Trump’s “unlikely voter” lead isn’t as worthless as you’d think
A poll has Trump leading Biden among people who say they are not likely to vote. What does that really mean?
Scan the methodology of the latest Suffolk University/USA Today poll and it’s easy to predict the reaction from many pundits.
“Why,” they will ask, “would you bother polling people who say they are not likely to vote?”
It’s a good question. After all, most public political surveys focus on likely voters in an attempt to build a sample that approximates the makeup of the next electorate. That standard operating procedure makes plenty of sense for typical horse race polls that exist to show you the state of play at a given time.
But there’s more to any election than snapshot poll numbers can show by themselves, particularly in an election that 1) is still 14 months away; 2) features two well-known but deeply unpopular frontrunners; and 3) has two frontrunners with age and legal realities that could possibly push either or both out of the race. That’s where the Suffolk/USA Today poll comes in handy.
The headline is that unlikely voters are more likely to support former President Donald Trump than current President Joe Biden, 27.5% to 15% (though Undecided/Other still leads the field). In more traditional national polls of likely voters, Biden sits in virtual ties with almost any leading Republican primary candidate (and even with the non-leading candidates). This includes Trump. Naturally, these numbers come an imagined version of the 2024 electorate.
Given how close these margins are, a smart campaign looking for an edge will go outside this model by recruiting, registering, and mobilizing unlikely voters. And it might work: The Suffolk/USA Today poll shows that about 70% of the unlikely voters who are leaning toward one of the major party candidates claim that they will consider voting if they think it will put their candidate over the top.
In 2020, voter turnout was higher than in any presidential election in at least 40 years — but still at just two-thirds of potential voters. The Census Bureau projected that 40 million eligible voters didn’t cast a vote.
The Trump vs. Biden unlikely voter numbers show that Trump has more numerical potential to increase his raw vote totals in a hypothetical 2024 match-up. It also gives a taste of where we might see a difference between polls leading up to election day and actual returns. In both 2016 and 2020, Trump got help from a strong data and turnout operation that pushed his frenetic campaigns to outperform expectations set by public polls. In 2016, 10% of the electorate were first-time voters; in 2020 that climbed to 13%. In a close race, those at the margins of the electorate can tip the scales.
(All that assumes a standard, one-on-one presidential campaign. But if Trump loses the Republican primary, support among non-likely voters could animate a Trump third-party campaign. While sore-loser laws make it nearly impossible for Trump to win outright this way, that wouldn’t preclude ratings-grabbing media attention.)
Analyzing an election requires understanding what pushes people to vote. The corollary is that it also requires some understanding of what keeps people away, and what might change someone from a non-voter to a voter (or vice versa). Even when media outlets report poll numbers day to day, they are still just giving us snapshots and assumptions — and in the case of the 2024 race, it really is much too early for those snapshots to be relevant on their own. With their poll on the edges of the voting public, Suffolk and USA Today are doing something different, allowing us to consider the question, “What if…?”
