Teaching Forecasting Without Teaching Methods

Abstract

Election forecasting has become the centerpiece of media coverage of elections. Yet for all the attention paid to forecasting, public understanding remains low and increasingly distrustful. We can improve citizen knowledge and comprehension and increase student engagement by giving students the opportunity to develop their own election forecast. However, the complex methodology associated with forecasting provides barriers for courses that are not methodologically oriented. In this paper I outline a strategy on how to teach forecasting and have students produce their own without the need of using complex quantitative methods. Students engage in qualitative assessment of quantitative data and develop their own forecast. This pedagogical approach gives students the hands-on learning they need to understand the intuition behind forecasting and increases their comprehension of models of elections and voting behavior.

Keywords:

• Forecasting
• active learning
• methods
• undergraduate education

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 Election forecasting is a robust field of study and takes many different methodological and theoretical tracks. One excellent guide for teaching the ins-and-outs is Lewis-Beck (2005), which divides forecasting into systematic v unsystematic approaches, before and after forecasts, and also provides an evaluation of many specific forecasting styles used in the United States and internationally. Given the emphasis of my project, I focus on quantitative approaches to forecasting, and hone in on two major schools of forecasting: structural, i.e., forecasts based on long term and predictable forces, including incumbency, state of the economy, and presidential approval, among others; and public opinion, based on polling, campaign spending, twitter sentiment, and candidate characteristics, among others (Lewis-Beck and Stegmaier 2014). The general theoretical approach is one of elections being decided my major systemic forces or more immediate campaign and candidate effects. I use these theoretical approaches to examine questions of aggregation (local v national), methodology, and measurement. As a note, while I focus on quantitative forecasting, it is useful and engaging to include a few examples of non-systematic forecasts, for example, the height of candidates and the effect on victory (Judge 2018), along with other scientific approaches outside of your focus (for example, ensemble forecasting (Graefe et al. 2014)). I strongly recommend the PS Forecasting Symposia as excellent readings for students.

2 In off-election years, I have had students also make early day forecasts of the winners of important state-level primaries, especially Senate or Gubernatorial primaries, or of salient ballot initiatives.

3 For multi-candidate primaries, I usually have them use the baseline of 100 divided by the number of candidates as the minimum to win, or the percent of vote share of the previous primary winner. Either of these strategies could be applied to multiparty systems.

4 Thus, my class created the first official forecasts of the 2022 Missouri Senate race. In aggregate, my students produced results that neatly mirrored the top three candidate’s vote share for the incumbent party.

5 I would like to thank an anonymous reviewer for this suggestion; I also suggest, if you are teaching multiple sections of a class, to adopt this strategy and have the students compete across classes as well.

Additional information

Notes on contributors

Debra Leiter

Dr. Debra Leiter is an associate professor with the Department of Political Science at the University of Missouri-Kansas City. She received her BA from the University of California, San Diego, and her PhD from the University of California, Davis. An award-winning researcher and teacher, she studies comparative politics, primarily in Western Europe. Her research examines the intersection of elections, parties, and voting, and how context shapes political decision-making. Her teaching focuses on cross-national institutions and political representation, with a strong emphasis on undergraduate research. She also plays trumpet in Kansas City’s own Mid-America Freedom Band.