The National Election Pool (ABC, CBS, CNN and NBC) and Edison Research will once again conduct the Exit Poll of record in the United States. Using methodology that has provided reliable estimates and served journalists and scholars for more than a generation, we will interview voters across the country after they cast their ballots in this year’s critical midterm elections in the largest single-day data collection effort in the U.S. after the Census.
Our Exit Poll team will interview voters at hundreds of polling locations in all 50 states. The in-person interviews that form the core of our data guarantee that our polls and our projections reflect the views of actual voters nationwide. No poll conducted solely by telephone or online can ever promise that level of accuracy. Americans who tune in to our broadcasts or look at the poll results on our websites can be sure that they are seeing an accurate result that includes the opinions of people just like them.
The Exit Poll’s methodology and question wording is always being refined and edited to maintain the highest methodological standards and ensure that our practices are in keeping with the best practices of the survey research industry. Here are some of the new methodological efforts we will be implementing for this year’s Exit Poll.
Americans are increasingly voting early and dropping off absentee ballots in advance of Election Day. The Exit Poll incorporates these voters in two ways: by interviewing early and absentee voters by telephone in the final days of the campaign as the NEP has done since 2004, and in a new way this year, by deploying in-person interviewers to early polling locations in the days before the election in key states.
The results of the Exit Poll in 2016 demonstrated that knowing a voter’s educational background is critical to understanding American politics today. To ensure that the Exit Poll continues to accurately reflect the views of voters by their educational background, we have modified the question wording to more closely mirror the wording used by the Census Bureau and make it easier for voters taking the polls to accurately categorize their educational background and to better facilitate comparisons to Census data for later scholarship.
Changes to question wording on the Exit Poll are common. In the entire history of modern exit polling conducted by members of the National Election Pool dating back to 1972, just one question has been asked with the same question wording over that entire time.
The NEP has also added the ability to apply weights to Election Day interviews for voter characteristics that are not easily observable by the interviewers conducting the polls. In the past, exit polls were weighted only for such characteristics, including gender, age and race. Now, the Exit Poll will be adjusted for non-response by education and by age using a parameter developed by comparing past Census estimates of turnout among these groups and past Exit Poll estimates of turnout among these groups. While this adjustment does not meaningfully change the results of the Exit Poll in terms of the voting behavior of various subgroups, it does serve to adjust the poll’s estimates of the age and education makeup of the day’s voters in a way that is a best practice for polling in today’s environment.