The Silent Majority
The 2020 Republican National Convention painted a grim portrait of a desolate, crime-ridden wasteland should Democrats take the White House. Invoking “law and order,” Donald Trump transparently leans on Richard Nixon’s successful 1968 campaign strategy of weaponizing fear to turnout voters. But law and order isn’t the only tired Nixonian trope the President is banking on. The silent majority, used by Nixon to describe the Americans that were not anti-war demonstrators, was repurposed by the Trump campaign in 2016 to explain his low polling. Donald Trump’s surprise win suggests the polls missed something in 2016. And it’s convinced many that Trump voters everywhere, pressured by their peers into silence, didn’t speak up until Election Day.
These “shy Trump voters” — so the thinking goes — are the reason the polls failed to capture Donald Trump’s win. As Joe Biden maintains a lead on Donald Trump, a distrust of polling emerges from both sides of the race: an anxious fear on one side, and a quiet hope on the other.
But what actually happened in 2016? And what does it mean for 2020?
Politically Correct Polls
In 1982, the California Democratic Party nominated Los Angeles mayor Tom Bradley for the Governorship, pitting him against the former state senator and Republican George Deukmejian. Although Bradley maintained a significant lead in the polls ahead of the election and exit polls on the day of led some to project a Bradley win, he lost narrowly: 49.3% to 48.1%
Bradley was an African-American, and some believe a statistically significant portion of white voters told pollsters they were going to vote for him, when in fact they would not. According to this view, these voters only said they would vote for Bradley because they felt it was the socially preferred, or desired, answer pollsters wanted to hear; that they were going to vote for the African-American in the race.
Political scientists and pollsters are still divided on this, but those who hold this view of Bradley’s defeat lean on what’s called the social desirability bias- a cognitive bias phenomenon that occurs in survey respondents. The idea is that when pollsters ask questions, respondents answer in a way that presents themselves in a better light, even when that presentation does not reflect how they truly feel. The Bradley effect- applying social desirability to elections when race is involved- would seem then to explain discrepancies between polls and results in similar elections.
A New Spin
Fast forward to 2016. The Trump campaign attributed his bad polling to a new form of social desirability bias: That Trump voters felt shamed for their support of him so they lied in surveys, telling pollsters their votes would go to Clinton. A new spin on the same idea, the campaign claimed Trump voters were afraid to tell pollsters the truth.
But this wasn’t just your typical overcompensating hubris of campaigns that aren’t looking too hot: 71% of GOP insiders at the time were convinced that Trump’s support was being underrepresented in the polls. When election day came and the results began to turn against expectations (including Trump’s own), what was a way to scoff at poor performance in the polls became a real, possible explanation of the polls’ apparent failings.
But is it?
Understanding Polls
It cannot be overstated that polls are not forecasts; they are a snapshot of a given moment in time. They tell us how things look, not how they will look. Additionally, national polling in 2016 was not wrong — in fact, it was strikingly accurate. Polling averages like Real Clear Politics’ in the final days told us that if the given moment held, Clinton could win nationally by about 3 points. She won the popular vote by 2.1.
Polls also showed Trump closing the gap in critical states the week of then-F.B.I. Director James Comey’s letter to Congress, stating the department was reviewing additional emails in Clinton’s private server scandal only 11 days before the election.
That being said, there were problems with battleground state polls, and for a number of reasons. But the question here of silent Trump voters requires a closer look at one of those reasons in particular.
Shy Trump Voters In Wyoming?
A few weeks before the 2016 election, a Politico/Morning Consult study addressed the ‘shy Trump voter’ idea. The study found Donald Trump performed better in polls conducted online versus live interviews with pollsters over the phone. Live phone interviews showed Clinton with a 5 point lead, but in online polling her lead shrank by 1 point while support for Trump went up by the same.
Although the study did find a considerable gap between phone and online polling with college-educated voters — giving Clinton a 21 point lead in phone polls while only a 7 point lead online — the study concluded this had no meaningful impact. Statistically insignificant, and not enough to move the needle.
And this bore out in the election results. Trump did end up outperforming the polls in a handful of states. But if the shy Trump voter theory held, as FiveThirtyEight’s Nate Silver noted just after the 2016 election, we might expect this to be the most true in blue states. After all, wouldn’t a shy Trump voter feeling suppressed by the political environment likely be ‘shier’ in California than in, say, Wyoming?
But the opposite was true. Trump didn’t outperform his polling numbers in blue states. In fact, he underperformed polls in the bluest states, such as California where he underperformed by 4 points and Washington where he underperformed by about 3. As for the states where he outperformed the polls? Republican dominated states, including Wyoming and Utah — arguably the reddest states in the Union — where he outperformed by about 6 and 8 points respectively. If Trump voters were shy, they were shy in red states; the states he most likely was going to win anyway.
So while the Politco/Morning Consult study did show evidence for “small pockets of shy Trump voters,”it wasn’t enough to actually change the results. This conclusion was further affirmed when the American Association of Public Opinion Research published a more thorough study later on after the election.
Donald Trump’s surprise win didn’t happen because of a silent majority. But he may have won because of an undecided majority, and this is where a real threat to Biden’s prospects may lay in wait.
Undecided Voters in 2016
By the end of October 2016, 15% of the electorate was undecided. FiveThirtyEight’s forecast model was one of few that made space for this factor. As Nate Silver wrote on October 25th of that year:
“That’s one of the reasons why our models still give Trump an outside chance at victory. In theory, with Clinton at ‘only’ 46 percent of the vote, he could beat her by winning almost all of the undecided and third-party voters.”
