There are many myths about the 2016 election. On the left it is an article of faith that Donald
Trump could not have won without illicit coordination with the Russians. On the
right, there is the pervasive notion that because Trump beat the odds to win
the presidency, all polling is wrong and should be disregarded. Yesterday Nate
Silver of FiveThirtyEight managed to blow up both of these theories in a single
tweet.
The exchange began with a tweet
from Ben Collins of NBC News that asked, “What did the Kremlin's cutouts
know about targeting MI and WI? How did they know it? And is there data to show
they took action on it?”
Silver, the head of the FiveThirtyEight polling analysis
site, responded quickly with a tweet
that slammed the door on that particular aspect of Russian collusion. “The 538
model, which was based on publicly-available polling data, said the campaigns
should target WI and MI,” Silver wrote. “You didn't have to have any
proprietary info to know they were important states. You just had to look at
the data and not be huge dumbasses like the HRC campaign was.”
Silver followed up with a link to a FiveThirtyEight
article from February 2017, “Donald Trump Had A Superior Electoral College
Strategy.” The thrust of the article,
subtitled “How Hillary Clinton and the media missed the boat,” was that Hillary
made two key errors in the campaign. First, she focused on states where the
race was close rather than states that had the potential to tip the race. In particular,
the article points out that Clinton did not set foot in Wisconsin after the Democratic
primary. Second, she was overconfident and limited her focus to a narrow range
of states. Hillary’s main focus was on Florida, North Carolina, and Ohio.
Silver didn’t explicitly address the idea that because the
forecasts were wrong in 2016 that all polling is wrong, but it is implicit in
his statement that the Clinton campaign was made of “huge dumbasses” who ignored
polling data that showed that Hillary was in trouble. Many Republicans claim
that the media gave Hillary a 99 percent chance of becoming president on
election night, but FiveThirtyEight’s
forecast, which is still posted, gave Donald Trump a 29 percent chance of
winning. Under those circumstances, Trump was an underdog, but not
prohibitively.
With respect to the two states in question, FiveThirtyEight gave
Trump a 21 percent chance in Michigan
and a 17 percent chance in Wisconsin.
Many polls were within the margin of error in Michigan, but Wisconsin polling was
further off, showing a consistent albeit single-digit lead for Hillary.
As I
pointed out a few months ago, polls are snapshots rather than predictive.
One good technique for examining polls is to look for trends in the big
picture. The big picture of the polling average from 2016 is still available on
Real
Clear Politics in convenient graph form. If we look at the trend, we can
see Trump plunging in the polls about Oct. 10 then starting a slow rise on Oct.
20. There is a sharp increase between Oct. 28 and Nov. 2 that brought Trump to
within two points of Hillary, well within the margin of error of most polls. Going
into Election Day, the national polling average had the two candidates about
three points apart, a close race by any standard.
If we look at key
events that occurred in the campaign, we can see exactly what caused these
movements in the polling. Keeping in mind that polls are lagging indicators, we
see that Trump’s decline in early October followed the release of the Access
Hollywood tape on Oct. 7. The final presidential debate was on Oct. 19 and
Trump’s performance seems to have helped him in the polls, but not enough to
close the deal. The event that sealed the race in Trump’s favor occurred on
Oct. 28, the release of FBI Director James Comey’s memo to Congress that
detailed the discovery of thousands of emails that related to the investigation
of Hillary Clinton’s private email server. In a May
2017 article, Silver also made the case that Comey’s memo cost Hillary the
election.
There are lessons for both parties in Silver’s tweets and
articles. For the Democrats, candidates should not take the Rust Belt states
for granted. Traditional party loyalties may not be enough to carry a state,
especially in an election where everything seems to be going wrong for your
candidate. There is no substitute for getting into the field and making
appearances. Charisma, broad appeal outside the party,
and the stamina to campaign should be factors
in nominating a candidate.
For Republicans, the lesson is also that the Rust Belt states
should not be taken for granted. Michigan, Pennsylvania, and Wisconsin were all
decided by about one percent of the vote after being all but ignored by the
Democrat candidate. Just because they voted for Donald Trump once does not mean
that they will do so again. The Clinton campaign made mistakes that will
probably not be repeated by the next Democratic candidate. Even with these
mistakes, however, Trump still lost the popular vote and would most likely have
lost the electoral vote had it not been for James Comey. Donald Trump has the
stamina to campaign, but he lacks charisma
and popularity outside the GOP.
Despite claims from both sides, the 2020 election is far
from a sure thing for either party. The outcome will be determined by which
side better learns the lessons of the 2016 election and adapts their strategy
to a changing electoral environment.
Originally published
on The
Resurgent
No comments:
Post a Comment