Election 2024: What and Why

Election analyses this soon after the election are nearly all based on exit polls. Exit polls are subject to “sampling bias,” meaning that members of various social or economic groups aren’t necessarily represented in the polls in the same proportion as they are in the actual population. For instance, college graduates and younger voters tend to be over-represented. Also, the overall winner of the exit poll must match the winner of the actual election, and if it doesn’t the pollster will make statistical adjustments to the results, which can further exacerbate the over-representation.

In 2020, for instance, exit polls underestimated the share of voters who were white without a college degree, and overestimated the share of white women who voted for Trump. As Wired’s Gilad Edelman wrote, though “None of this is to say that any of the emerging narratives about how various groups voted this year are wrong. We just don’t know yet.” More accurate analyses must wait until states finish updating their voter files – some time next year – and public information about who did and did not vote will be available. At that point studies validated against voter files will be released, such as Harvard’s Cooperative Election Study and the Pew Research Center’s validated voter survey.

As of this writing, the Cook Political Report’s National Popular Vote Tracker, which uses official sources from the states, shows Trump with 49.83% of the vote, vs. Harris’s 48.28%. A Trump victory, yes, but not the grand mandate his minions have trumpeted, and in fact – as we noted on our home page – in the history of presidential elections only five popular vote winners have prevailed by less.

So with the caveats above in mind, let’s take a look at the exit polls.

What

The Associated Press found that Trump won a bigger percentage of the vote in every state and Washington, D.C., and more actual votes in 40 states than he did in 2020. Turnout was somewhat lower than in the 2020 election (153 million vs 158 million in 2020), but approaching that historic high. In three key northern urban counties –  Wayne County, MI (Detroit metro), Philadelphia County, PA, and Milwaukee County, WI – Harris won fewer votes than Biden collected in 2020, and Trump won more than he did in 2020. Statewide in Michigan and Pennsylvania, and also in Arizona, Harris received fewer votes than Biden did in 2020.

In Arizona Harris’s vote total was more than 60,000 votes lower than Biden in 2020. Trump improved his margin of victory in every county, compared to 2020, including almost 56,000 more votes in Maricopa County where Phoenix lies. These effects combined enabled Trump to carry the state this year, which he narrowly lost in 2020.

In Georgia, Nevada, North Carolina, and Wisconsin, Harris won more votes than Biden did in 2020, but Trump increased his total by more. In Georgia, for example, Harris received more than 73,000 more votes than Biden did in 2020, but Trump’s total grew by over 200,000. The AP found that even some of the most Democratic-leaning counties shifted votes toward Trump compared to 2020. Vote totals in Michigan’s Wayne County, for example, shifted 9 points toward Trump.

The AP also found that, although Harris “won majorities of Black and Latino voters, it wasn’t enough,” to offset Trump’s gains. Approximately 30% of Black men under 45 voted for Trump in 2024, compared to roughly half that in 2020. And approximately 50% of young Latino men voted for Harris, down about 10% from the share that voted for Biden in 2020. 

From among the five issues presented in the Edison Research exit poll – the state of democracy, the economy, abortion, immigration, and foreign policy – approximately one-third of voters selected either the state of democracy or the economy as their most important issue. Harris collected 80% of those choosing the state of democracy, and Trump 80% of those choosing the economy. Of the 14% of voters who chose abortion as their most important issue, Harris won 74%. But her margin there was offset by the 11% of voters who selected immigration as their key issue, 90% of whom voted for Trump. Also of the 4% of voters identifying foreign policy as their most important issue, Trump won 57% (compared to Harris’s 37%).

Pre-election polling by Marquette University Law School and the University of Maryland Program for Public Consultation found that attitudes toward immigration – and particularly the mass deportations promised by the Trump campaign – varied based on language and details used to describe the potentially affected population or the selection and deportation process. Some analysts have suggested that these results mean that the Harris campaign might have been able to limit the popularity of Trump’s deportation plans if they had focused more on publicizing its negative aspects. Instead Harris et al. highlighted Republican obeisance to Trump and Vance in opposing the bipartisan immigration and border legislation in congress. Trump’s rhetoric was typically ambiguous, initially telling supporters he was fine with being blamed for the bill’s failure to pass, but later opposing it more directly while falsely claiming it would allow 5000 illegal immigrants to enter the country every day.

