The history of polling in national elections has a troubled history. Presidential polling started back in 1936 with the Literary Digest, a national magazine of the time, which sent out to their 10 million patrons a “straw” ballot asking them who they planned on voting for in the 1936 presidential election. Some two and half million ballots returned, and based on their survey, they predicted Alf Landon would beat Franklin D. Roosevelt 57% to 43%. Of course, the actual vote was considerably different. The final vote was Roosevelt’s 62% to Landon’s 37%. It escaped them at the time that the average voter was generally poor, whereas the patron of Literary Digest was generally well-to-do. And this was in the midst of the Great Depression…

Yet, it happened again in the 1948 presidential election when Thomas Dewey, then governor of New York, ran against the incumbent, Harry Truman. Three major polls (Gallup included) predicted Dewey would win. The Gallup poll came closest with Dewey’s 49.5% versus 44.5% for Truman, which was almost exactly the opposite, as Truman beat Dewey 49.6% to 45.1% in votes and won 303 electoral votes to Dewey’s 189. This failure of the press is famously remembered with Harry Truman holding up the Chicago Tribune newspaper with the headline, “Dewey Defeats Truman.”

We would like to think that, over the years, polling has become a more accurate science. However, history has not been kind to pollsters. Even when they are right, such as Jimmy Carter over Gerald Ford in 1976 and George W. Bush over Al Gore in 2000, the pollsters predicted much larger victories. And in 2016, the pollsters considered Hillary Clinton a lock over Donald Trump. Big surprise!

But just like the Peanuts comics where Lucy promises Charlie Brown that this time, she will not pull the football away from him when he tries to kick it, the pollsters tell us that, positively, this time for sure — we KNOW who is going to win this election.

Yeah, right. There are over twenty different national polls, and they range in mid-September from a three-point advantage for Trump to a four-point advantage for Harris. So, help me out here. How can these astute, brilliant number crunchers search for the same thing (popularity of candidates) and come up with such different results?

This is because all polling agencies use statistical adjustment called “weighting” to make their survey sample align with the broader population’s key characteristics. Some use as few as three characteristics (e.g., age, race, gender), and others use as many as twelve characteristics (adding to age, race, gender, other dimensions such as political party, religion, education, etc.). This ‘weighing’ must be consistently modified because the society itself is constantly changing. And when the actual results come through after the election, as sure as the sun will rise in the east, pollsters will admit they ‘missed’ a sizable chunk of the voting citizenry, which is what scuttled their predictions.

There are great benefits for these organizations to win the “our poll is closer to the actual vote than your poll” contest. But, they also know that before the actual vote, they can get valuable headlines by being an ‘outlier,’ showing one candidate or the other moving the needle. The alphabet media, ABC, CBS, NBC, NPR, PBS, etc., all know they can get attention (and thus raise THEIR market numbers) by highlighting dramatic movements in the ongoing contest.

As November 5th gets closer, more and more attention will be given to the pollsters. But, here is the bellwether of polling that really matters. Ask Al Gore how he felt about getting the majority of votes in 2000 or how Hillary Clinton experienced the same in 2016. John Quincy Adams, in 1824, could brag that he had more votes than Andrew Jackson. In 1888, Benjamin Harrison could high-five himself by receiving more votes than Grover Cleveland. But, as pleased as these presidential candidates were about receiving the overall majority of votes, that does not win a presidency.

At the end of election day, the electoral college vote determines who wins the election. I know many hate it, but the concept of the Electoral College is a cornerstone of our republic. The top ten states in population in the USA have more votes than the other 40 states and Washington DC combined. As Rush Limbaugh said so well, “State sovereignty is key here in the Electoral College — and if you’re going to start divvying up the power of each state’s elections, you are destroying state sovereignty.”