Can election forecasting make people better citizens?

February 25, 2026 - Eric Gonzalez Juenke

MSU Political Science Associate Professor Eric Gonzalez Juenke, a Good Authority Fellow, writes that a simple game might make people more interested in campaigns. The original article can be found here. 

Every March, millions of average Americans get together to pick 63 winners for a sport many of them do not understand or follow. March Madness has become a national pastime, even though few people who play it watch any college basketball before the tournament begins.This collective ignorance doesn’t stop Americans from a deep dive into power rankings, statistical forecasts, injury reports, and Cinderella picks from a variety of outlets. The popular bracket game makes the NCAA tournament more interesting to both casual fans and non-fans – and creates social incentives to learn more about college basketball.

This same idea motivates Election Madness, a game I created to help my American politics undergrads become interested in 21 federal elections in 2020. That year, more than 4,000 students from across the country, along with students in France and Iceland, competed to see who could correctly pick the most federal election winners. We repeated this in 2024 with similar numbers of participants.

In March, primaries in several states will kick off the 2026 midterm elections. This semester, I am running a smaller version for my own students. Can a simple pick’em game help players become better citizens? It looks promising.

A healthy democracy requires some basic civic knowledge

Americans don’t know much about their government, the research suggests. Political scientists have understood this for decades, but it is still surprising to see the lack of attention most Americans give to a political system many say they hold sacred. For example, a Pew Research survey in 2023 asked U.S. adults six basic questionsabout the federal government. The average respondent could only answer three questions correctly. An Annenberg Public Policy Center survey in 2025 revealed that about 1 in 3 adults cannot name the three branches of federal government; 1 in 5 cannot name a single first amendment right. And these are some of the most basic questions about the workings of the American government.

These gaps in knowledge can have debilitating effects for individuals – and also for democracies more broadly. Political science research suggests that uninformed and disengaged voters regret their votes more often. And these voters are also more likely to tune out of politics, which helps create a cycle of political apathy and cynicism.

Election innumeracy – and distrust

Americans’ trust in government is at an all-time low, according to survey findings. Part of this distrust is fueled by polarized partisan elites taking advantage of public misperceptions and cynicism. We witnessed the extreme dangers of this political ignorance on Jan. 6, 2021. But we’ve also seen many other meaningful, and less-violent, examples of election surprises. Perversely, the glut of election forecasts in the last 15 years may have contributed to these surprises. That’s because pundits, polls and forecasters convey a false sense of precision, while voters don’t fully understand probability and numbers. Many voters simply do not understand election forecasts and this ignorance affects their feelings about the results. The popularity of subjective betting markets adds a new twist, and could lead to further distrust in polls, election outcomes, and more election conspiracies.

Can games like this help teach basic principles?

Instructors have used games for decades to teach students about American government. Simulations like Model United Nations or simple cooperation games on a computer, for example, allow students to experience political success and failure in a safe environment. Political role playing and other games encourage active learning that leads to more engagement and interest, which can help promote a greater understanding of the material.

Election Madness is built on some basic principles that make it useful for teaching. First, it is simple. Students pick the winners of about 20 elections, hoping to earn points that are similar to March Madness’ seedings. Upset picks can earn more points than picking favorites. There are no other complex rules for students to sort through – this is like Tetris, not chess. The simplicity lowers the barriers to entry, making the game valuable for introductory students and easy for teachers to use.

Second, like March Madness, Election Madness simply generates candidate odds with corresponding points, and the game is adaptable for students and teachers alike. For example, some of my students made their own friendly pools to make the game a little more “interesting.” Some professors used the game to go in-depth on specific campaigns and candidates, creating research opportunities. I use it to teach about polling and forecasts, and to familiarize students with higher-quality resources. The students are engaged because these lessons are no longer abstract. They’re using real information as practical guides to help them with their picks.

Does Election Madness work?

What did students learn? Using a survey, I compared introductory political science students who did not play the game (1,700) to those who did (275). This was not a true experiment because there was no random treatment, but the results are still informative.

Did playing the game correlate with a better understanding of forecast probabilities? Unfortunately, no. Students continued to think of forecasts as predicted electoral vote share (i.e., that a 65% chance, incorrectly, equals 65% of the vote). This was a frustrating finding, but it reinforces the difficulty of thinking about probabilities. One caveat is that I did not devote specific attention to teaching this tricky concept directly in class because I wanted to test their learning without interference. This will change in the fall 2026 version.

There were some positive outcomes, however. More students were likely to think elections make the government “pay attention to what people think” compared to students who did not play (74% vs. 63%, respectively). More of the players were “very” interested in campaigns compared to non-players (41% vs. 33%) and more of the players were likely to “trust public opinion polls” (53% vs. 47%). These correlations are promising, considering that these are self-selected political science students who already share many of these values at different levels of intensity. This fall, my colleague and I will conduct a more rigorous analysis of various learning outcomes during the 2026 midterm elections.

Games like this encourage learning

As we head into March Madness season, it is fascinating to reflect on how a simple bracket game made average Americans go wild for college basketball. This contest creates space for sports networks, betting sites, and office managers to make and nurture communities of learners both big and small.

Games like Election Madness can do the same for politics. Winning requires moving outside of one’s partisan bubble, learning how polls and forecasts actually work, and looking beyond the “cheap talk” of random online fanatics. Learning more about elections and candidates, and doing these activities with friends or within a work community, may enhance deliberative democracy in a world full of noisy misinformation.

In fall 2026, the public will be able to play along with thousands of students and teach themselves about polling, candidates, campaigns and elections, becoming better-informed voters in the process. This type of engagement will not fix America’s political problems, but it might help Americans understand the challenges of democracy better. And that understanding is a crucial first step to becoming a good citizen.

Eric Gonzalez Juenke is a 2025-2026 Good Authority fellow.