“Polling is an inexact science,” said statistician Andrew Gelman of Columbia University.

Duh! Yeh! We have found that out now in two consecutive presidential elections. I would also add that it is not a “science,” even though the pollsters use scientific modeling. So what’s to be done? Stop polling? Well you know that will never happen, not when there is money to be made and not when the public loves to hate polls.

In 2016, the result of a number of polls led the experts to believe that there was a 90% chance that Democrat Hillary Clinton would win the election. She didn’t. Real estate entrepreneur and media star Donald Trump won a shocking victory to become the 45th president of the United States.

In the four years since Trump has been president, several of his close insiders have said that based on the polls in 2015 and 2016, even he didn’t think he would win. Since then, he has trashed polls that showed him behind and tweeted unflinching support for polls which found him doing well.

In 2020, the chastened pollsters said they did their homework and figured out where they went wrong four years earlier. The majority predicted a fairly easy win by Democrat Joe Biden over Trump.

They also said down-ballot candidates would ride Biden’s coattails, allowing the Democrats to win control of the U.S. Senate and pick up nearly a dozen seats in the House of Representatives.

It was not an easy Biden victory as the polls had been predicting. In fact, Trump has refused to concede the election two weeks after the balloting. Trump’s legal team has filed multiple lawsuits in key states such as Pennsylvania, Wisconsin, Georgia and Nevada contesting the fairness of the election and charging fraud, although there has been no proof shown.

Biden leads Trump in the Electoral College vote count, with all states called, 306-232. (The 306 electoral votes is the exact number Trump had in victory in 2016.) Biden also had more than 5 million more popular votes than Trump, a number expected to swell before the final figures are certified.

The polls were wrong again in many states. Most glaring was Florida, where the polls predicted a Biden victory with a 3%-4% margin, whereas Trump took the state by that amount.

In Pennsylvania, the state that put Biden over the top on Nov. 7, four days after the general election, pollsters Terry Madonna and Chris Borick had Biden winning by six and five points, respectively, but he will probably wind up winning by about 2%.

In Wisconsin, where all of the votes have been counted but not certified, Biden is leading by about 20,000 votes, or a little more than half of 1%; the polls had Biden winning by 7%. In Michigan, Biden is leading by 2.9% while the polls predicted he would carry the state by 7%.

There was anything but a Biden coattails effect. He appears to have won the presidency with the weakest House of Representatives coattails of any president since John F. Kennedy in 1960. Studying the figures, it was obvious that many of those who voted for Biden split their tickets to vote for Republican down-ballot candidates.

The snapshot of the situation right now is that Democrats will pick up one U.S. Senate seat, which means the Republicans have a 50-48 lead with two runoffs in Georgia still to be decided in early January. If Republicans win one or both, they will retain control of the Senate. If Democrats win both, it would mean a 50-50 split, with Vice President-elect Kamala Harris as Senate president casting the deciding vote in any tie.

In the House of Representatives, the Republicans are poised to pick up as many as 11 seats, leaving the Democrats with a much slimmer margin.

Based on their strong showing in 2018 when they flipped five Senate seats and added 11 House seats, Pennsylvania Democrats had high hopes for building on those successes. There were even some pie-in-the-sky expectations that maybe they could take control of both houses of the General Assembly. Those pipe dreams evaporated quickly, and with them any hope that maybe this time Democrats could lead the effort in redrawing the legislative redistricting map that must be done after the U.S. Census figures are certified.

Republicans won the open House seat created by Democrat Neal Goodman’s retirement in Schuylkill County. Republicans filled two House seats in Lehigh and Northampton counties formerly occupied by retiring Republicans Justin Simmons and Marcia Hahn.

At the congressional level among Pennsylvania’s House members, it ended up just as it began: nine Republicans and nine Democrats. After the census figures are certified, however, it is expected that Pennsylvania will lose one seat in Congress, so this could lead to a fight between two incumbents in 2022 to see who prevails.

First-term Democratic U.S. Rep. Susan Wild retained her seat over Republican Lisa Scheller by 3.8%, nowhere near the 13-point margin that the polls predicted.

Pollsters have begun soul-searching again to try to come up with answers for their second questionable showing in a row. Some say it was the high turnout; others say it was “shy” Trump voters, meaning that some surveyed intentionally said they would vote for Biden when they wound up voting for Trump. There also were many other postelection excuses, but I am wondering whether we just should ignore the polls as being too inexact and untrustworthy.

By Bruce Frassinelli | tneditor@tnonline.com