Pity the poor pollsters?
I doubt that sentiment exists anywhere in the U.S. after the 2020 polling was again, like in 2016, way off the mark nearly everywhere. Sure, the final presidential vote tallies, which at this hour remain in flux in a few key battleground states, do not indicate an unexpected winner of the White House this time.
But, after all the post-2016 rhetoric about refined polling procedures that would pay off, Tuesday’s results still show state-by-state pro-Biden predictions in the presidential race that were pure fiction, plus Senate and House projections that were absolutely wrong.
In Michigan, the polls showed Joe Biden leading by eight or nine points heading into Election Day. Instead, the former vice president apparently won by a sliver of the vote over the President Trump.
Incumbent Democratic Sen. Gary Peters supposedly pulled away in the final days of the campaign season to establish a several-point lead over Republican challenger John James. But the pollsters were embarrassed as Peters trailed throughout election night and did not secure a narrow victory until Wednesday, when the absentee votes from Detroit and Wayne County kicked in.
The prognostications by nearly every pollster that called for a continuing shift toward Biden and the Democrats in the Detroit suburbs were mostly off the mark. In Oakland County, the well-publicized pro-Democrat trend that began in 2016 never blossomed into a 2020 avalanche.
In Macomb County, Trump not only won again, after providing the winning statewide margin for the president in 2016 over Hillary Clinton, but his stunning margin of 12 points four years ago only shrunk to eight points on Tuesday. Meanwhile, James bested Peters in Macomb by 5 ½ points. In the state House, prophecies of Democratic gains never materialized.
Suburban Dem wave never materialized
Without naming names, as it appears Michigan has no pollsters that deserve accolades, their abilities to project trends and results among subsets of the electorate – suburban voters, the college-educated, blacks and Hispanics – seem to be lacking, at the least.
Michigan poll results released in the last week or two before Tuesday showed double-digit leads by Biden in the Detroit suburbs and, in one case, a survey found Biden cruising to a 15-point win in Macomb.
Clearly, small sub-samples have a much wider margin of error than statewide numbers, but when a pollster comments on his overall numbers and offers examples of a Biden surge due to suburban women or Macomb County’s bellwether voters, they need to be engaged in data-driven analysis, not pure speculation.
Nationwide, the 2020 polling averages painted a picture of Biden leading by nearly double-digit levels. The Democratic nominee was inaccurately favored by significant margins in battleground states ranging from Nevada to North Carolina.
In Wisconsin, where pollsters last week determined that the presidential race was essentially over, Democratic Gov. Tony Evers reportedly took one look at a Washington Post poll showing a Biden lead of 17 points and tossed it in the garbage. The final margin in the Badger State was 49.6-48.9 percent.
Elsewhere, pundits producing poll-driven stories proclaiming that the Democrats had a shot at carrying Florida, Texas and Iowa now are red-faced.
None of that panned out. There was no Blue Wave across the nation. The former vice president did not alter the political landscape or put an end to Trumpism. The predictions of a Biden landslide among Democratic activists now seems nothing less than silly.
Polling predictions for Congress way off
If you take a long look at the races for the House and Senate, the polling forecasts of significant Democratic gains were wayward at best.
In the Senate races, Democrats look to chart only a two-seat net gain, with nationally publicized “likely winners” that would flip the Senate flopping at the polling place. Chief among them are South Carolina Democrat Jaime Harrison, who spent a fortune on his election bid and was expected to eke out a win over GOP incumbent Lindsey Graham, who appeared to be on the ropes. Harrison lost by 10 points.
Similarly, Republican Sen. Susan Collins of Maine was viewed as a dead woman walking over the past two years, an easy pickup for the Dems. Contrary to the polls, Collins won re-election by a comfortable margin. In addition, Senate election results in Iowa, Montana, North Carolina and Georgia didn’t quite match the pre-election polls, which led to a distinct GOP advantage.
Even in Arizona, where Democrat Mark Kelly was posting double-digit leads in some polls a month out from the election, he is now locked in a tight, 4-point contest against beleaguered GOP Sen. Martha McSally.
It’s interesting to note that Trump’s internal pollsters probably viewed this race a sure loser for Republicans as much as any other polling firms. At a now infamous Oct. 30 Trump campaign rally in Goodyear, Ariz., the president announced to the crowd that he would give McSally one minute on stage if she hurried to the microphone. Trump glibly said that the supporters on hand “don’t want to hear” from her.
As for the House, conventional wisdom, backed by polling numbers, indicated a 10- to 15-seat gain for the Democrats. Instead, the GOP scored some House gains and the partisan margin in the lower chamber may shrink to just six or eight seats comprising the new Democratic majority.
Regardless of the final dramatic acts that play out over the next few days in the slow-moving counts that will determine the presidency, the Senate and the House, it seems fairly obvious that the failures of a polling system that doesn’t comprehend the highly polarized era ushered in by Trump, as well as cultural issues such as the overwhelming emphasis on cell phones, has failed us in 2016 and again this year.
It’s hard to imagine how long until the pollsters regain their credibility. I don’t think four years will do it.