According to exit polls, 13% of voters remained undecided until the last week of the election. Undecided voters broke for Trump in Michigan by 11 points and by 17 points in both Pennsylvania and Florida. In Wisconsin, undecideds cast their ballots for Trump by nearly 30 points.
In sum, four states tipped and changed the election because final-week voters broke for the Republican candidate overall by 3 points. The result? 76 electoral votes for Donald Trump, handing him the race.
It was undecided voters, not shy voters, who gave Donald Trump the election.
There has been plenty of commentary as to why undecided voters flipped to Trump at the midnight hour. But one thing we can be certain of is that by September, both candidates were disliked by 63% of registered voters. So-called October Surprises, the explosive news stories strategically dropped near the election, have more sway with undecided voters. Indeed, the Hollywood Access tape leak, the Wikileaks DNC email drop, and Comey’s letter made 2016 seem like a contest of which candidate could out-scandal the other.
All of these factors combined made for a severely unstable race in the Electoral College, introducing an uncertainty in voter opinion any time either candidate lost a news cycle. We may never know just to what extent Comey’s letter impacted Clinton’s chances, but forecast models like FiveThirtyEight’s saw a dramatic shift: Clinton’s previous-week lead on average in swing states, then at 4.5 points, dropped to 1.7. When the election came, Donald Trump swept in with winning voters who disliked both candidates by 17 points.
We could also discuss the faults in Clinton’s campaign strategy, especially in Pennsylvania, Michigan, and Wisconsin. For a long time following the election, most of us blamed one of two factors to explain Trump’s surprise win: Comey’s letter or Clinton’s campaign. If I’m being transparent, I fell into the latter. But in reality, all of these factors combined contributed in some way or another, and given the unprecedented dislike for both candidates, undecided voters held massive decision power.
But if the mythical silent majority is just that — a myth — our question still remains: What about 2020? Can undecided voters still deliver for Donald Trump in November?
Undecided in 2020
The short answer is Yes. But the conditions aren’t quite the same.
Undecided voters in 2020 are a dramatically different demographic than they were in 2016. They are majority Millennial or Gen Z voters and are more likely to be non-college educated. Most of them are women and most of them are people of color. This means different messaging will have to be tailored to reach these voters.
Undecided voters this year also currently make up about 7–10% of the electorate. That is not only lower than it was at this same time in 2016, but it is lower than it was in the last week of the election. But we should make no mistake here: it’s still enough voters to be more swayed by episodic events or decide at the midnight hour and swing the election.
However, Joe Biden isn’t as disliked as Hillary Clinton was. Biden is performing significantly better than Clinton, even at her strongest moments, in both national and state polling. His lead is currently more stable. The national environment is also currently playing in favor of Democrats on the generic ballot, a backdrop Clinton didn’t have. For voters who dislike both candidates this year, they lean toward Biden by 40 points, more than 20 points more than the President won them by in 2016.
This is why the Trump campaign is scrambling in its messaging. Joe Biden doesn’t scare certain voters or off-put other voters the same way Clinton did, especially the voters Trump won in 2016 that he is currently hemorrhaging: suburban voters.
Donald Trump continues to rely on his greatest hits, tired tropes, and an old bag of tricks- or whatever analogy you want to use. But it’s a mistake; none of his punches are landing. If law and order was going to work, for instance, the events in Kenosha were the ripest of conditions for it. But it appears to have only strengthened the positions voters had already chosen. Biden’s lead remains fundamentally unchanged.
A Silent Minority
The Trump campaign insists that its winning strategy is mobilizing nonvoters in 2016 that would have voted for him if they had. And, to be clear, in swing states these nonvoters do prefer him. But the primary demographic of these eligible voters, white Americans without a college degree, is shrinking rapidly.
Turning out nonvoters is really difficult to do, and it’s no good place to be if you’re simultaneously needing to claw back Americans who voted for you the last time (emphasis on need — suburban voters will decide this election) and turn out nonvoters to make up the difference.
When you’re down to as fine of electoral math as that, then you’re already underwater.
The President additionally cannot seem to let go of his attack on Biden’s mental acuity, despite its failure to stick. It’s a poor strategy: Trump keeps lowering the bar for Biden, giving the former Vice President opportunities for easy wins. If Trump continues down this path toward the ever-anticipated debates, the first being on September 29th, he is only setting Biden up for another easy win. Trump will be left to defend, overcompensate, and praise his own mental acuity, person, woman, camera, T.V.
The Myth of Donald Trump
Anti-Trumpers and die-hard supporters alike have fallen prey to the idea that Donald Trump is invincible. It would seem so, given his surprise win and the feeling that scandals throughout his Presidency don’t ever seem to stick to him. Indeed, his approval rating remains remarkably stable as it has throughout his Presidency.
While it is true this Presidency goes through scandals like scandals are going out of business, remarkably stable, however, depends on context. Because his approval rating is also remarkably and consistently low for any President. So the point needs to be made:
Donald Trump won in 2016 because of four states, two of which he barely broke 1% and one of which he won by less than 1%.
That isn’t invincible.
None of this is to say however that there is no chance the President can pull off another upset win. With 7 weeks to go, this election is far from over. Trump has closed in on the gap in some key states like Florida.
If 2016 can teach us anything, it’s that nothing is ever certain and that every voter should vote as if their candidate is 5 points behind, no matter what the polls say.
Most importantly, more than a mass of closeted Trump voters waiting in the wings, we should be prepared for anything but for a simple reason:
It’s 2020.
And if Tiger King wasn’t the weirdest thing to happen, then anything can happen.
