Commentary by Professor Thomas Scotto, Head of the Department of Government & Public Policy

Polling firms, betting markets, and campaign data suggest that the 2024 US presidential election is neck-and-neck. As a political scientist who has taught American politics abroad for nearly two decades, I see two key questions emerging: Why is America heading toward its third consecutive election that could be decided by a margin smaller than a Premier League crowd; and how much confidence can we place in polling data, and what factors could lead to an outcome that diverges from predictions?

The Competitiveness of American Presidential Elections

Political scientists generally agree that competitive elections are a net positive for citizen engagement and the quality of a nation’s democracy. When elections feel like a foregone conclusion, candidates often spend fewer resources on campaigning and sharing their policy positions. In competitive elections, however, candidates must work hard to prove to voters that their leadership and ideas make them the best choice.

The unique structure of the Electoral College adds complexity, making American presidential elections a mix of competitive and non-competitive elements. For instance, in 2024, the Cook Political Report identifies 36 out of 50 states as “Solidly Democratic or Republican.”

History and expert opinion suggest majorities of voters in states like Connecticut are almost certain to support Kamala Harris, while Donald Trump is considered unbeatable in places like Oklahoma. According to 2023 population estimates, roughly 60% of Americans live in states where the outcome is virtually assured.

Meanwhile, traditionally conservative states like Florida and Texas lean Republican but are not entirely out of reach for Democrats.

This leaves just seven “toss-up” states in 2024: Pennsylvania, Michigan, Wisconsin, North Carolina, Georgia, Arizona, and Nevada. Fewer than one in five American voters live in these critical states, making their and not the majority of Americans’ choices pivotal in determining the next president.

The Impact of Micro-Targeting in Swing States

The competitive states are both geographically and demographically diverse, featuring unique groups such as growing Latino populations in the South and a significant Muslim population in Michigan. This diversity forces candidates—especially Vice President Harris—to address issues that resonate with these communities, such as Middle Eastern conflicts.

At the same time, the small number of “in play” states allows campaigns to “micro-target” voters who could tip the outcome. Campaigns send highly tailored messages to motivate specific groups within these states to vote.

For example, the Harris campaign in Michigan may send different literature to undecided voters in Detroit, a city with a large immigrant community, compared to single women under 30 across the state. Critics argue that micro-targeting undermines broad issue deliberation in campaigns and stirs intense emotions on divisive issues, contributing to political polarization.

Why Competitive, Swing-State-Focused Elections Are Likely to Persist

The factors driving competitive, swing-state-focused presidential elections are unlikely to change soon. While most Americans—and a growing number over time—support eliminating the Electoral College, the constitutional barriers to doing so are formidable. Although demographic shifts are significant, they won’t reshape politically “safe” states quickly, and demographic change doesn’t always translate to political realignment.

For instance, while many expected Texas to become a “swing state” due to its growing Latino population, a sizable portion of Latinos have increasingly aligned with the Republican Party, which will likely keep the state in Trump’s column.

In short, while only a few states remain highly competitive, American presidential elections are likely to stay close, with considerable implications for candidates and voters alike.

What About the Polls?

In both the United Kingdom and the United States, the polling industry has faced strong criticism over the past decade. Discrepancies between poll predictions and final election outcomes, coupled with high public expectations, have fuelled this critique.

The public rightly expects pollsters to sample a representative group and ask questions that reveal respondents’ true intentions—in this case, their preferred presidential candidate. But as any poll only samples a portion of the population, it has a margin of error, usually around plus or minus three percent.

Polls also typically use a 95% confidence interval. So, if a poll reports a tie between Harris and Trump at 48% each, the actual support could vary between 45% and 51%. Therefore, if the final polls show a tie but one candidate wins 50%-46%, this is within statistical expectations and should not be seen as a failure of polling.

However, polling faces challenges beyond statistical margins of error. The theory behind sampling assumes that pollsters understand the population they are sampling, but this has become harder to achieve. Not all eligible, registered citizens vote, so pollsters often include only “likely voters” in their sampling, and identifying these voters can be tricky! Traditionally, a “likely voter” is someone who has voted in previous elections.

But in places like Wisconsin and Pennsylvania, Donald Trump’s ability to mobilize “forgotten men”—voters who had previously sat out elections—can introduce bias in polls if these voters are excluded.  However, in 2024, some are wondering if pollsters may be overcorrecting, over-surveying those who voted for Trump previously but may sit this election out.

Not only does this short discussion of polling suggest that predicting elections lies somewhere between statistical science and the creative arts, but it underscores that this election, at the end of the day, will be won by the candidate who gets people to the polls.  Both former President Donald Trump and Vice President Kamala Harris are well-known figures, running costly campaigns in a polarized America.

While the outcome remains uncertain, it’s likely that post-election discussions will focus on which candidate better motivated their base rather than who convinced the last few undecided voters.