Dr. Dorian Hunter Davis
Associate Professor in the School of Communications at Webster University in St. Louis, Missouri. He teaches and conducts research on social media in news and politics.
Email: doriandavis39@webster.edu
U.S. Election 2024
25. Seeing past the herd: Polls and the 2024 election (Dr Benjamin Toff)
26. On polls and social media (Dr Dorian Hunter Davis)
27. How did gender matter in 2024? (Prof Regina Lawrence)
28. The keys to the White House: Why Allan Lichtman is wrong this time (Tom Fisher)
29. Beyond the rural vote: Economic anxiety and the 2024 presidential election (Dr Amanda Weinstein, Dr Adam Dewbury)
30. Black and independent voters: Which way forward? (Prof Omar Ali)
31. Latino voters in the 2024 election (Dr Arthur D. Soto-Vásquez)
32. Kamala’s key to the polls: The Asian American connection (Nadya Hayasi)
33. The vulnerability of naturalized immigrants and the hero who “will fix” America (Dr Alina E. Dolea)
34. Did Gen Z shape the election? No, because Gen Z doesn’t exist (Dr Michael Bossetta)
35. Cartographic perspectives of the 2024 U.S. election (Prof Benjamin Hennig)
Just before Election Day, USA Today posted on X (Twitter) about the results of a Selzer poll showing Kamala Harris leapfrogging Donald Trump to a 3-point “lead” in Iowa. If true, that would have been a remarkable change from 2020, when Trump won the state with 53% of the vote. That result fueled speculation that other pollsters showing tied results in the swing states were “herding,” or weighting their results to match everybody else’s, inspired fawning memes about pollster herself, and gave Harris supporters a last-minute confidence boost. But that post, and posts about the same poll from several other major news outlets, left out at least one important detail – that Selzer’s results showing Harris ahead were inside the margin of error. As we know today, Trump went on to win Iowa and the presidency.
This speaks to one of the challenges of reporting polls on social media. In traditional news stories, the Associated Press recommends that journalists include enough information about polls for “readers and viewers to evaluate the results for themselves.” In one example from The AP Stylebook, that amounts to five methodological details – the sponsor of the poll, population, sample size, selection procedure, and dates the poll was conducted. The AP also mentions several other details like margin of error that could be useful as well.
But there are no such guidelines for poll disclosure on social media. And, as a result, there’s less of it. Just for a quick example, I analyzed over a hundred posts about presidential match-up polls from 15 major news outlets on X during the month before Election Day. Only about 10% included four or more methodological details. And, while mentions of a poll’s sponsor, population, and some indication of when it was conducted were relatively common, just about 10% referenced margin of error.
Admittedly, there are good reasons for leaving some details out. For one thing, there’s less space on social media. For example, on X, where a lot of political news breaks, accounts that aren’t subscribed to X Premium have a 280-character limit. And even apps that allow more text or video tend to reward brevity. Journalists and news media also tend to use their social media accounts more as promotional tools, driving traffic to their websites. And, in that context, it makes sense to treat their posts about polls more like teasers than comprehensive stories with all the relevant methodological details. And making their social media posts appealing and accessible to consumers is an important consideration as well. Many people are more interested in the results of a poll than the methodological details and may not know how to assess some of those details, like weighting procedures, anyway.
On the other hand, there are compelling reasons for journalists to publish more context for polls on social media. For one thing, those details can make it easier for consumers to interpret the results for themselves. Even reputable news sources can misinterpret or misrepresent a poll. In fact, research has shown that journalists often frame stories about election polls in terms of change even when the polls themselves show no change outside the margin of error. We saw a version of that phenomenon this year on social media. For example, this Fox News post about Harris taking a ”slim lead” over Trump in a NYTimes/Siena poll referred to results that were inside the margin of error. You had to read the article on Fox News’s website to find that out though. This kind of framing can drive media narratives about potential outcomes and even affect voter behavior.
As a compromise, I would encourage journalists and news media to apply AP standards for covering polls to their social media. And that can be done without sacrificing brevity and readability. For example, The New York Times attached an image including six details – the sponsor, population, sample size, phrasing of the question, dates conducted, and margin of error – to this post about one of its own match-up polls. Admittedly, applying AP standards wouldn’t eliminate confusion about polls on social media altogether, but it would be a step toward more transparency.
I’m not dismissing the challenges that news media face in determining how to package their stories for social media. There are different and competing interests at stake, and legitimate reasons that some outlets would hesitate to share the minutia of polling data on their social media. But social media are news access points for millions of people now, and journalism should adjust to that reality.