Making Sense of Pre-election Polls: How Not to Feel Overwhelmed This Fall

Here are four straightforward suggestions to help navigate the deluge of pre-election polls in this year's presidential race. They are offered to help avoid the whiplash and confusion that a profusion of polls can produce. The suggestions that follow are drawn from my research into pre-election polling and its failings.

1. Avoid focusing too keenly on any single poll. It may be tempting to do so, especially to focus on results released by the few pollsters who did well in 2020. Or to ignore the data of pollsters who were notably poor performers four years ago. But past polling performance is no guarantee of accuracy in future elections.
   
   Instead, try to consider polling results broadly, as presented by an aggregator of polls such as RealClearPolitics.com. The RealClear site is updated frequently with fresh polling data. For weeks, the polling averages of RealClear have signaled a close race between former President Donald Trump and Vice President Kamala Harris.
   
   That's been the trend; it's not necessarily a prediction. After all, Trump won the presidency in 2016 in part because late-deciding voters in Michigan, Pennsylvania, and Wisconsin swung their support to him in the campaign's closing weeks. Trump carried each of those battleground states narrowly. And with them, the presidency.
2. Keep in mind this campaign's great known unknown. Which is whether the polls will underestimate Trump's support, as they did in 2016, and did again in 2020, when they collectively turned in their worst performance in 40 years. A number of pollsters have since modified the techniques by which they seek out would-be respondents. For example, some pollsters have turned to a combination of phone- and web-based techniques in hopes of improving their samples.
   
   It won't be known till after the election whether pollsters succeeded in reaching enough Trump supporters to measure his support accurately. Many pollsters thought they had accomplished just that in 2020, after making adjustments following the 2016 surprise outcome. Although the polls pointed accurately to Joe Biden's winning the presidency in 2020, they largely understated Trump's support.
3. Pay small notice to political defectors and apostates. Personalities like former Congresswoman Liz Cheney and her father, former Vice President Dick Cheney are both Republicans who despise Trump and have endorsed Harris. Neither Cheney, however, has a constituency that will translate into millions, or even thousands, of votes for Harris. That holds true for former Congresswoman Tulsi Gabbard, a onetime Democratic presidential candidate who is backing Trump. Likewise, Robert F. Kennedy Jr. probably won't influence cadres of supporters to vote for Trump, whom Kennedy endorsed after his independent candidacy withered during the summer. Political apostates aren't all that uncommon in American politics. Former Democratic presidential nominees Al Smith and John Davis endorsed Republican Alf Landon in 1936 - when Landon lost to Franklin D. Roosevelt in a landslide.
4. Eschew the cliché that polls are merely "snapshots in time." It's the laziest aphorism in survey research and is invoked not infrequently as a metaphoric shield for pollsters whose pre-election surveys go awry. In such cases, "snapshot in time" becomes a rationalization. The cliché also can be a convenient ploy for politicians and their supporters to dismiss or scoff at poll results that contradict their partisan preferences. As I wrote this spring in a commentary for the Hill, "snapshot in time" is meant "to suggest impermanence - that polls taken weeks or months before an election have limited predictive value. … It will be heard many times before the campaign ends."
   
   The sheer number of polling "snapshots" together comprise a panorama that offers useful insights about a presidential campaign. The panorama of polls has signaled consistently that the 2024 race may be close, but with no political landslide ahead.

W. Joseph Campbell is a professor emeritus of communication and author of seven non-fiction books, all of which were published during his 26 years on the SOC tenure-line faculty. His most recent work is "Lost in a Gallup: Polling Failure in U.S. Presidential Elections" (University of California Press, 2020 and 2024).