Registered voters who didn’t vote on Election Day in November were more Democratic-leaning than the registered voters who turned out, according to a post-election poll from SurveyMonkey, shared with FiveThirtyEight. In fact, Donald Trump probably would have lost to Hillary Clinton had Republican- and Democratic-leaning registered voters cast ballots at equal rates.
Election-year polls understandably focus on likely voters. Then, after the election, the attention turns to actual voters, mainly using exit polls. But getting good data on Americans who didn’t vote is more difficult. That’s why the SurveyMonkey poll, which interviewed about 100,000 registered voters just after Election Day, including more than 3,600 registered voters who didn’t vote, is so useful.1 It’s still just one poll, and so its findings aren’t gospel, but with such a big sample we can drill down to subgroups and measure the demographic makeup of nonvoters to an extent we couldn’t with a smaller dataset.
Let’s look first at the most newsworthy finding: Registered voters who identified as Democrats and independents were more likely than Republicans to stay home.
| 2016 REGISTERED VOTERS | ||||
|---|---|---|---|---|
| SELF IDENTIFICATION | SHARE REGISTERED TO VOTE | VOTED IN 2016 ELECTION | DIDN’T VOTE | DIFFERENCE |
| Democratic | 44% | 44% | 35% | +9 |
| Republican | 43 | 46 | 32 | +14 |
| Neither | 13 | 10 | 33 | -23 |
Given how closely party identification tracks with vote choice, the disparity in turnout probably cost Clinton the election. SurveyMonkey did not ask non-voters whom they would have voted for, but we do know that more than 90 percent of self-identified Democrats who cast a ballot voted for Clinton and more than 90 percent of Republicans voted for Trump. Moreover, voters who didn’t identify with or lean towards either party were slightly more likely to prefer Clinton to Trump. That means that had the non-voters cast a ballot in accordance with their party identification, Clinton’s advantage over Trump nationally would have expanded by about 2 to 3 percentage points. That almost certainly would have been enough to flip enough states for her to win the Electoral College.
The large gap in party identification between registered voters who cast a ballot and those who didn’t also helps to explain why pre-election polling underestimated Trump. Pre-election polls suggested that the gap between these two groups would be smaller than in 2012; the SurveyMonkey data suggests it was larger.
The biggest reason given by non-voters for staying home was that they didn’t like the candidates.2 Clinton and Trump both had favorable ratings in the low 30s among registered voters who didn’t cast a ballot — both had ratings in the low 40s among those who did vote. That’s a pretty sizable difference. So why was Clinton hurt more by non-voters? Trump was able to win, in large part, because voters who disliked both candidates favored him in big numbers, according to the exit polls. Clinton, apparently, couldn’t get those who disliked both candidates — and who may have been more favorably disposed to her candidacy — to turn out and vote.
The second pattern that jumps out in the SurveyMonkey data: Non-white and Hispanic Americans were more likely to stay home than white voters.
Of all voters who cast a ballot in the general election, 25 percent were black, Hispanic, Asian, or a member of another minority group. But those voters were 42 percent of those who didn’t vote. Drilling down a little further, black voters made up 11 percent of voters who cast a ballot and 19 percent who didn’t. This disparity really hurt Clinton because black voters (by 82 percentage points) and Hispanic voters (by 40 percentage points) overwhelmingly favored her, while white voters went for Trump by a 16-point margin in the SurveyMonkey poll.
The turnout rate for black voters was substantially higher in 2012, the last time Barack Obama was on the ballot. According to the Census Bureau’s Current Population Survey,3 black Americans made up 13 percent of voters and only 9 percent of registered non-voters in 2012. In other words, black voters actually made up a larger percentage of voters who cast a ballot than those who didn’t in 2012, which is the opposite of what occurred last year. Whites, on the other hand, made up about the same percentage of registered voters who cast a ballot (74 percent) and those who didn’t (73 percent). The higher number of black non-voters in 2016 probably had a big impact.
Next up: Younger voters were more likely to stay home than older voters.
That matches a similar pattern from 2012, according to the Current Population Survey. That probably didn’t help Clinton, but it’s not as harmful as you might think because the difference in voting patterns between the oldest age cohort (a group Trump won by 12 percentage points in the SurveyMonkey data) and youngest (a group Clinton won by 30 percentage points) voters isn’t as large as it is between racial groups. Overall, the age breakdown of 2016 voters looks about the same as four years ago.
More harmful for Clinton was which young voters stayed home: minorities. Among white voters, voters 18-29 years old made up 30 percent of voters who did not participate in the November election. Among young Hispanic voters, that climbs to 43 percent. Among young black voters, it was an even higher 46 percent. That generally matches the findings of the voter data released in some Southern states showing that young black voters were especially likely to stay home in this election. Younger black voters were far more likely to support Bernie Sanders in the primary, suggesting that there simply was not the enthusiasm for Clinton’s candidacy as there was for Obama’s in 2012. Clinton’s favorable rating, for instance, was about 10 percentage points lower among the youngest black voters compared to the oldest black voters in the SurveyMonkey poll.
Perhaps most important is the group that voted in much larger numbers than in 2012: white voters without a college degree. (Trump won this bloc 63 percent to 32 percent.) Generally speaking, college graduates are more likely to vote than non-college graduates, even when controlling for race. According to the Current Population Survey, whites without a college degree made up 44 percent of voters who cast a ballot in 2012, and 58 percent of registered voters who didn’t vote.
These may have been some of the “missing” white voters that RealClearPolitics Sean Trende has written about, but in 2016, they weren’t missing. In the SurveyMonkey data, white non-college graduates made up 48 percent of 2016 registered voters who didn’t vote, substantially lower than 2012. They made up 47 percent of voters. It’s pretty remarkable that a group of voters that is shrinking as a percentage of the population made up a larger share of the electorate in 2016 than in 2012. But Trump made a clear appeal to this group, and some voters who stayed at home in previous years may have felt they had a greater voice in 2016.
Simply put, Trump got more of his voters to turn out than Clinton did. That’s quite a turnaround from the pre-election conventional wisdom that the Clinton campaign had the better turnout machine. Of course, Clinton’s turnout operation may well have nudged many reluctant voters to the polls, but either way, it wasn’t enough. The polling numbers from SurveyMonkey indicate that Clinton was hurt dearly by the voters who decided not to vote.