With President Joe Biden and Donald Trump having locked up enough delegates in the 2024 primaries to secure their candidacies in November’s Presidential election, it seems like an appropriate time to look at the latest polling data and electoral-map projections. With nearly eight months to go, the polling shouldn’t be taken as gospel, obviously, but it’s sending a fairly consistent message, and it illustrates the challenge facing Biden.
In the 2020 election campaign, Biden led in the poll averages throughout and ended up winning the popular vote by 4.5 percentage points. But, as of Saturday morning, Biden is trailing Trump, albeit only by 1.7 points, in the RealClearPolitics poll average. In eight national polls carried out since Biden’s energized State of the Union speech, and recorded in the 538 database, two have shown him leading Trump, and six have shown him behind. (In almost all the surveys, the margins were small.) Even more concerning for Democrats is a series of new polls from key battleground states that showed Trump retaining a lead in some places that Biden won in 2020. According to the RealClearPolitics poll averages, on Saturday morning, Trump is ahead of Biden by more than five points in Arizona, Georgia, and Nevada, by 3.5 points in Michigan, and by one point in Wisconsin.
It’s worth noting straight away some of the large asterisks that should be attached to the current polling data. The first is that there’s a long way to go. At this point in 1980, Ronald Reagan was tailing Jimmy Carter by double digits in the Gallup poll, and, at this point in 2016, Donald Trump was trailing Hillary Clinton by eight points in the RealClearPolitics poll average. Needless to say, Reagan and Trump both ended up winning. In 2020, in contrast, the early polls and the final result turned out to be pretty similar.
To put it another way, early polls aren’t necessarily a reliable predictor of the final results. This week, in a useful survey of the historical data going back to the postwar era, G. Elliott Morris, the polling director of ABC News, including 538, pointed out that, on average, the polling figures from March of an election year have diverged from the final result by about eight points. Morris wrote that there is “plenty of precedent for a candidate like Biden to gain enough ground to win the election.”
Morris also discussed the reliability of the polling data, a matter that Biden and some of his aides raised with my colleague Evan Osnos, in a recent Profile of the President that was published in this magazine. One big issue is that, right now, response rates to pollsters’ queries are extremely low—one or two per cent, in many cases—and the people who do respond may not be representative voters. (It’s widely assumed that they are more committed than the average voter.) Because of this problem, pollsters often have to reweight their data to get what they think is a representative sample, but this calculation introduces a potential for a “non-sampling error” in addition to the regular sampling error attached to random samples. Putting both of these things together, and other potential issues, Morris says the zone of uncertainty for a poll with a sample of about a thousand people is now six or seven points.
Does this mean we should ignore the polls? No. But, rather than obsessing about individual surveys, we should pay more attention to the poll averages, which have less uncertainty attached to them. They still provide the best snapshot available of public opinion. Also, despite their flaws, individual polls can still provide some valuable information. For example, even if a certain poll is (inadvertently) off one way or another, then, as long as its error remains roughly constant, successive iterations of the survey should indicate how public opinion is shifting over time, or if it’s shifting at all.
One striking thing so far this year is that there has been little movement. On January 1st, the RealClearPolitics poll average showed Trump leading Biden by 2.3 percentage points, compared to the current 1.7 points. This stability expands to many other poll findings. For example, in the first five polls recorded in the RealClearPolitics database this year, Biden’s average approval rating on handling the economy was just 38.8 per cent. In the latest five polls, his average rating was thirty-nine per cent, basically unchanged. That is despite the fact that some other measures of economic sentiment, such as consumer confidence, have picked up since late last year. (The two most recent reports from the Conference Board and the University of Michigan showed the consumer confidence indexes dropping back a bit, but in both surveys the indexes were well above last October’s levels.)
Conceivably, it may take more time for changing perceptions of the economy to feed through to the President’s ratings. And, as the media narrative shifts squarely to Biden versus Trump, and the real possibility of Trump returning to the White House, the over-all dynamics of the race could start to change. In a podcast with The New Republic’s Greg Sargent, earlier this week, the Democratic strategist Simon Rosenberg compared the situation to a basketball game in which one team is losing by two points after the first quarter. Rosenberg cited a number of factors working in Biden’s favor. In a series of electoral contests since 2020, most recently a special congressional election on Long Island, Democrats have performed better than Republicans, Rosenberg noted. He also mentioned the Democrats’ lead in fund-raising, the formidable grassroots operations that the Party has built up in key states, and Trump’s tendency to say potentially damaging things, such as his suggestion earlier this week (subsequently retracted by his campaign) that he was open to cuts to Social Security and other entitlement programs.
Trump’s gaffe, if that’s what it was, served as a reminder of how unique this election is. For the first time in more than a century, it features a sitting President and a former President. Never before have there been two major candidates who were so old, or, since the dawn of polling, two candidates who were so unpopular. (According to the 538 poll average, Biden and Trump both have favorability ratings below forty-three per cent.) The unprecedented nature of the election is another reason for exercising caution in interpreting it, but, in all likelihood, as in 2016 and 2020, the outcome will rest on the results in a handful of states.
The immediate task facing Biden’s campaign is to secure his current polling leads in some of the states that he won handily in 2020—Minnesota, New Hampshire, New Mexico, and Virginia—and bring together the Democratic coalition in places such as Michigan and Nevada, which the Party has carried in its past three Presidential victories, going back to 2008. If the Biden campaign can accomplish these tasks—which have been complicated by the President’s position on Gaza, and by the presence of third-party candidates—it can look to eke out additional victories in Wisconsin and Pennsylvania which could carry it across the two-hundred-and-seventy-vote threshold for victory in the Electoral College.
Nearly eight months out, such a scenario is largely speculative: lots of things could change the electoral map. But, considering the precedent of recent elections, this year’s contest was always likely to be a close one. Given the threat of another Trump Presidency, and all the chaos it would bring with it, plus the extremism of the Republican Party, particularly on abortion, the Biden campaign is exuding confidence that the dynamics of the race will eventually change, and that Biden will prevail. But, between now and Memorial Day, strategists from both political parties—as well as many concerned voters—will be watching the polls closely to see if, after the end of the primary season and the start of the campaign proper, there is any solid evidence of this shift happening. ♦