Damn lies and statistics. Coming out of a honeymoon for pollsters in the Obama years when industry professionals like Nate Silver gained celebrity status for their work, polling has since come up short dramatically in the last two presidential cycles. Polls once again proved inaccurate in the 2020 election cycle and experts are still discussing why. Donors should also be paying attention as questions persist about the reliability of polling, which plays a big role in how money is raised and spent in elections. Whether you’re giving directly to campaigns or to national organizations, you may be operating under wrong assumptions if you rely on public polling. Or the money might be misused if strategists are working from poor data.
During the 2020 election cycle, large and small donors alike pumped unprecedented funds into races that leading Democrats signaled were most in play—priorities developed from polling. Those off-base numbers led to tens of millions of dollars misallocated and disappointing results for Democrats on election night.
The (persistent) problem with polling
A recent study from the American Association for Public Opinion Research analyzed some of the reasons polls were off yet again in 2020, but notably, were unable to identify a clear explanation. The study found national presidential polls were the worst in 40 years, and state-level polls were the worst in 20 years, with most national polls underestimating support for President Donald Trump. Among the possible explanations are conservative voters’ ever-growing distrust of the media and polls themselves, making them unwilling to answer the phone or give a detailed answer to a caller—something the study noted is difficult to confirm since you can’t know who isn’t responding.
But the report noted that even worse than the presidential polling in many states was the Senate race polling, which on average was off by six points. The phenomenon attracted countless headlines bemoaning poor data shortly after Election Day in 2020, and experts have spent the time since trying to pin down what happened.
“Even seven months after the fact, you’d think you’d be able to know exactly what happened,” Josh Clinton told Politico—he’s chair of the AAPOR election task force. “We’ll have to wait and see what happens—which isn’t a particularly reassuring position. But I think that’s the honest answer.”
Bad polling in 2020 is particularly disappointing given how bad polling was in 2016 and the efforts made to correct the problem in between the two election years. When Trump won the presidency, it shocked just about everyone and the early conclusion was that Trump and his supporters were so chided by the public that people kept quiet about supporting him—a cohort representing what is sometimes called the “secret Trump voter.” Pollsters tried to account for demographic divides that may have contributed to skewed numbers in 2016 when conducting 2020 polls, but the figures still didn’t reflect what the vote totals ended up being in November.
“For me, there was a little bit of déjà vu all over again, watching some of these results come in,” said Robby Mook,president of the House Majority PAC and Hillary Clinton’s 2016 campaign manager, speaking to a video panel at the Harvard Kennedy School just after election night in November. “I kind of come out of this with a lot of questions as people think about how to strategize for the midterms.”
Mook went on to echo the theory that Trump voters were uniquely avoiding talking to pollsters—noting that 2016 and 2020 had similar problems not seen in the 2018 midterms. While Trump may have had a special effect on polling during his two campaigns, his distrust of the media and polls themselves has become commonplace in the Republican Party, even with him out of office. If the theory that conservatives are unwilling to cooperate with pollster calls is what’s driving the bad numbers, there’s reason to expect this problem will persist in the coming cycles.
What donors can do
Whether polling can be fixed or not, there are still steps that donors can take to avoid misdirecting donations due to bad data. To start with the most obvious: Don’t trust individual polls, which are unreliable. Donors should always be sure to look for aggregated averages instead of the results of just one sample. Politicians often use individual polls to create urgency for their campaigns—for example, a fundraising appeal from a Democratic group supporting Kentucky Senate candidate Amy McGrath last year noted that the “latest poll” had her within “striking distance” of incumbent Republican Mitch McConnell (the email did not link to the poll or describe its methodology). Jamie Harrison’s 2020 senate campaign also routinely cherry picked polls to tell a misleading story. As a general rule, donors should never rely solely on polling numbers provided by any candidate or organization as part of a fundraising appeal.
Another takeaway from recent history is that donors should prioritize giving to electoral groups that really know what’s happening on the ground. Before giving money to a major organization that may have limited connection to the electorate in a given place, and therefore be more likely to rely on polling, donors should explore grassroots and smaller organizations where their money will go directly to organizing work and has fewer variables. Donors should inspect to see which organizations are taking proactive steps to improve the polling they’re working from, or else see if the group is using a spending strategy that’s less dependent on polling for 2022, and particularly for 2024.