Skip to main content

The Value of Getting It Wrong

This was the year political scientists fumbled in their forecasts. What can we learn from that?

By Seth Masket


Donald Trump speaks during a campaign rally at the American Airlines Center on September 14, 2015, in Dallas, Texas. More than 20,000 tickets were distributed for the event. (Photo: Tom Pennington/Getty Images)

In many ways, 2016 is proving to be the year that political science fell flat on its face in predicting and explaining American presidential nomination contests. Is there some value in having made a very public forecast and gotten it wrong? What can we learn going forward?

Several political scientists wrestled with these questions at a panel at last week’s meeting of the Midwest Political Science Association in Chicago. The University of California-Los Angeles’ Lynn Vavreck opened with a slide showing quotations from dozens of political scientists, journalists, and other political observers from mid-2015 stating confidently that Donald Trump would not be the Republican presidential nominee or come anywhere close to winning that title. Panelists had a bit of fun at the expense of Georgetown University’s Hans Noel, representing the authors of The Party Decides (which describes a very party-centered nomination system biased against wealthy famous celebrities with weak commitments to party goals). But one hardly needed to have subscribed to that book’s framework to have reached the conclusion that Trump’s campaign was going nowhere. A lot of political observers (myself included) were quite certain of that. Parties just simply don’t nominate people like that for president. And the case for Trump last summer was based on the notion that whoever was leading early in the polls would ultimately win the nomination, something which is very rarely true.

Political science, like most academic disciplines, is far better suited to explanation than prediction.

Yet here we are, with Trump almost certain to go to the Republican National Convention this July with a plurality (thought not necessarily a majority) of pledged delegates. And political journalists who had long been skeptical of political scientists’ theories on party nominations are enjoying quite a bit of schadenfreude watching the discipline eat crow. But, in fact, there’s quite a bit to learn from this moment.

For one thing, it shows us that people are actually paying attention to political science today. In 2004, many American politics scholars fumbled in their forecasts of the Democratic presidential nomination contest. They initially underestimated Howard Dean’s appeal and later overestimated his staying power. John Kerry was a plausible and conventional enough nominee, but few saw him emerging with the prize before it happened. That year was a belly flop for many in political science, but the political media largely ignored it. No one cared. The increased engagement of political scientists through blogs and news sources over the past decade has brought political science into the national discussion. When we fail, people notice.

This is, in the long run, good for the discipline. It means we have increased incentive to get it right next time. American political scholars are already engaged in debates and research over just what 2016 means and whether their theories need to be revised or thrown out, or whether Trump is just a bad case for an otherwise pretty sound theory.

It’s also a reminder that political science, like most academic disciplines, is far better suited to explanation than prediction. Very few of our publications offer predictions for the ways political events will unfold. Rather, we’ll occasionally use forecasts to test our understanding of the political world. Political observers have offered many explanations for what’s important in presidential nominations, from money to charisma to debate performances to momentum. Political scientists are trying to systematically approach the question and figure out just what is important and what isn’t. Using this evidence to make a forecast is a good way to test whether the theory is right. And a success or failure of the test is an important update of the theory. Indeed, we learn a lot more from failures.

Finally, this year is raising many important questions that will inform political inquiry for years to come. If political science theories about nominations really are failing in 2016, is this because political scientists never really understood what was going on, or because the political world is really different this year from what it has been before? If it’s the latter, what exactly has changed?

Has social media transformed the way campaigns are run? Is the Republican Party just going through a crisis that has temporarily prevented it from making coordinated decisions? And how will the parties attempt to reform themselves to prevent all this from happening in future years?

This is, in some ways, an embarrassing moment for political science, but it is also one rife with possibilities. We will know a lot more about American politics at the end of 2016 than we did at the beginning. That’s a pretty good year.