Eunji Kim sparked Chris Wiggins’ interest during an event on artificial intelligence and democracy when she brought a fresh perspective to the discussion of memes and their role in information ecosystems. Reflecting on the event, Wiggins, an associate professor of applied mathematics at Columbia Engineering, remarked, “A string of fellow technologists — including myself — got up and made various technical claims about how technology is useful for understanding the information ecosystem. Then Professor Kim got up and said, in academic terms, none of you has any idea how Americans actually interact with information.”
This moment of academic candor has since led to a collaborative teaching effort between Kim and Wiggins. Together, they are co-teaching a new course at Columbia University titled “Persuasion at Scale: Causal Inference, Machine Learning, and Evidence-Based Understanding of the Information Environment.” The course brings together data science and political science to examine how information influences public opinion and behavior. Both professors are members of Columbia’s Data Science Institute (DSI).
Kim, an assistant professor in political science, highlights a common challenge in computational social science: the risk of drawing incomplete or inaccurate conclusions without considering the political and social context behind the data. “It’s very common for researchers in computational social science to use big data to draw conclusions about society,” Kim explained. “But if you don’t consider political context and meaning that influences your data, the analysis will be incomplete and your conclusions could be wrong.”
The course is designed to bridge this gap by equipping students with the skills to critically analyze the impact of political communication, ranging from campaign advertisements to partisan media and social media content. As Wiggins pointed out, “Persuasion is happening at scale on information platforms, so we now have the chance to understand this question statistically.”
A significant component of the course is machine learning, a branch of artificial intelligence. Students will delve into the role of machine learning in content recommendation and moderation systems, which are fundamental to modern information platforms. Additionally, they will employ these techniques to analyze and interpret complex datasets.
The course blends theoretical insights with practical applications. Students will explore the academic literature on political persuasion while also engaging with statistical methods to study real-world data. By doing so, they will examine the effects of partisan media, social media, advertising, and political campaigns on public opinion. The course also takes a historical perspective on how persuasion strategies have evolved over time.
“When we actually bring data to these questions and look at them objectively, we sometimes find that conventional wisdom isn’t supported — or that it’s wrong,” Kim stated. For instance, there is a widespread belief that Americans are deeply entrenched in partisan echo chambers, where half the population consumes conservative media like Fox News, while the other half tunes in to liberal outlets like MSNBC. However, Kim noted, “If you look at actual behavior-level data, the extent to which echo chambers exist is very limited because most people do not consume news to begin with. Consumption of news content is very low relative to other media, like sports or entertainment.”
Another surprising discovery in political science is the limited impact of political campaigns on voter decisions, even when those campaigns spend enormous sums of money. “There’s a lot of discrepancy between what people believe versus what empirical social science has been discovering,” Kim said.
To address such discrepancies, students will learn causal inference techniques, enabling them to differentiate between correlation and causation when working with real-world data. This methodological rigor will help students draw more accurate and meaningful conclusions.
The course is part of Columbia’s Provost’s Cross-Disciplinary Frontiers Initiative and is open to undergraduates from across the university. With an anticipated enrollment of about 70 students, the professors hope to foster an environment of interdisciplinary collaboration. Kim emphasized the importance of such cross-disciplinary learning, stating, “Engineering students don’t often take classes in political science, and our own social science students do not often take many math classes. These types of classes are critical for them to learn how to fix the many complex problems facing our society.”
For Wiggins, the course represents an opportunity to bring mathematical and statistical precision to a field often dominated by assumptions and anecdotal evidence. “I think it’s useful to zoom out and see how persuasion — whether it’s political persuasion or marketing — has some universal aspects that we can understand using mathematics,” he said. By combining this mathematical framework with historical and contextual insights, the course aims to empower students to move beyond sensational anecdotes and instead adopt a methodical approach to understanding persuasion.
The professors also aim to challenge the conventional wisdom that often pervades discussions of political communication. By applying statistical methods and analyzing real-world data, students will be better equipped to question prevailing assumptions and develop evidence-based conclusions.
Ultimately, “Persuasion at Scale” is more than just a course on political communication; it is a training ground for the next generation of interdisciplinary thinkers. Students from diverse academic backgrounds will learn not only the technical skills required for data analysis but also the critical thinking needed to navigate the complex intersections of technology, media, and politics.
By encouraging collaboration across disciplines, Kim and Wiggins hope to create a new wave of scholars who can tackle pressing societal challenges with a blend of quantitative rigor and contextual understanding. As Wiggins concluded, “By combining that context with the language of probability, we hope to enable students to look past inflammatory anecdotes in order to think methodologically and historically.”