CIVICA Applied Data Science Seminar Series

Session 4 - Spring 2023

Session 4 - Spring 2023 will take place on Wednesday, 4:00 PM CET / 3:00 PM GMT / 11:00 AM ET 22 March, on Zoom Meeting

About The Series

The CIVICA Data Science Seminar series is a unique multi-disciplinary series focused on applications and methodologies of data science for the social, political, and economic world.

Session 4 Spring 2023 Details

Given that emotional engagement occurs prior to conscious awareness, it is difficult to measure using self-reports, leading scholars to rely on physiological measures. Unfortunately, these measures often must be collected in laboratories using convenience samples, leading some to question their external validity. Nonverbal paradata could provide a potential remedy.

In this session of the CIVICA Data Science Seminar Series, Prof. Bryce Jensen Dietrich, Associate Professor at Purdue University will show that nonverbal paradata can effectively measure emotional engagement, something that is predictive of vote choice. In doing so, this talk will also show how multimodal neural networks, specifically those using text, audio and video data, can be used to better understand information processing. Neural networks mimic the human brain, meaning the underlying weight structure could yield insights into how people internally process audiovisual information, among other things. To underline this latter point, this talk will also show ways in which class activation maps can be used for inferential in additional to descriptive purposes.

Seminar Speaker

Prof. Bryce Jensen Dietrich
Prof. Bryce Jensen Dietrich

Bryce Jensen Dietrich is currently an Associate Professor of Political Science at Purdue University. He is also a Research Scholar at the Center for C-SPAN Scholarship & Engagement. Previously, he was an Assistant Professor of Social Science Informatics at the University of Iowa and a postdoctoral research fellow at Harvard's Kennedy School and Northeastern University. His research uses novel quantitative, automated, and machine learning methods to analyze non-traditional data sources such as audio (or speech) data and video data. He uses these methods to understand the causes and consequences of non-verbal political behavior, such as vocal inflections and walking trajectories, especially in relation to descriptive representation and implicit gender/racial bias.


Dr. Erica Thompson

Welcome Introduction Dr. Erica Thompson, LSE

Setting the scene: Brief intro to the speaker and his talk

Prof. Bryce Jensen Dietrich

Seminar Session Prof. Bryce Jensen Dietrich, Purdue University

Using Multimodal Neural Networks to Better Understand How Voters Process Audiovisual Information

Research Discussion. Lead Institution

Q&A / Discussion on the research


Upcoming seminar in the series and other announcements

Register here