Skip to main content

Art and Democracy

In partnership with the Johnson Museum of Art, the Center for Global Democracy at Cornell University facilitated the creation of a democracy-learning art exhibit to help prompt new thinking and discussion on political events and processes. Across ten learning sessions, 45 Cornell students participated in viewing, learning about, and discussing the pieces. The exhibit featured topics from political revolutions around the world, political leaders, civil rights, reflection on identity in America, the roles of nationalism and patriotism, as well as authoritarianism and violence.

Roger Shimomura's "Desert City," 2010. Acrylic on canvas.
Roger Shimomura’s “Desert City,” 2010. Acrylic on canvas.

We studied the effect of the learning experience through a short questionnaire given to one group before the discussion and our second set of participants after the discussion. For a comparative baseline, we create an online sample of university students and apply a matching technique to produce a similar sample of students. 

Our findings confirm, even through this small sample, that attending the session and viewing the democracy-related art strengthens support of and commitment to democracy, particularly the rule of law and the need for political opposition.

This image shows three points with confidence intervals representing the averages for three groups in the study - Control, Treatment 1, and Treatment 2. The y-axis reads Satisfaction with Democracy. The key findings is that Treatment 1 groups members have lower average satisfaction score (0.30) compared to Treatment 2 group members (0.42), a statistically significant difference.

We also observe a significant difference between those measured after only viewing the art and those who viewed and then discussed the pieces. In the latter group, we find average lower degrees of negative partisanship, in particular, students being less sympathetic to the view that rival partisans are a threat to the country.

This image shows three points with confidence intervals representing the averages for three groups in the study - Control, Treatment 1, and Treatment 2. The y-axis reads Negative Partisanship. The key findings is that both the Control group and Treatment 1 group averages are higher, 0.38 and 0.37 respectively, than the average score of Treatment 2 members at 0.24. The Treatment 2 average is statistically significantly lower than the other two groups.