Prime Only Members-Only Article
Editor's Brief

Post-Traumatic Stress, Polls and Trump Magical Thinking

The Leaked Poll Saga Has Forced a Reckoning with Trump's Profound and Enduring Unpopularity
WASHINGTON, DC - MAY 06: U.S. President Donald Trump speaks as he hosts the U.S. Military Academy football team, the Army Black Knights, in the Rose Garden of the White House May 6, 2019 in Washington, DC. President... WASHINGTON, DC - MAY 06: U.S. President Donald Trump speaks as he hosts the U.S. Military Academy football team, the Army Black Knights, in the Rose Garden of the White House May 6, 2019 in Washington, DC. President Trump hosted the Commander-in-Chief’s trophy champion to honor their win in 2018. (Photo by Alex Wong/Getty Images) MORE LESS
June 18, 2019 9:01 a.m.

We’ve all been watching this Trumpian series of lies, denials and special pleadings tied to the leak of internal Trump campaign polls from March. The numbers showed Trump being defeated decisively for reelection and Trump himself has repeatedly denied exist. Campaign manager Brad Parscale eventually conceded the polls did exist. But he insisted that they’re months out of date and are too early in the campaign cycle to count in any case.

To a significant degree, he’s right. We’re almost a year and a half before the election. But the whole spectacle illustrates a reality I mentioned a couple weeks ago: the political nation is in a collective state of denial about the depths of the President’s unpopularity and his uphill challenge seeking reelection. The denial isn’t based on nothing. The 2016 outcome was a shocking as well as a calamitous result, one based on significant albeit sometimes overstated polling misses in the key states of Pennsylvania, Michigan and Wisconsin. Still, it’s warped our collective understanding of just what’s happening in the 2020 cycle.

Let me share portion of a note I got in response to that earlier post (“The Alternative Scenario: Trump Loses and It’s Not Even Close“) from a highly respected public pollster.

I’ve been turning a similar line over in my mind about how to balance evidence that this is a decidedly uphill battle for re-election (bad approval, 2018 state shifts, a more attractive Dem than Clinton) against a good economy and most presidents get re-elected absent a recession. And the uncertainty due to our errors in WI, MI and PA where over 100 polls had Trump behind regardless of method, mode or weighting. I think the answer is late decisions and an unattractive Dem alternative, but the trauma of being wrong with so many polls makes one especially cautious about 2020.

At this point I’m going to trust the data and try not to be unduly cautious. But hard to ignore last time entirely.

The trauma of being wrong really captures it. Trump has pulled off the seemingly impossible once. We can’t discount the possibility. It’s a snake bite feeling that goes beyond rational caution. But we also can’t overstate it, and in so doing make luck and uncertainty and Trump’s own sort of evil genius into a kind of magical power.

Let’s go back and review the precise nature and extent of the error. The consensus of polls on the national popular vote was only off by about 1% — pretty close to dead on. But that’s not how we elect Presidents. Trump won Florida, North Carolina, Ohio and Iowa. But those were all considered well within the range of possible outcomes. It was Wisconsin, Pennsylvania and Michigan that provided the decisive margin and the big shock.

But how far off were they? Let’s look at the final RCP Averages in each race. In Pennsylvania the final poll average was Clinton +2.1 and the actual result was Trump +.7 For a four way race, which is of course what it actually was, it was a slightly closer Clinton +1.9. For Michigan it was a final average of Clinton +3.6 and the final result Trump +.3. (Four way race had Clinton +3.4.) In Wisconsin the final average was Clinton 6.5% and the final result Trump +.7.

The results shouldn’t have been that surprising in Pennsylvania and Michigan. Those are misses but well within the inherent variability of polling. (After the fact studies have pointed to a late surge of support for Trump that wasn’t captured by the polls. Others point to the high number of undecideds and the complexity of a four person race.) It’s Wisconsin that stands out as really out of nowhere and a major miss. But it’s important to look at the actual data and not the shellshocked feeling so many people had. In retrospect there was more than a little reason to think there was a problem in Michigan and Pennsylvania. People were shocked because it was very hard for people to imagine Trump actually getting elected, quite apart from poll numbers. Whether it was experience or wishful thinking or experience, many simply assumed that partisan muscle memory would keep these states in the Democratic column. That was helped along by the fact that a number of the late polls that pointed to a tightening were either Republican polls or ones that tend to be GOP-friendly.

