Campaign Roadmap
Campaign Roadmap
Polls can't predict our destination. But they can tell us about our direction.
We’ve reached the point in the campaign where we are being flooded with new polling data every day. And where there is polling data there is polling analysis, sometimes featured prominently in breathless headline news stories.
Past experience leads me to expect that some of it will not be particularly good. Perhaps even most of it.
Journalists tasked with interpreting poll results for readers tend to base their analyses on head-to-head horserace numbers and are susceptible to falling into several traps:
Relying exclusively on their organization’s polls. News organizations that commission polls invest a lot in them for the purpose of using the results as a peg for headline news stories, so reporters elevate them to breaking news status. But no single poll should be considered newsworthy in its own right.
Ignoring apparent outliers. When a poll result is at odds with the preponderance of other results, it should be regarded as an outlier unless and until other polling corroborates it. But that doesn’t make for a compelling story, so it’s rarely done.
Overlooking assumptions. Every poll makes assumptions about who’s going to vote, and in an age of hyper-partisanship—when turnout can determine the outcome—the accuracy of these assumptions can be critical. Uncertainty surrounding polling assumptions should inform any analysis, but doing so means watering down the headline.
Treating horserace numbers as predictive. Expressly or implicitly, journalists suggest that we can discern who’s going to win the election from head-to-head results. The accuracy of polling should improve as we get closer to Election Day, but it is always a better gauge of the present than the future.
This last point is important. Polling is better regarded as a roadmap than a crystal ball. It can tell us which way we’re pointing better than it can tell us about our destination.
We should look beyond the horserace figures in any given poll and consider other characteristics of public opinion that can lend depth to our understanding of the race.
Good public opinion data (and that covers only a subset of what’s reported in the press) can tell us if the election is moving in a particular direction. It can tell us about the intensity of support for the candidates. It can tell us whether we’re in a stable or unstable political environment, which in turn can help us determine the likelihood that things will change in these final weeks.
The prevailing analysis of the 2024 election is that it is tied. If you look only at top line polling data showing the election within the margin of error, it would be accurate to classify the election at this point as too close to call.
But that doesn’t necessarily make it a toss-up.
Dig beneath the head-to-head numbers and you’ll find a close contest that’s playing out amidst conditions more favorable to Kamala Harris than Donald Trump. Despite what Harris may say at her rallies, it is actually Donald Trump who is the underdog.
Direction
Polling averages paint a clear picture: Kamala Harris has been gaining support—gradually but steadily—since she entered the race in July. Although head-to-head numbers remain within the margin of error, she has improved her standing relative to Donald Trump during the brief time she has been a candidate.
On July 26, five days after Joe Biden dropped out of the presidential race, Harris held a slim 0.2 percentage point lead over Trump in the fivethirtyeight.com polling average. Within five days that lead had grown to a full point, then to two points by August 8. By the start of the Democratic National Convention it had expanded to three, where it settled for the remainder of the summer.
In an absolute sense, a lead of 0.3 points and a lead of 3.0 points are effectively the same thing—a measure of a race that’s statistically tied. But from the standpoint of how public opinion is moving, it suggests an improvement in Harris’ position.
While we can’t look at these numbers and say with accuracy that Harris has a three point lead, we can say that her support has been building and she has been holding on to the voters who have moved to her campaign.
This growth is reflected in the share of the vote she is receiving in polling averages. On July 26, she was averaging 44.2% of the vote. By July 31, it was 44.7%. On August 8 it was 45.4%, and it crossed 46% on August 12, 47% on August 21, 48% on September 5, and hit 48.4% last Sunday.
Donald Trump’s growth has been more flat. He has been stuck in the 44-46% range since Harris entered the race, clocking in at 43.9% on July 26 and 45.6% on Sunday.
The gradual improvement we’re seeing for Harris is a reflection of her consolidating, exciting, and expanding her base, and while we cannot predict that it will continue, there remains room for growth in the final weeks of the campaign.
