Donald
Trump was at the top of each of the last 10 polls in Iowa, but his lead
failed to hold up on caucus night Monday. In the end, his seven-point
lead in polling averages amounted to a three-point loss to Ted Cruz.
That
10-point swing was enough to make Mr. Trump’s defeat the biggest
polling error in an early primary since Hillary Clinton defeated Barack
Obama in New Hampshire in 2008. But even that measure understates the
extent that the polls misjudged Mr. Trump’s strength.
Mr.
Trump was at 31 percent in the final polls, but finished with just 24
percent. In our data set of early primary polls from New Hampshire and
Iowa since 2004, no candidate underperformed the final surveys by as
much as Mr. Trump. Mrs. Clinton, for instance, mainly beat Mr. Obama by
outperforming her polling, not because Mr. Obama fell short.
It’s
probably not a coincidence that the candidate who underperformed the
polls by the most is also the one who had a mediocre turnout operation
and enjoyed seemingly nonstop media coverage.
It’s
always hard to figure out why polls are wrong, but this time the stakes
are higher. Republican strategists have hoped for months that Mr.
Trump’s lead was an illusion. The results in Iowa at least raise the
possibility that they’re right — which would call into question Mr.
Trump’s advantage elsewhere.
This
time there is evidence to support one of two possibilities for why
polls overestimated Mr. Trump: Voters broke strongly against Mr. Trump
in the final days or the electorate was more conservative and more
religious than polls anticipated.
In general, there are three basic ways polls go wrong:
■ an unrepresentative sample that doesn’t accurately reflect the population it’s trying to measure.
■ a flawed likely-voter model that misjudges the composition of the electorate.
■ late events or changes in the race after the poll was conducted that would move voters.
This
year, there is extremely strong evidence to support the “late movement”
scenario, some evidence to support the likely-voter problem and little
evidence to support the sampling problem — even if it can’t be ruled
out.
The
case for late movement away from Mr. Trump is very strong. The entrance
polls showed that voters who decided over the last week broke heavily
against him, and for Mr. Rubio.
What’s
more, the only two surveys conducted after Jan. 29 provided a certain
amount of confirmatory evidence. They aren’t the best polls in the
world, but they showed Mr. Trump leading by just one percentage point,
with 27 and 20 percent of the vote — his two lowest tallies in more than
a week. The same surveys showed Mr. Rubio rising to 19 and 22 percent —
his two strongest showings of the campaign.
If
late movement explains Mr. Trump’s defeat, it’s good news for
pollsters. There’s nothing they can do about the possibility that
undecided voters will break away from a candidate at the last minute.
It
doesn’t necessarily answer whether Mr. Trump’s lead should be expected
to hold elsewhere. On the one hand, you can make a solid case that the
final weeks of the race should naturally pose a very difficult challenge
for him. He has benefited from tremendous media coverage and has faced
relatively few attacks. This changed in the final few weeks before Iowa,
as voters started to tune in and focus on other candidates, and as Mr.
Cruz began attacking Mr. Trump far more forcefully than he had been
attacked before.
This
would be the worst-case possibility for Mr. Trump. It would mean his
support really might evaporate ahead of future contests, as voters focus
on other candidates and as he faces even more attacks.
But
it’s also possible that voters broke late because of Mr. Trump’s
strange decision not to participate in the final debate. The entrance
polls did not ask voters whether they were less likely to vote for him
because of that decision, but it’s certainly possible.
This
would be the best-case explanation for Mr. Trump: If skipping the last
debate did it, then he can simply avoid making the same mistake again.
On balance, though, there’s not a lot of evidence to support this
conclusion — and entrance polls showed Mr. Trump losing among voters who
decided over the last month, not just over the last few days.
The Electorate
There
is also a case that Mr. Cruz benefited from a more evangelical
Christian electorate than pre-election polls assumed. The entrance polls
showed that evangelicals were 63 percent of the electorate, while most
pre-election polls showed a much lower tally. The final Selzer poll, considered to be one of the most reliable polls, had evangelical voters at 47 percent of the electorate.
Mr. Trump lost evangelical Christians to Mr. Cruz by a 12-point margin, 33 to 21 percent.
Mr.
Cruz would not be the first conservative candidate to outperform Iowa
polls. In fact, it’s pretty common. Pre-election polls showed a close
race between Mike Huckabee and Mitt Romney; Mr. Huckabee won by nine
points. They showed Rick Santorum trailing Mitt Romney significantly; he
won, although one could attribute that to late momentum, not a poor
measure of the electorate. Pat Robertson outperformed the polls in 1988.
The list of moderate candidates who underperformed is just as long, and Mr. Trump did best among self-identified moderates.
One
easy explanation is that the most conservative voters are among the
party’s most committed, longtime activists who regularly attend
caucuses, and that polls conducted by mainstream media organizations do a
poor job of incorporating them into their likely-voter models. That’s
because they don’t buy the list of past caucus attendees, which is
collected by the state parties and sold at a considerable cost. When
asked during a phone survey, voters often misreport how much they’ve
participated in the past.
But
this is probably not the whole explanation. Mr. Trump, after all, only
won 29 percent of the nonevangelical vote — even less than the 31
percent he held overall in pre-election polls. If that’s right, Mr.
Trump would have underperformed the pre-election polls, even with a
100-percent nonevangelical electorate.
What’s
more, the proportion of voters who were “very conservative” was
actually fairly low, at 40 percent of voters — well beneath the 47
percent from 2012. Mr. Cruz won “very conservative” voters by a far
greater margin than he did among evangelicals.
And
the turnout was, as Mr. Trump might put it, huge. Nearly 187,000 people
turned out, or more than a 50 percent increase over 2012 levels.
Pre-election polls all showed that Mr. Trump would benefit from a
stupendous turnout, but few anticipated a turnout of anything near this
level. Even Steve King, the Iowa congressman and Ted Cruz backer, said
turnout needed to be around 135,000 people or lower for Mr. Cruz to win.
Given the huge turnout, Mr. Cruz might not have even targeted as many
voters as he actually won, making it harder to argue that it was his
vaunted field operation that put him over the top.
So
it is a stretch to argue that Mr. Trump lost simply because of the
turnout, given that it vastly exceeded just about every pre-election
benchmark for a Trump victory.
In
the end, we rarely know exactly why polls are wrong. None of the public
pollsters will be able to go back and see whether the people they
thought would vote actually did. They could call voters back and find
out whether they voted and for whom, but voters tend to over-report
voting and over-report voting for the candidate who won, so it’s not
clear whether there’s much to learn.
What’s
clear, though, is that Mr. Trump woefully underperformed — even with
the benefit of a big turnout. This type of a miss usually can’t be
easily explained by one single factor. Multiple causes were probably at
play. Some may be specific to Iowa. Others might not be, and may
represent a lasting problem for Mr. Trump.
It could be exactly what his opponents have been hoping: a sign that he’s not as strong as he and the polls have been saying.
No comments:
Post a Comment