"Election results were consistent with polls"

December 15, 2020
Adirondack Daily Enterprise

Robert Schnibbe

In the final election results, Biden won 51.3% to 46.8%, by +4.5% with 7 million votes. The polling averages had Biden around +7% or +8% nationally the night before Election Day. With a historically normal sized polling error of 3% to 4%, you would expect Biden between +4% and +11%.

The turnout was a record high for both parties. If Trump had received only as many votes as he did in 2016, inverting their own bogus statistical argument, Biden would have won +12% — including Texas and the down-ballot toss-ups.

The RealClearPolitics national vote error underestimated Trump by -1.1% in 2016 and -2.7% in 2020, and Obama by -3.2% in 2012. Their average error for the same battleground states was 3.3% in both 2016 and 2020. With the exception of Florida, most states were roughly unchanged in the errors by poll aggregators, in both size and direction.

Forecast models like FiveThirtyEight and DecisionDeskHQ only gave Trump 10% odds of winning. It was not because they expected the polls were “right.” It was that if the polls were only as wrong in his favor as 2016, he would still lose the battleground states with narrow margins.

This is illustrated by subtracting the 2016 errors from the 2020 state polling averages. Biden wins a few states by less than 1%, and wins in the low 300s in the Electoral College. He won 306-232.

The Electoral College is the only reason Trump had a chance without a “Dewey defeats Truman”-sized upset. In spite of popular perception, the Electoral College does not consistently favor one party. Kerry almost won with 3 million fewer votes. But Republicans have lost the popular vote in seven out of the last eight elections. It was stacked toward Trump, but not by enough.

FiveThirtyEight still had Biden at 90% odds when Trump was falsely declaring victory at 2:30 a.m. on election night. It was clear by 9 a.m. Biden had won Wisconsin and Michigan, and with leads in Arizona and Nevada, was probably the president-elect with 270 electoral votes.

Unfortunately, what has happened since then is equally unsurprising, and very dangerous. These are mostly illegitimate election challenges. They are really “autogolpe” attempts, a coup by subverted institutions.

Without any actual evidence of widespread fraud, and now trying to overturn the election because the mere possibility of fraud makes the results “unknowable,” success would mean millions of votes can be ignored in every election with retroactive disenfranchisement by courts or state legislatures. Electoral chaos.

Imagine if Pennsylvania had counter-sued the Texas case, to which Rep. Elise Stefanik signed in amicus brief, instead requesting the Supreme Court invalidate the electors for Texas, Florida, Ohio, and North Carolina entirely.

Odds Trump won them are only one in a gazillion! The conspiracy was orchestrated by the Russian G.R.U., international bankers, and the lizard people of the Flat Earth Society. Hundreds of hearsay affidavits will be presented at the Waldorf Astoria Plumbing & Heating off I-86!

That’s where this road ends — with parties willing to end democracy, refusing to admit they lost.

Robert Schnibbe lives in Saranac Lake.

For fact checking purposes:

Table comparing state polling errors from FiveThirtyEight and RealClearPolitics:

Table comparing Trump 2016 vs. Biden 2020 vote totals in key states:

Forecast modeler projections and red/blue lean map estimates for Electoral College outcomes:





Poll averagers:

FiveThirtyEight 2020: https://projects.fivethirtyeight.com/polls/president-general/national/

RCP 2012: https://www.realclearpolitics.com/epolls/2012/president/us/general_election_romney_vs_obama-1171.html

RCP 2016: https://www.realclearpolitics.com/epolls/2016/president/us/general_election_trump_vs_clinton-5491.html

RCP 2020: https://www.realclearpolitics.com/epolls/2020/president/us/general_election_trump_vs_biden-6247.html

External Links

(1) "Election results were consistent with polls" Guest Commentary; Adirondack Daily Enterprise, December 15th 2020 (source)

(2) "Take polls seriously, not literally" Letter to the Editor; Adirondack Daily Enterprise, October 17th 2020 (source)