Why College Students Aren’t Well Represented in Voter Polls

Your newsfeed will soon be flooded with polls showing how “likely voters” will cast their ballot in November’s midterm elections.

But there’s a good chance those polls don’t accurately reflect the views of traditionally-aged college students.

Despite historic turnout in recent elections, the voices of Americans aged 18 to 23 aren’t included in political polls because of the methods pollsters use to decide “likely voters,” the methods they use to measure public opinion, and the proportion of young voters compared to the overall voter pool.

Still, it’s important to understand how this demographic is included in polling, as their opinions often drive political discourse and media narratives, Abby Kiesa, deputy director at the Center for Information and Research on Civic Learning and Engagement, told BestColleges.

To understand why students are less likely to be considered in major election-year polls, it’s important to understand how polling companies determine a “likely voter.”

A “likely voter” is a popular polling demographic used closer to elections, according to the American Association for Public Opinion Research. To determine if someone is a “likely voter,” a pollster asks a set of questions specifically aimed at categorizing how likely a person is to vote; those that respond “correctly” to a certain number of questions qualify.

Gallup’s model has been in place since the 1950s and uses the following seven questions to determine whether a respondent is a likely voter:

While the model weighs responses differently for those aged 18-22, the questions are still biased against first-time voters, Kiesa said. For instance, consider the question: “Do you know your voting location?” That’s not a question most first-time voters know until very close to the election, she said.

“Some of these questions are questions that if you’re a brand-new voter, regardless of age, you’re not going to know,” she said.

Brian Schaffner, a pollster and professor of civic studies at Tufts University, told BestColleges that the voting location question is slowly being phased out from most pollsters’ set of questions as mail voting becomes more widespread.

Likewise, the methods that pollsters use to contact the public may also leave out many younger voters, Kiesa added.

Many pollsters still use phone calls to gauge public opinion, she said. Younger people are much less likely to answer a call from an unknown caller than older Americans. Additionally, many pollsters use voter registration records to reach out, which means younger voters who can’t or don’t register until soon before an election won’t get called.

Improving “likely voter” models is a focus of Schaffner’s more recent research at Tufts University.

It’s not necessarily that these models aren’t accurately predicting the percentage of young Americans that will vote in an upcoming election, Schaffner said. In fact, a 2016 Pew Research Center study found that the standard Gallup method of polling accurately predicted that voters aged 18-24 would make up 5% of the electorate.

Instead, Schaffner told BestColleges, the problem is trying to predict which of the 18- to 29-year-olds polled will actually vote, rather than how many of those people will vote.

“One thing that we’ve seen in research is that even when young people say they’re definitely going to vote, a lot of them don’t end up voting,” he said. “Only about half who said they were definitely going to vote actually did.”

One solution Schaffner’s research points to is to assign to each person polled a score assessing their probability of voting. For example, a traditionally college-aged student may receive a 0.4 weighted score, while a 65-year-old who has voted in every election for the past 20 years would be assigned a 0.9 to balance the scale.

Still, the biggest predictor of who is likely to vote is who has voted in the past, he said. That means there will always be a risk that first-time voters are undercounted. However, they make up such a small percentage of the electorate, it usually doesn’t have a major impact on how accurate polls are.

The 2020 election saw historic youth turnout, Schaffner said. However, polls conducted in advance weren’t too far off from final election results. Turnout increased across all demographics, so the youth vote didn’t sway the result any more than what polls predicted.

A review of the 2020 presidential election from the Institute for Democracy & Higher Education found a 66% student voter turnout rate in that election. The national voting rate was about 67%, according to the U.S. Census Bureau, so college students roughly mirrored national trends.

Schaffner’s research shows that people under 25 who are students or recently graduated college are between 6-7 percentage points more likely to actually cast a ballot than others in that age range.

Kiesa stressed that youth voter polls are vital in determining the priorities of generations not regularly represented in larger polls.

These youth polls have shown that issues like climate change, student debt, and diversity and inclusion are top of mind for younger voters. National polls, meanwhile, tend to focus more on issues that older voters prioritize since they are more likely to be counted by pollsters.

A way that some groups have attempted to bridge the gap, Kiesa said, is through panels with the same group of people repeatedly questioned on different topics. These panels mix all ages in a way that accurately portrays U.S. demographics to capture a more accurate snapshot of public opinion.