Split Ticket

President Joe Biden trails former President Donald Trump by one point in a survey conducted by Split Ticket and Data For Progress. The poll was conducted online between July 1–3, 2024 with a sample size of N=2,067 likely voters, and was weighted to be representative of age, gender, education, race, geography, and the recalled 2020 vote.

Our survey found Trump at 41% of the vote, with Biden at 40%, Independent Robert F. Kennedy Jr. at 10%, and minor party candidates Cornel West, Jill Stein, and Chase Oliver each at 1%.

We also polled a two-way, head-to-head race as well. Here, Biden leads Trump 47–46. Critically, we find an electorate that is quite stable and polarized, just like it was in the last election; Biden wins his own 2020 voters by a margin of 91–5, while Trump wins his 2020 voters 93–4. It is the 2020 nonvoters, however, that are far more fluid; our survey found that they tilt towards Trump by a margin of 38-33, with nearly 30 percent of them undecided.

This is a notable change from the way that less-engaged voters tended to lean in previous cycles. For instance, in 2020, they were estimated to be significantly more Democratic than the rest of the electorate. The Republican advantage among these voters is a relatively new development, and it has been one of the cornerstones of Donald Trump’s polling strength in this cycle.

This slender two-way lead is still significantly worse for President Biden than his 4 point win in 2020. Although Trump’s numbers mirror his 2020 vote share, Biden’s have dropped by 3 points. If the core tipping-point states remain to the right of the nation, as they were in 2016 and 2020, then such an outcome would probably make Trump a slight favorite to win the electoral college.

But this result is also actually better for the President than the apocalyptic picture that some other polls have painted in the wake of his disastrous debate, which has sparked nonstop talk of replacing him among party officials — for example, a recent New York Times/Siena College poll found Trump up by six among likely voters.

In the event that Joe Biden exits the race, the likely Democratic nominee would be Vice President Kamala Harris. Our poll finds her running even with Donald Trump in a two-way, head-to-head matchup, with both candidates tied at 46 percent.

Notably, Harris and Biden perform very similarly among every single demographic group, with minimal divergence anywhere, and while Harris’ vote share is one point below Biden’s, her unfavorables are also 3 points better. While there are some minor differences in favorability among subgroups (Harris’ net favorability is higher than Biden’s among voters under 45 and Black voters, for example), this actually does not translate to extra votes in the general election, where Harris performs similarly to Biden among these same groups. This suggests that at the moment, she would begin at a similar baseline if she were to replace Biden atop the Democratic ticket as the nominee, with a nearly-identical coalition.

Generic Ballot

While the presidential matchups were all virtual ties, the generic congressional ballot told a slightly different story. Here, Democrats have a small, but clear advantage over Republicans, as they lead 48–45. Our results seem to align with similar findings in other polls, suggesting that currently, congressional Democrats might not be paying as much of an electoral penalty on persuasion as President Biden is.

Our survey also found that while both parties are viewed unfavorably by the electorate, the Democratic Party still has higher favorables than the Republican Party overall. 52% of respondents had an unfavorable opinion of the Democrats, while 46% viewed them favorably. Opinions of the Republican Party, however, stand at 57% unfavorable and 41% favorable.

But the presidential nominees are a different story: Biden is seen unfavorably by a 59–40% margin, while Trump comes in at a somewhat smaller 57–43%. This favorability gap may help explain why Biden trails Democrats on the generic ballot.

An Experiment With Issue Polling

In this poll, we tested the impact of some of the issues on Americans’ minds going into November. Predictably, the economy ranks as the most salient topic, with 30% of respondents rating it as their most important issue when deciding which candidate to vote for. This held true among Republicans, Democrats, and Independents.

But these types of questions can mislead — indeed, if you looked at 2022 polls, they would have suggested that abortion and democracy would have minimal impacts on voters’ minds, despite actual election results telling a completely different story. Although one general issue can be reported as the most important, voters often have a more holistic view of the issues at stake.

To get a better sense of how such polarizing issues may influence voters, we decided to try an experiment similar to the one run by the New York Times a year ago. We polled hypothetical congressional matchups between generic Republican and Democratic candidates, varying the positions that each party’s candidate took on a key issue for each question.

To get a baseline, we first polled a generic Democratic candidate against a generic Republican candidate for Congress. As mentioned earlier, we found the Democrat up 48–45, leading the Republican by a margin of 3 points. After this, we split up the sample into two separately weighted groups, giving each group a different hypothetical matchup with candidates that took different positions on issues. Comparing the two revealed fascinating insights into how voters responded to specific issues and stances taken by candidates.

First, we tested abortion — an issue at the top of many voters’ minds since the Supreme Court overturned Roe v. Wade in 2022. While only 8% of the overall electorate considered abortion their most important issue when it comes to choosing a candidate, our experiment suggested (perhaps predictably) that most respondents are not on board with more extreme Republican rhetoric on abortion access.

Our first matchup featured a Democrat advocating for the codification of Roe and a Republican arguing that abortion policy should be left up to the states. The generic Democrat led 48–45, the exact same as the D+3 generic ballot topline. Unsurprisingly, there was also a clear gender divide, with women preferring the Democrat 51-41 and men opting for the Republican 49–44, practically identical to the splits seen in the unspecified head-to-head.

