In an effort to better understand cognitive dissonance, my team as part of a course entitled “Psychology of Music” developed and conducted an experiment to study this phenomenon through the use of audiovisual material.
Our goal was to identify and understand some of the major factors that contribute to cognitive dissonance between auditory and visual stimuli.
We began a comprehensive study of background literature, including research in the purely psychological studies of cognitive dissonance and in the more specialized field of music psychology through the lens of audio and visual imagery. Each of these topics of study- cognitive dissonance, visual imagery, auditory stimuli, as well as emotion and engagement responses to stimuli- began to deepen our understanding of the overarching topic we hoped to address. Using these topics of previous research, we began to see some overarching trends emerge. These trends allowed us to better define the factors we particularly wanted to study. Consequently, we designed an experiment comprised a set of stimuli that varied over two features: familiarity and dissonance modification.
Methods & Hypotheses
We selected a set of popular and unpopular video clips and chose three auditory modifications categorized as comedic, dark, and popular for each. We asked that participants rate how well the video and audio “matched” using a sliding scale and asked that they give a pleasure rating after each clip. We hypothesized that, for unfamiliar clips, participants would give a lower match rating to the audio modifications. Previous research suggested that experiment participants could, regardless of familiarity with stimuli, accurately match audio to its video counterpart. So we also expected to see that participants would largely be able to match original audio to its video clip. Lastly, we hypothesized that a difference in rhythm and affect could lower the matched rating between video and modified audio.
We analyzed responses to find similarities in the participants’ reasoning behind rating certain clips as more or less well-matched. We found that, indeed, many participants pointed to the difference in rhythm or affect between video and audio, confirming our hypothesis.
In the open-source software JASP, we submitted our data to a 2x4 Repeated Measures ANOVA. Post-hoc t-tests with hold correction were planned for any significant main effects or interactions. We found that participants could indeed quite accurately match original audio to its video counterpart. We found that the opposite of one of our hypotheses was true: we believed that participants were more likely to rate a lower match if they were familiar with the clip but instead they were more likely to rate the dark and comedic modifications higher if they were familiar with the clip.
Additionally, we found a significant relationship in the participants rating of pleasure to the audio modifications (P-value of <.001) but the relationship between familiarity and pleasure was not significant (P-value of .952).
These results may be useful toward application in the film and music industries, including advertisement, in which audiovisual materials and dissonance between audio and video used could affect the viewer’s perception of the clip.
I have omitted identifying information to protect the privacy of my team members.