The Arts Engagement Project is a study of ~4000 undergraduate students at the University of Michigan that asked questions about the impacts, precursors, barriers, frequency, and perceptions of co-curricular arts engagement in college.
Drawing from a subset of open-ended questions and responses, we created ontological topic models for each question and an interactive decision support tool to facilitate collaborative interpretation of the topics. We also measured a variety of linguistic and psychographic factors associated with the responses using three different dictionary-based tools. We then combined these measures with the measures of topic prevalence in a principal components analysis to explore the relationships between the topics, tone, and structure of the responses.
The results — particularly when coupled with an analysis of their relationship the demographic profiles of the respondents — provided important insights for the design of arts-based experiences in higher education, as well as the impacts of the arts on student learning and engagement.
The Arts Engagement Project site demonstrates new, interactive, tree-based visualizations of the topics that emerged from the study to facilitate semantic zooming between the topic prevalences and source responses (i.e. the data). The site is also an early demonstration of Jupyter Book, an open source project for building beautiful, publication-quality books and documents from computational material.