The second matchup pitted the same pro-choice Democrat against a Republican advocating for a federal ban on abortions after fifteen weeks. Here, the divide was much larger, as the generic Democrat led 52–41. Interestingly, the Democratic boost was concentrated among women, who backed the Democrat by a margin of 56–35 in this scenario (a swing of 11%); by comparison, Republicans won men 48–47 (a swing of just 4%). Collectively, the results indicate considerable, albeit expected, gender polarization over abortion. 

Overall, the fact that abortion didn’t clock in near other hot-button issues (like the economy or immigration) on the standard “issue importance” question should not deceive. Our head-to-heads demonstrate clearly that abortion remains important to the electorate and that, as we’ve written before, Republicans would play a losing game if they chose to focus on extreme abortion policy.

If you’re still not convinced, keep in mind that we also polled preferences on the legality of abortion, and 61% of respondents said that abortion should be legal in most or all cases.

Next, we tested views on threats to democracy, which ended up being the third most important issue among respondents. We used January 6th as a proxy for these kinds of threats. First, we polled a matchup between a Democrat characterizing Trump as a threat to democracy and a Republican pushing for the public to “move on from January 6th and the 2020 election”. The Democrat led 47–46, which is among the narrowest measured congressional ballot toplines in our entire survey. 

But when we pitted the same Democrat against a Republican who claimed that the 2020 election was stolen from Donald Trump, the Democrat led 52–42, with Independents driving most of the change. This aligns with most of our previous work suggesting that election denial is spectacularly unpopular, and that like abortion, it is a hot-button issue Republicans should try to avoid (and one that Democrats might seek to elevate).

Finally, we touched on the foreign policy issue recently at the top of many voters’ minds: Israel-Palestine. Our first matchup tested a Democrat and a Republican who “both agree with Israel’s current approach to the conflict in Gaza”. In this case, the generic candidates tied 44–44. The second matchup saw the same Republican facing a Democrat supporting “an immediate ceasefire and a halt of military aid and arms sales to Israel”. Interestingly, the Democrat led 49–43, with Independents and 2020 non-voters driving the bulk of this shift.

We want to stress that these campaigns do not treat issues in isolation from each other, and that news cycles can often be beyond a candidate’s control. But these questions do give us some insight into how voters react to topics when they gain salience, and how they may move in response to certain political stances taken by candidates on hot-button issues.

A Methodological Note

If you’ve read this far, perhaps you’re willing to go a little bit further. We’re going to get very detailed here to discuss a key point in this survey: the crosstabs in our survey, and how to reconcile some of them with the actual composition of the poll subsamples.

In order to better guard against some of the issues we see in survey crosstabs, we chose to “nest” our weights, similar to an experiment we did in our last poll. In the past, we found that it made our subgroup estimates a bit more stable, especially with certain demographics. You can find the weights here, and the poll crosstabs here. But like anything, these results still require careful contextualization. 

For example, our Black electorate supports Biden by a two-way margin of 77–23 in today’s survey, which would represent a titanic swing right from 2020’s 91-9 split in his favor. But in our sample, this group only backed Biden 81–19 in 2020 (excluding non- and third-party voters), according to their self-reported recalled vote. This means that the swing we measure towards Trump is actually 8 points in margin.

That is a sizable shift, of course, but it is not nearly as big as the 28-point swing that you would get when comparing post-election estimates to pre-election polling. What this result really suggests is that Democrats are on track to slip some more with Black voters in this election, but that the expected swing would be a slight acceleration of the slow shift towards the Republican Party since Barack Obama’s high-water marks with this group.

On the flip side, Biden’s relatively strong numbers with white voters are at least partly due to a more Democratic white sample. For instance, post-election estimates had Biden losing white voters by 12, whereas in this survey, he loses them by just 10 in our survey. On the face of it, that would imply that white voters are swinging to the left, in the face of national trends. However, this sample’s recalled vote was only Trump +10 in 2020, meaning that while Biden is certainly holding up better with white voters than he is with non-white voters, his standing is not as good as a simple, naive comparison between 2020 post-election estimates and today’s poll crosstabs suggests.

Another interesting group remains voters under 45. In our sample, this subgroup’s two-way recalled vote was Biden +16 in 2020. However, for the 2024 election, their two-way vote preference is just Biden +7. This is because although Biden and Trump both get 94% of their own 2020 voters in the two-way vote, Trump actually wins the under-45 2020 non-voters by a few points, and this is driving a good chunk of his gains with this demographic in our survey. (Notably, all of these groups have an N of greater than 100 in our survey, which is why we feel comfortable reporting on them, despite much higher margins of error).

And while our Latino sample looks fairly healthy for Biden, with a lead of 20 among voters of Hispanic or Latino descent, it’s actually a rather large swing away from him — this sample’s 2020 two-way recalled vote was 70-30, showing that Democrats may actually risk serious losses with this group. In fact, Biden only wins his own 2020 Hispanic voters by a two-way margin of 89–11, which is a fair bit lower than the numbers he gets with the rest of his 2020 voters.

We stress that this type of crosstab “noise” in sample composition is actually entirely normal for most polls. The survey is weighted to be representative of the overall electorate in terms of demographics and partisan lean; however, certain groups may be more Republican in the poll than in reality, which gets countered by certain groups in the sample being more Democratic. This ends up having minimal effect on the overall poll topline; it just means that we need to be a bit careful when extrapolating from crosstabs and stick to broader, directional inferences, as always.