Exit poll results on abortion as an issue present a confusing picture. Most Americans want abortion to be legal under all or most conditions, and this statistic is apparently somewhat independent of party identification. The AP Vote Cast survey conducted after the election found that in Arizona, Missouri, and Nevada – states that passed abortion rights amendments to their constitutions – 30% of those supporting these measures also voted for Trump. Analysts have speculated that voters who supported abortion rights but voted for Trump may simply have prioritized other issues such as the economy or immigration. But a September New York Times/Siena College poll suggested another factor: only 27% of Republicans (and 49% of voters overall) thought Trump would support a nationwide ban on abortion.

Finally, voters cited dissatisfaction with the economy as a major factor in their decision to support Trump vs. Harris – this despite strong growth, easing price increases, and a strong job market in the lead-up to the election. We will discuss this and other perception-vs-reality issues below, but one set of factors influencing this gap between perception and reality is that the way Americans view the economy has changed since the COVID19 pandemic. Among other factors: personal savings grew during the pandemic as social activity declined, but in 2024 the personal savings rate has fallen to one of the lowest levels since the Great Recession; the inflation rate grew during and after the pandemic, and, while in September 2024 it dropped to its lowest point in 3 years, as ABC News put it “Even if inflation is getting better, Americans aren’t done being ticked off that it was bad to begin with.”

Why

The American Prospect’s David Dayen offers arguably the most concise TL;DR analysis of the 2024 US presidential election.

The temperature of the world has been taken. And with a few notable exceptions, and to the extent that those elections were free and fair, the result has been largely the same: Virtually every party that was the incumbent at the time that inflation started to heat up around the world has lost.

… Populations experienced the shock of the pandemic and the further shock of high prices, grew angry, and took it out on the ruling party.

This, Dayen adds, is “irrespective of the causes of that price increase” (italics in original). And, as we’ve noted above, “people don’t give ruling governments credit for taming inflation if they were around for its rise. The price level is permanently higher than people think is fair” (italics in original). Inflationary effects are compounded in the US, Dayen suggests, by the generally high cost-of-living, especially as it affects middle class families: housing, health care, child care, etc.

The Atlantic’s Annie Lowrie expands on this point in her November 11 piece titled “The Cost-of-Living Crisis Explains Everything.” Lowrie refers to deliverism – a term coined by commentator Matt Stoller in 2021 to refer to “governing well and establishing a record that the electorate needed to win actually feels.” Yet with the economy having grown consistently for the past two years, a May Harris-Guardian poll found 55% of Americans believed the economy was shrinking or in a recession. The Washington Post’s Catherine Rampell suggests that one set of reasons for the gap between perceptions of the economy and reality is differences between technical and informal usage of economic terms. “Recession,” for instance, to an economist means that the economy is shrinking as measured by output, employment, etc.; informally people may label the economy as a “recession” if something about it feels bad – high gas prices, for example. In the same way “inflation” to an economist refers to prices increasing, while informally people may use the term to refer to prices that are merely high.

Lowrie notes that, regardless of the real economic gains, Democrats failed to acknowledge the strain on average households, with real median income dropping below its pre-COVID high, and the poverty and jobless rates moving up during the Biden administration’s tenure. In Lowrie’s view, Democrats failed to acknowledge adequately the cost-of-living hardships for average Americans, and the focus on the technical details left some voters feeling like they weren’t being heard. She acknowledges, however, that “[a] better economy might not have delivered the gains that Democrats once could have relied on,” and that today’s voters seem more motivated by cultural or “identitarian” issues. Moreover, as Rampell observes, today’s media consumers have a “bad news bias,” especially when the source is social media, being more likely to engage with news items that evoke outrage. (More on this below.)

Perhaps not surprisingly, Republicans were more likely during the Biden years to believe that the economy was in decline. The May Harris-Guardian poll found a “vast majority” of Republicans believing that the economy was shrinking, inflation increasing, and overall economic conditions worsening; less than 40% of Democrats polled held the same views.

One set of factors affecting this partisan gap in perception has to do with the distribution and consumption of news. A November report from the Civic Health and Institutions Project found that “friends and family” were the leading source of election information among the 25,000 adults surveyed (29% compared to 26% for news media). Republicans and the 18-24 age group were particularly susceptible to this behavior. Significantly voters in the 18-24 age group, while still favoring Harris overall, shifted some support to Trump compared to the 2020 election (young women 65/33% Democrat in 2020 vs 57/40% in 2024; young men 56/41% Democrat in 2020 vs 42/56% in 2024).

