Episode 21 - Lost in a Gallup
An interview with journalist, professor and author, W. Joseph Campbell about the co-history of journalism and polling
I’m going to try a slightly different format today, because I’m out actually doing interviews and focus groups in person with real people. And because there’s quite a bit on my mind, as well as new episodes dropping weekly, though not always on Tuesday because, well, I’m in field. So — Coupla Things.
Thing One: Default to (Dis)Trust
I’m finally getting a moment to write this post from an airplane seat on my way to Chicago to facilitate a creative consumer workshop. It’s only the second time since the pandemic that we’ve had a project where everything aligned and we could be in field, meeting real people, listening to them talk about their lives. I love it.
But I will say that things have changed since before Covid. It’s not just that clients have come to rely on remote methods of conducting research, and the way the industry has had to shift around their preferences. It’s also the way people behave in focus groups and workshops.
It’s not like what’s been reported about classrooms and other public places — people are still very nice to each other. But it takes longer to warm up.
There’s a process in focus groups that participants usually go through, a process known as Tuckman’s stages of group development. It looks like this:
You used to see people enter the group through that norming phase, but now, forming, storming, and norming are happening all at once. People come into the room excited, curious, shy, distrustful, uncertain, and even a bit defensive. They don’t default to trust — with me or with each other, as quickly as they used to.
As they start to transition out of the muddle and into the norming and performing part of the experience, things improve a lot — people are open, they disagree agreeably, it’s like the Old Days. But there is something else that I’ve noticed.
We used to notice in a lot of groups — especially anything related to raising kids and health — a lot of misinformation or knowledge gaps. That was always pretty normal — people are busy and can’t be subject matter experts in everything, so sometimes they confuse things, and sometimes they’re just guessing (and guessing wrong).
But now, I notice, it’s not just misunderstandings or knowledge deficits. It’s worse. We hear more conspiracy theories, more mistrust, more disinformation; we hear more about online influencers and grifters; and we hear people simply throwing up their hands, living with a combination of skepticism and YOLO.
And I’m just talking to them about consumer goods.
It feels like our default-to-trust, prosocial evolutionary tendency is not just getting hacked — it’s breaking. While we do want people to be skeptical, to ask questions, and to be critical thinkers, we don’t want people to surrender to the idea that no one and nothing can be trusted. That way lies antisocial behavior, and that is, by definition, madness.
Thing Two: How Polling & News Grew Up Together
I had the exceptional pleasure of talking at some length with the journalist, professor and author, W. Joseph Campbell about his recent book Lost in a Gallup.
The book is delightful — it lays out some central characters, founding fathers of modern polling, who made all the same mistakes and gambles that modern pollsters do. And it also lays out the familiar patterns reporters and columnists repeat election year, after election year.
Our “heroes” are George Gallup, Elmo Roper, and Archibald Crossley. Each, in their way, borrowed from the still-young methods of market research, and disrupted the old political pulse-taking method: the straw poll. Prior to the first scientific polls in the 1930s, it was the Literary Digest’s massive mail-in polls of readers — part of an effort to get attention and subscribers — that set the “gold standard” for political polling. But the Digest’s luck, inevitably, ran out after a few presidential cycles, and they called a clear victory for Alf Landon in the 1936 election; in the end, FDR won in a landslide. Meanwhile, Gallup’s polling came within 2 points of the final result. He and his fellow upstarts quickly formed partnerships with news organizations, establishing a symbiotic relationship that continues to this day.
In our conversation, we discussed a variety of themes that have remained relevant for almost a century:
Notable Polling Failures and Their Impact
Despite advances in methodology, polling failures have occurred regularly throughout history, often with significant consequences. The infamous “Dewey Defeats Truman” miscall of 1948 stands as a stark reminder of the pitfalls of overconfidence in polling. More recent examples, such as the underestimation of Donald Trump’s support in 2016 and 2020, highlight the ongoing challenges in accurately capturing voter sentiment. These failures inevitably lead to periodic soul-searching within the industry and efforts to improve polling techniques, though they often solve the last cycle’s problems and fail to anticipate how polling and the electorate are changing beneath their feet.
The Evolution of Polling Methods and Technologies
As technology and social behaviors have changed, so too have polling methods. The transition from in-person interviews to telephone polls marked a significant shift, while the decline of landlines and the rise of cell phones presented new challenges. Today, pollsters experiment with various approaches, including online panels, text-to-web surveys, and multi-mode polling combining different methodologies. The search for a new "gold standard" to replace random-digit-dial telephone polling continues, with no clear consensus on the most reliable method.
The Role of Exit Polls in Election Coverage
Exit polls emerged in the 1970s as a powerful tool for election night forecasting, quickly becoming a cornerstone of media coverage. However, high-profile missteps, such as the premature calling of the 1980 election and the 2000 Florida debacle, have led to increased caution in their use. The impact of early projections on voter behavior remains a subject of debate, with research suggesting that such calls may have less influence than previously thought.
Challenges in Accurately Representing Diverse Electorates
One of the persistent challenges in polling is ensuring that samples accurately reflect the diversity of the electorate. Underrepresentation of certain groups, such as non-college-educated voters in recent elections, can lead to significant errors in forecasts. The increasing polarization of the electorate and the phenomenon of "shy" voters who may be reluctant to share their true preferences add further complexity to the polling process.
Best Practices for Interpreting Poll Results
Given the inherent uncertainties in polling, consumers of poll data are advised to approach individual polls with caution. Focusing on polling averages, rather than single polls, can provide a more reliable picture of the electorate's mood. It's also important to consider the track record of specific pollsters and to be aware of potential sources of error. Ultimately, while polls can provide insight into public opinion and electoral trends, they should be viewed as estimates rather than precise predictions of election outcomes… no matter how much journalists and politics junkies wish it otherwise.
Resources
Lost in a Gallup by W. Joseph Campbell
This book explores the complex and often contentious relationship between polling and journalism throughout history, highlighting major polling failures and their impact on elections and public perception.
The Pollsters by Lindsay Rogers
Published just after the infamous "Dewey defeats Truman" polling failure of 1948, this book offered a timely critique of polling methods and their limitations.
Precision Journalism by Philip Meyer
Meyer's influential work advocated for the application of social science research methods in journalism, helping to establish a more data-driven approach to reporting.
Getting it Wrong by W. Joseph Campbell
This book challenges popular media myths, including the widespread belief that the War of the Worlds radio broadcast caused mass panic, demonstrating how such myths persist despite contrary evidence.
Our Guest
W. Joseph Campbell is an American writer, historian, media critic, and analyst. He has written seven solo-authored books, the latest of which, Lost in a Gallup (2020), examines polling failures in U.S. presidential elections. The book has been praised as “well-written, impressively researched, and detailed,” as a “fast-moving narrative history,” and as a “bracing reality check.”
Campbell is professor emeritus of communication at American University in Washington, D.C. He earned his doctorate in mass communication at the University of North Carolina at Chapel Hill. Before entering the academy, he was a professional journalist for 20 years. He earned his undergraduate degree at Ohio Wesleyan University which awarded him its Distinguished Alumni Citation at a ceremony in 2014.
Subscribe to Cross Tabs
Don't miss an episode! Subscribe to Cross Tabs on your favorite podcast platform: