Substantive, upper division classes in political science provide huge opportunities to connect substantive questions and themes with data to encourage greater data literacy, critical thinking, and understanding of the social science process. For these reasons, I transformed my public opinion and political behavior course to one that focused on learning how to ask and answer theoretically rich and interesting questions about citizen attitudes and decisions using public opinion data.
I have found that while data presented in books, articles, and lectures is helpful to students, requiring them to examine the data directly is a whole new ball game in learning. I present a lot of public opinion data in my public opinion class to make my substantive point. The articles and texts the students read also contain a lot of data graphs and data tables to make their substantive point. But being a passive actor and listening to someone talk about and describe data does not provide the bang for the buck that actively engaging with data does. Students certainly hear me and look at the numbers in the tables and graphs, but when they work with the data they develop a deeper knowledge and understanding of both the questions we are asking and the conclusions. Simply put data are more interesting when they are connected to substantive questions instead of simply examples of methodology. Therefore, the course provides a social science methodological foundation in the principles of science, observation and comparison, in a concrete way connected to an interesting subject, in this case, political behavior.
To accomplish this, I integrate traditional pedagogical approaches– lectures, readings, and discussions– with “data labs” that are designed to reinforce the substantive material. The data labs are meant to teach students how to manipulate data and how to use relatively simple statistical methods to find patterns in data and make meaningful inferences to a population. The assignments build upon one another, starting out very simple. Simple is important because it is easy for students to get overwhelmed and frustrated with issues related to data science including computer programs, computer code, syntax files, output files, data files, reading output, understanding output, writing about and discussing results, and applying observations to the real world. The point of each assignment is to increase understanding of the subject material, while building data analytic skills. Thus, each assignment increases in complexity and requires more student engagement and interaction with data.
Because students enter the class with a knowledge of frequencies and averages I use this took kit as a foundation to observation and comparison on which we can build. While during the first assignment some students get lost and frustrated by the third assignment students begin working independently, and by the final assignment students are telling me they are adding this skill to their resumes.
Beyond data literacy the course has a moral theme threaded throughout about the need for intellectual humility in the political world. The course, therefore, begins with a solid theory of human behavior from which we develop hypotheses to explain the political world. The most important theory to understand attitudes and behaviors at the level of mass politics is the theory of motivated reasoning, which argues that most of the time people are motivated to engage in biased reasoning to get the right answer and consequently that we need to have intellectual humility in the political world. To understand how motivated reasoning impacts human behavior we examine the role partisanship plays in structuring attitudes. For example, we have an entire lab exercise devoted to looking at economic voting which has both an economic and partisan component to it because partisans tend to think the economy is better when their party holds the presidency. We examine this relationship and find that while indeed partisanship plays a role in structuring those attitudes when we control for party we can still find a moderate relationship between evaluations of the national economy and vote for the president.
By the end of the semester the students feel like they have accomplished something; they have worked very hard and learned quite a lot. They understand how partisanship structures American politics and leads to polarization because they read the literature, participated in discussions, and most importantly crunched numbers that showed it. They are also happy because they largely have grades of As and Bs, which they have justly earned through completing and fixing assignments. In addition, the material they covered substantively is more concrete to them. The data assignments reinforced what we read and discussed, promoting a deeper understanding of the material. They also feel they have a better understanding of data, how to manipulate it, how to ask questions using data, and how to read and interpret data results.
Dr. Lonna Atkeson is the LeRoy Collins Eminent Scholar in Civic Education & Political Science and director of the LeRoy Collins Institute. She is a member of the MIT Data and Election Science Board and an Associate Editor for Political Analysis. Her research focuses on election science, survey methodology, public policy, voting rights, public opinion, political psychology, and political behavior. You can learn more about Dr. Atkeson here.