Viewed another way, as a Data for Progress study found, Harris led by 6% among voters who consumed “a great deal” of news (52/46%) and by 3% among those who consumed “a lot” of news (50/47%); Trump led by 7% among voters who consumed “a little” news (49/42%), and by 19% (!) among those who reported their news consumption as “none at all” (51/32%).

As a 2021 study published in the journal Science found, political conservatives perform worse than liberals at distinguishing truths and falsehoods. “This is partially explained by the fact that the most widely shared falsehoods tend to promote conservative positions…,” the researchers found. Their study also established that “conservatives have lower sensitivity than liberals,” meaning that on average they were less able to distinguish truths from falsehoods – and that liberals experienced “bigger improvements in sensitivity than conservatives as the proportion of partisan news increases.”

In the last weeks of the campaign The Atlantic’s Adam Serwer travelled through the South interviewing Trump voters. He found what he calls “a tendency to deny easily verifiable facts about Trump.” For instance, a voter complained to Serwer that the media was comparing Trump to Hitler – this despite former Trump chief of staff retired Marine general John Kelly recounting Trump having said that “Hitler did some good things.” Another voter complained to Serwer about Trump being labeled racist. “Just because the media says he’s racist doesn’t mean he’s racist,” they said. Serwer notes the irony of this voter’s assessment in the context of Trump’s many racist remarks during the campaign, as well as his history promoting the false claim that Obama was not born in the US dating back to 2011.

A more recent report in Science from researchers at Princeton and elsewhere found:

Compared with trustworthy news sources, posts from misinformation sources evoked more angry reactions and outrage than happy or sad sentiments. Users were motivated to reshare content that evoked outrage and shared it without reading it first to discern accuracy. Interventions that solely emphasize sharing accurately may fail to curb misinformation because users may share outrageous, inaccurate content to signal their moral positions or loyalty to political groups.

Speaking to the Washington Post, study co-author Molly Crockett suggests that “people care about more than just accuracy,” when sharing information on social media. “... I think outrage is a way to signal important aspects of your political values, your moral commitments, your social identity,” Crockett says. Contrary to the phrase coined by conservative pundit Ben Shapiro, that “facts don’t care about your feelings,” Crockett and her team found that “In this information environment, your feelings don’t care about facts.”

In a related study from 2018 researchers examined the premise that “rational” voters should not base electoral decisions on “events outside the control of politicians, such as natural disasters or economic shocks.” The study showed that

[S]uch events can affect electoral outcomes, even if voters are rational and have instrumental preferences. The reason is that these events change voters' opportunities to learn new information about incumbents.

Former Democratic policy advisor and current political blogger Taegan Goddard compares the Trump team’s use of social media in the 2024 campaign to the Obama campaign in 2012. The Obama team used Facebook, in particular, to mobilize younger voters and drive turnout “among a demographic historically difficult to engage.” The Trump 2024 campaign borrowed Obama’s playbook, Goddard suggests, substituting podcasts for Facebook. “By planting Trump in podcast spaces where political talk was often a sidebar to broader cultural conversations, his team activated new voters who might otherwise have stayed home,” Goddard writes.

Trump appeared on The Joe Rogan Experience, The Shawn Ryan Show, and The Lex Fridman Podcast, among others. He also appeared on livestreamed events with influencers “ensuring he dominated the digital airwaves where younger Americans spend their time.” By contrast, Harris confined her appearances to “traditional outlets and a conventional slate of interviews with mainstream figures.” The result, according to podcast tracking service ivy.fm is that Trump was featured or mentioned in 70,000 podcast episodes, compared with Harris’s 13,000.

In a November 14 op-ed piece for The Hill, BlueWing Impact’s Andy Oare writes:

Democrats must invest in progressive media while simultaneously being willing to venture outside their comfort zones. The mainstream media, while important, can no longer be relied upon as the primary conduit to voters. With trust in traditional institutions at historic lows, authentic engagement through diverse media channels isn’t just an option — it’s a necessity.

The bigger question for the US and our institutions, though, is not why Trump won, but how he came to be able to run at all.