Editor’s Note: Rich Thau is the president and co-founder of the research firm Engagious, which specializes in message testing and message refinement for trade associations and advocacy groups. He is also the moderator of the Swing Voter Project. Follow him on Twitter @richthau. The views expressed in this commentary are his own. View more opinions on CNN.
Death and taxes are certain. Polling forecasts are not.
We re-learned that lesson last week when swing states like Florida and Wisconsin performed far better in pre-election surveys for Joe Biden than with actual voters on Election Day. Trump won Florida despite polls putting Biden in the lead there and the President narrowly lost Wisconsin, where pollsters had predicted a much larger margin for Biden.
If setting the public’s expectations about election outcomes is a meaningful exercise, then the public opinion profession needs to take stock and be more precise the next time.
To help achieve that goal, here’s a modest proposal: Ensure that publicly released quantitative research, like polling, is continuously coupled with qualitative research like focus groups, in-depth interviews, and ethnographic research.
The insights gleaned from open-ended exchanges are far different — and sometimes far more useful — than those generated from closed-ended questions in surveys. When used in tandem, these methodologies provide insights that none alone can offer.
I know this firsthand. For the past 21 months, I’ve been living a split-screen existence in anticipation of the election, conducting my own focus groups in swing states while religiously following the polls.
As I explained in an op-ed for CNN.com in July, my firm’s focus group research was repeatedly out of sync with election survey results at the time. Pollsters were regularly telling us then that Democratic nominee Joe Biden was far ahead. I wasn’t then — nor am I now — looking to pick a fight, as the sample sizes of pollsters are much larger than mine. That said, as a focus group moderator, I consistently heard strong support for President Donald Trump from a critical sliver of the electorate.
That sliver consisted of so-called “Obama-Trump” voters across swing counties in Florida, Iowa, Michigan, Minnesota, Pennsylvania, Ohio and Wisconsin. We held monthly focus groups with them from March 2019 through November 2020 as part of the Swing Voter Project.
Throughout those focus groups, with rare exception, we invariably found approximately two-thirds to three-quarters of respondents siding with President Trump in a hypothetical matchup with former President Barack Obama, or in the eventual matchup with former Vice President Biden.
The fact that Biden was not winning back a higher proportion of one-time Obama voters always struck me as noteworthy. By their comments, those supporting Trump were dug-in and not likely to flip to Biden.
For reference, focus groups are early detection systems of shifting public opinion. Before something important appears in polling, it often surfaces first in focus group conversations.
This led me to feel confident writing the following to clients last month:
“Based upon what I’m hearing, about three-quarters of these swing voters will stick with the President. It’s TBD whether having Biden win back that other one-quarter means he will secure enough votes to win the election. But even if Trump loses (which the polling suggests will happen), it does not mean Trumpism — and the economic nationalism that propelled him — is dead. America ignores at its peril the people I’m talking to monthly — the working people of the upper Midwest.”
It’s critical to understand why so many “Obama-Trump” voters continue to support Trump, even though Biden won the popular vote.
They think a businessman is best suited to turn the country around economically. They feel Covid-19 was not Trump’s fault, and believe he’s doing the best he can to contain it. They unconsciously conflate the Black Lives Matter protesters with the rioters attacking federal buildings and retail shops. And they dismiss defunding the police as endangering their security.
These voters told me they want America finally to be put first; they oppose immigration and trade policies that they say give benefits to foreigners at their expense. They fear socialism. And they want a non-politician who relentlessly fights back, after witnessing too many officeholders fold in the face of special interests.
These voters may sound like typical Fox News watchers, but the overwhelming majority are not. Instead, many are people who get their news disproportionately from local television, regional websites and Facebook. Compared to the kinds of people who seek out news from national cable channels, many swing voters reside in a national politics desert.
One new and eye-opening trend has emerged recently that merits our attention: The prevalence of widespread conspiracy-mongering by these “Obama-Trump” voters, which CNN’s Dana Bash has reported on.
A newly released study from the University of Southern California suggests a correlation between sizable polling discrepancies in certain swing states, and a disproportionate interest in QAnon conspiracy theories.
Get our free weekly newsletter
In focus groups with voters in Ohio in late September and Michigan in early October, respondents offered a wide range of wildly unsubstantiated allegations questioning the morality of people in Hollywood, the severity of the pandemic, and the quality of Vice President Biden’s performance in the first debate.
My admonition remains: Pay a lot of attention to those voters who don’t pay much attention at all. Also pay a lot of attention to what they see on social media, especially Facebook and YouTube. They are issuing an important warning about America’s future.