their emotional responses to the messages.

Finally, we conducted a focus group of approximately 100 adults (non-students) who watched the generic and Arizona advertisements in an experimental laboratory. These subjects answered the same survey questions about the commercials asked of the two survey sample respondents. In addition, while these subjects watched the commercials on laptop computers, their facial expressions were recorded via the webcam on the computer. We then measured their expressed emotion with facial expression software. The software (i.e., Emotient FACET) allows assessment of emotional reactions in an unobtrusive and reliable way. Emotient FACET tracks 20 action units (e.g., nose wrinkle) in real time, allowing the extraction of 7 basic emotions (i.e., joy, anger, surprise, fear, contempt, sadness, disgust).

Most studies of emotions in political psychology rely on self-reports, such as the American National Election Studies (ANES) survey questions assessing people’s emotional reactions to the major party candidates (e.g., how angry did the advertisement make you feel). However, self-reported measures have limitations: (1) people are not always conscious of the emotions they are experiencing; (2) respondent’s memory of their emotional state is often fraught with errors; (3) individuals may be hesitant about accurately reporting their emotional reactions. By comparing self-reported measures with the Emotient FACET data, we can examine similarities and differences across these alternative measures. Furthermore, the two measures of emotions, taken across three distinct samples of respondents, with variability in the Arizona and national advertisements, provides us analytical leverage to examine how people’s tolerance of negativity influences their emotional reactions to negative advertisements.

Authors

• Kim L. Fridkin, Arizona State University

• Patrick Kenney, Arizona State University