During this time, there has been a corresponding surge in Harris’ favorability scores. An NBC News poll released last weekend found a jump of 16 points since July in the percentage of people who view her favorably, the largest increase they have recorded since George W. Bush in the days following the September 11 attacks.
When she entered the race, Harris was viewed positively by only 32% of registered voters and negatively by 50% (for a -18 net favorability score). By mid-September, the share of voters viewing her positively had soared to 48% versus 45% who view her negatively (a favorability score of +3).
For any political figure to have net positive favorability is remarkable in our polarized age, and the movement toward Harris is a noteworthy indicator of the upward direction of her campaign. While it is possible to get people to vote for candidates they view unfavorably, it is much easier to get them to vote for candidates they like.
Intensity
Public opinion also has an element of intensity—how strongly people feel about something, like how excited they are to vote for a candidate.
Although an enthusiastically cast vote counts as much as a reluctantly cast vote, measures of intensity can be used as a proxy for turnout, with candidates who benefit from enthusiastic supporters likely to find it easier to mobilize them to vote.
When Joe Biden was the presumptive nominee, Republicans held an intensity edge over Democrats. Then he dropped out and intensity indictors flipped upside down.
Back in March, Gallup found that only 55% of Democrats were enthusiastic about voting, trailing Republicans by four points. But enthusiasm among Democrats shot up once Harris became the presumptive nominee, and by the time the Democratic convention rolled around, a near-record 77% of Democrats were excited to vote—fourteen points higher than Republicans.
These figures are rivaled in recent times only by the excitement surrounding the 2008 Obama campaign.
Looking beyond polling data, we see intensity reflected in voter registration surges relative to the same point in 2020 among young people, women, and voters of color. These spikes are attributable to Harris entering the race (with an additional massive surge following Taylor Swift’s endorsement).
We see intensity in the volume of small dollar contributions to the Harris campaign, along with an influx of new donors who hadn’t given to Joe Biden. In July, two-thirds of the $310 million Harris raised came from new donors, and 94% of her contributions were $200 or less. In August, 1.3 million first-time donors contributed to her $361 million haul.
And while it is a subjective standard, we see intensity reflected in the size and enthusiasm of Harris’ crowds. This, of course, is Donald Trump’s favorite measure of self-worth, and although we have seen plenty of losing candidates hold massive rallies, taken together with the fundraising, registration, and enthusiasm figures, Harris’ rock-concert events offer a readily accessible example of what intense supporters look like.
Stability
There was a time when convention bounces really bounced. Michael Dukakis went from dead even to an 18 point lead after the 1988 Democratic convention, only to see his opponent George H. W. Bush take a four point lead with his convention bounce and go on to win the election by eight.
In 1992, Bill Clinton experienced a 23-point convention bounce. In 1976, Jimmy Carter’s bounce boosted him from a 13 to a 36-point polling lead before it came crashing down. Carter ended up winning the election by only two points.
Swings of this magnitude were commonplace prior to our hyper-polarized era, when a large number of voters were open to being swayed by candidates from both parties.
The electorate today is far less volatile. For all the changes we’ve experienced since Joe Biden left the arena, the head-to-head contest between Biden and Trump is not all that different from the head-to-head contest between Harris and Trump, with the candidates stuck in the mid-to-high 40% range. The four percentage point improvement in Harris’ numbers since she entered the race counts as significant movement in today’s political climate.
That means as we enter the last weeks of the contest, things are more likely than not to remain close to where they are today. If Biden’s dramatic departure wasn’t enough to shake things up by more than a few points, it’s difficult to imagine what will.
When we look at these elements of public opinion together, we can see why Harris goes into the last weeks of the campaign with an advantage. The election has been moving her way since she became a candidate, her supporters are more fired up, and the stable political environment should help her endure whatever surprises await around the bend.
Polling may not be able to tell us who is going to win or by how much, but it can tell us which candidate is benefitting from the direction, intensity, and stability of the race.
Polls may show Harris and Trump running neck and neck, but in the larger public opinion environment things are not as close.
Intensity is very hard to measure beyond asking ;people "are you very enthusiastic" and "do you intend"?