Rebutting A Right Wing Talking Point — Polls Are Largely Accurate
One of the common talking points on the right is that the polls are wrong in light of Trump’s surprise “victory” in the electoral college in 2016. Trump’s enablers like to repeat this talking point because Trump is an historically unpopular president and he trailed Biden and Warren in a recent Fox News poll by 10 percentage points. In that same poll, Bernie Sanders led Trump by 9 points.
The reality is that this tired GOP talking point is simply wrong. The final Real Clear Politics (RCP) average in 2016 predicted that Hillary Clinton would win the popular vote by 3.3%. Seven of the nine final polls projected a Clinton victory in the popular vote by anywhere from 2 and 4 percentage points. As it turned out, Clinton ultimately won the popular vote by 2.1%.
Similarly, the final RCP electoral college projections in 2016 were essentially accurate and forecast a very close election. Clinton had a safe lead in states that added up to 203 electoral votes and Trump led in states that added up to 164 electoral votes. Fourteen states with 171 electoral votes were considered to be too close to call. Florida, Michigan and Pennsylvania were all in that category. (Wisconsin was leaning to Clinton.) As it turned out, Trump drew to an inside straight in the electoral college and took just about all of the states that were in the too close to call category.
The reality about 2016 is that the polls were largely accurate. The race became very close after then FBI Director James Comey violated his agency’s rules and released his infamous letter on October 28, 2016. At that time, Clinton led in the popular vote by a 6 to 8 point margin. The Comey letter reduced that margin to two points and cost Clinton the crucial battleground states.
The people who got it wrong in 2016 were the pundits — they misread the polls. The NY Times gave Clinton a 90% chance of winning. The most cautious forecast came from Nate Silver. The 538 election guru was criticized in the run up to the 2016 election for giving Clinton a 70% chance of winning. Silver even wrote a piece that contended that Clinton’s battleground firewall wasn’t as strong as Obama’s. Apparently, the pundits (like most Americans) simply couldn’t believe that a buffoon like Trump could get elected.
As a result of the 2016 outcome, there was some skepticism about the polls in the run up to the subsequent 2018 mid-term elections. However, once again, the polls turned out to be largely accurate. The final RCP average predicted that the Democrats would win the House national popular vote by 7.3%. Eight out of the last eleven polls forecast a Democrat margin of between 7 and 9 percentage points. Once all of the 2018 votes were counted, it turned out that the Democrats had their best showing since 1974 and won by 8.4%.
The only real polling errors in modern times were in the 1948 and 1980 presidential elections. In 1948, polling was in its relative infancy and the final poll was taken about two weeks before the election. The final Gallup poll gave GOP nominee Tom Dewey a five point lead over Harry Truman. Obviously, no poll was taken that detected Truman’s surge in the last fourteen days of the campaign.
The conventional wisdom in November 1980 was that the election was too close to call. Apparently, the pollsters failed to detect the big movement toward Reagan in the last few days before the election. Carter’s pollsters didn’t discover this surge until the day before election day. Reagan won the popular vote by a 51% to 41% margin over Jimmy Carter.
In most election cycles, the polls have been accurate. I think we can safely assume that Trump’s re-election prospects are in serious jeopardy and that support for impeachment is surging. Nonetheless, we can take nothing for granted in 2020. Already, Trump has indicated that he will recruit foreign help for his 2020 campaign and smear any opponent. All of us need to dedicate ourselves to helping Democratic candidates up and down the ballot in 2020. Now let’s get it done!