Ahead of the 2024 Presidential Election, many Americans are left wondering if they should pay attention to polling
Ask anyone for their 2024 Presidential election prediction and, regardless of who they think or hope might win, you’re likely to hear one thing: it will be close. And if we take a look at the polls, that claim is backed up. ABC News’s website FiveThirtyEight, which collects and aggregates polling data from across the country, currently has Harris leading narrowly with 48.1% compared to Trump’s 46.7%. But how reliable are the polls and predictive models we see so often on the news, social media and headlines? The answer is complicated.
For one thing, there are more political polls now than there ever have been. According to data from the Pew Research Center, the number of national public pollsters more than doubled from the year 2000 to 2022. That increase likely has many causes, but one is the increase in 24/7 cable news coverage.
“Before the emergence of the 24-hour news cycle polls weren’t shown as frequently because news segments had less air time,” said Morgan Govekar ’26, a student in PSC-211 Election Polling & Public Opinion, which has been studying the history of election polling. “But the emergence of the 24-hour cycle really created a necessity for larger corporations and any news outlet in general to have a constant source of content and polls for people to watch.”
Additionally, sample polling has some inherent flaws that, no matter how well the study is conducted, still affect the results. One is marginal error, which is a statistical calculation that assesses how much random sampling error exists in a given survey. Basically, it’s a formula that takes into account the fact that a sample of 1,000 people is not going to perfectly represent the views of the whole population from which those 1,000 people were selected. The larger the sample size is, the less margin of error exists.
Another flaw with polls is where the sample is drawn from. This is an especially important point in a country of roughly 330 million people where certain geographical areas and institutions are skewed toward one political ideology.
“If you wanted to measure how the American population as a whole was going to vote and all you did was poll college students at all the giant universities, you’d have a giant sample,” said Professor of Political Science Shamira Gelbman. “There would be a tiny margin of error, but those college students aren’t representative of the whole population of voters.”
Polls can also have demographic issues, like weighting. A sample, even a diverse one, is unlikely to accurately reflect every demographic in a population. Some groups will be overrepresented and others underrepresented. Weighting allows survey collectors to adjust the edges of their data and address imbalances to more accurately reflect the whole population. Weighting is common practice in polls, but it is by no means perfect.
“One thing we discovered in the 2016 election was that, with all the weighting of the poll results being done, pollsters weren’t weighing educational attainment, and that turned out to be a really strong indicator of how people, especially white people, voted in the election,” said Gelbman. “Some people argue that had pollsters made those adjustments, they might have been able to anticipate more how the election eventually turned out.”
Even if national polls were perfect, they still wouldn’t tell us with certainty who will win the election. The reality of our electoral system is that the winner of this year’s election will likely carry more of the “swing states” than their opponent. A candidate may be up in national opinion polls, but if they lose important battleground states, our electoral college system might just crush their dreams of the White House.
So, are the polls this year accurate? As well as polls can be, but when a race is this tight, it’s basically impossible to be confident about a winner.
“Have pollsters overcorrected from prior years?” asked Gelbman. “Maybe, maybe not. We won’t know until the election. Since the start of polling, there have been years where the polls are dead on, and some years where they’re way off.”
“Honestly, from my perspective, I think polls are great, but they’re not the ballots being cast,” said Govekar. “Polls should be used as a way to gauge where the majority of people are at.”
So, ultimately, forget the polls and just get out and vote! In a tight race like this one, you never know if your vote will make a difference. And as a reminder, if you vote on Election Day in Montgomery County you are invited to participate in PSC-211’s exit poll.