James Martherus and Quin Monson, Department of Political Science

Groups like Seeds of Peace that seek to promote conflict resolution often attempt to do so by humanizing an outgroup. For example, to reduce conflict in Israel, I might gather Jewish Israelis and Arab Israelis and have them engage in activities designed to produce mutual empathy between groups. Many academics, including my mentors, have studied how these humanization treatments work among different groups (See Gubler 2014, Gubler, Monson, and Karpowitz 2014) and find that for a subset of people, humanizing treatments fail. My project attempted to determine the characteristics of those individuals for whom humanization fails and why it fails. Using prior research as a guide, I decided to test the effectiveness of humanization treatments by measuring levels of ingroup-centric beliefs and cognitive dissonance. I hypothesized that a) I could create increased levels of humanization by showing members of an ingroup counterstereotypical images of an outgroup, b) Individuals with high levels of ingroup-centric beliefs would react different to the pictures versus individuals with low levels of ingroupcentric beliefs, and c) the framing of the pictures would have a significant effect on how effective the treatment was.

To test these hypotheses, I designed a survey experiment with a control group and three treatment groups. First, respondents in all groups answered a series of questions designed to measure levels of ingroup-centric beliefs. This measure was created by Gubler, who combines items that measure ingroup glorification, outgroup denigration, and victimhood to measure ingroup-centric beliefs. The control group was then shown 10 pictures of scenic waterways in Utah and asked to rate them from 1 to 7. All three treatment groups were shown 10 pictures of Hispanic residents of Utah. The first treatment was told the individuals were simply Hispanic residents of Utah. Treatment 2 was told they would be looking at pictures of illegal immigrants before seeing the pictures. Treatment 3 was told they had looked at pictures of illegal immigrants after seeing the pictures. After rating the pictures, all respondent groups were asked questions designed to measure various factors like humanization, cognitive dissonance, empathy, and policy attitudes. I then ran various analyses using basic OLS regression.

Analysis of this data revealed that the treatments failed to produce significant levels of humanization, thus disproving my first hypothesis. Additionally, there were very few significant differences between the high ICB group and the low ICB group, disproving the second hypothesis. However, there were some differences in how certain individuals reacted to different treatments. For example, when high ICB individuals received treatment one, they were less likely to support policies that harm Hispanic immigrants.

There are a few reasons I believe the treatments did not produce the intended effects. First, some research suggests that a self-reported measure of cognitive dissonance is not very accurate. When we ask people what emotions they are feeling, they tend to skew their answers positively. To remedy this, my mentor and I plan on running a similar experiment where we place respondents in an fMRI and attempt to measure levels of cognitive dissonance by looking at brain activity. Second, I believe our control condition was compromised. Our control was meant to measure levels of humanization and dissonance before any stimuli occurred. However, we had all respondents answer a few questions designed to measure levels of ingroup-centric beliefs at the beginning of the survey. I believe this measure produced some level of cognitive dissonance in respondents. If this is the case, we are comparing our treatments to another treatment rather than comparing treatments to a control. Third, I believe the treatments were not strong enough. We had assumed that simply being asked to look at counter stereotypical photos would create cognitive dissonance for high ICB individuals, but the photos seem to have been too mild. In fact, high ICB individuals rated the pictures about the same as the low ICB individuals.

While the project did not produce the expected results, it was extremely useful. This project is part of a larger set of projects that are meant to help refine certain experimental treatments. The insights I gained from this round of experiments has allowed my mentor and I to create much stronger treatments to be used in the future.