Data and Democracy

Chapter 7. Election Forecasting in the Media

Natalie Jackson

Natalie Jackson is Senior Polling Editor at Huffington Post, coordinating the Pollster section of the site. Her primary focus is on polling coverage and methodology, statistical methods, and using polls to forecast elections. Natalie has a PhD in political science from the University of Oklahoma, with heavy emphasis on statistics, survey methodology, and American politics.

Election forecasting has been around for many years in academic circles, but in the last few presidential election cycles it has seen tremendous growth in the media. Networks and other media outlets have gathered polls and made electoral projections for decades, but recently advanced statistical models have become more prominent in elections coverage. In 2008, Nate Silver debuted his FiveThirtyEight blog with his projection that then-Senator Barack Obama would defeat Senator John McCain, and the market for that kind of statistical model has expanded in each subsequent election cycle. In 2014, The Huffington Post, The New York Times, FiveThirtyEight (by then its own media outlet), The Daily Kos, and The Washington Post all had forecast models for the midterm senate races.

The process of election forecasting is complex, but the general public seems to love forecasts. A challenge, however, is getting that audience to understand that a forecast—for example, Candidate A has a 60% chance of winning—is not as simple as it sounds.

The Basic Mechanics of ...