These are shocks, real shocks, consequential shocks. I don’t mean to discount them or argue them away. But they are more limited than people remember. And they’ve led to a pervasive species of denial and magical thinking that shadows all discussion of the 2020 election.

At one level, it’s good to run like you’re behind. Every Democrat should fight like the election is ultimately winnable (demoralization is never good) but a critical, challenging struggle requiring universal effort. But I’ve also wondered about how much the persistent belief in Trump’s magical powers enervates the opposition and vitalizes Trump himself. Remember that Trump and Trumpism are ideologies of aggression and dominance. Weakness is not only disappointing for followers. It undermines the whole premise of the movement. Not taking cognizance of this is not only misleading and enervating for Trump’s opponents. It deprives them of a significant offensive capacity.

Here we should ask too: why were these numbers leaked? There’s good reason to think it’s because people inside the Trump campaign realized that Trump’s actual situation is so different from what he claims and what a lot of people – including a lot of his supporters – seem to think that it’s actually become a danger to the campaign. The Post said as much in a sharp recounting out this morning.

There are also reasons to think that the campaign is confecting improved poll numbers to satisfy the President’s demand for good news. On Friday the Post’s Aaron Blake had a good dissection of Trump Campaign Chair Brad Parscale’s explanation of the bad polls. Along with comments from spokesman Tim Murtaugh, Blake argues convincingly that what the campaign claims is a dramatic turnaround since March has come in large part because they simply goosed the numbers. They’ve apparently done subsequent polls with a so-called ‘informed ballot’.

Basically this means prepping respondents with tendentious claims about your opponent. So instead of asking whether a respondent prefers Trump or Biden you start by telling them that Joe Biden is going to take away your health care and give all your money to MS-13. As Parscale says in his statement: “Since March, we have seen huge swings in the President’s favor across the 17 states we have polled, based on the policies now espoused by the Democrats. For example, the plan to provide free health care to illegal immigrants results in an 18-point swing toward President Trump.” In other words, it appears that the campaign is getting these numbers by baiting respondents with various anti-Democratic attack lines.

This tweet from campaign Director of Communications Tim Murtaugh, as Blake notes, seems to confirm that interpretation.

The key line: “The new data tested issues Dems are running on. In all 17 states polled, POTUS leads.”

This kind of polling can have a legitimate purpose. Campaigns will test lines of attack using this method to see how much they move respondents from the clean numbers. Here though they seem to be using this method to get more palatable results.

President Trump got the message and is pushing the new message of big leads in all 17 swing states at every turn.

Taken together this internal polling story seems to have spurred an inflection point in perceptions of Trump’s standing. An accumulation of data that simply overwhelmed the storyline or Trump’s unbeatable base and his magical powers. Internal polls of unreported methodology almost two years before election day (remember these were in March) shouldn’t upend perceptions like that, certainly not on their own. But I don’t think they have. It’s more that the spectacle of the leaks and denials has focused people’s attention on more widespread, public and probative evidence of Trump’s unpopularity: more than two years unable to climb above support in the low 40s, a midterm wave, narrow majorities of voters expressing an absolutely unwillingness to ever vote for him again, steep unpopularity in his 2016 breakout states. US Presidents tend to get reelected. The myth of Trump’s magical powers and his unbeatable base is the joint creation of his most dieheard supporters and most dogged critics. But it’s not borne out by the evidence.

Latest Prime
Masthead Masthead
Founder & Editor-in-Chief:
Executive Editor:
Managing Editor:
Senior Editor:
Special Projects Editor:
Editor at Large:
General Counsel:
Head of Product:
Director of Technology:
Associate Publisher:
Front-End Developer:
Senior Designer: