In 1964 I cast my first presidential ballot for Arizona U.S. Sen. Barry Goldwater.
Since that election, 15-four-year cycles, I’ve been a registered Republican, a registered Democrat and a registered independent. I have lived in New York, California, Washington and Pennsylvania. At no time did I ever miss in-person voting which must, I assume, qualify me among pollsters as “a likely voter.”
Yet during the last six decades I have never received a telephone call from a pollster asking me for whom I planned to vote. Moreover, after I inquired, I learned that no family member, friend, neighbor or work colleague has been polled.
Who, then, is polled? Given my longstanding experience as a confirmed but never polled voter, I wonder what the nonstop fuss in print media and television is all about: “Harris is up two points in Wisconsin, but down two points in Michigan!” or “Trump is up four in North Carolina and gaining in Arizona.” Comparable stories not only have headlined but consumed most of the print ink or broadcast air with one talking head after another chattering predictable points that depend on their political leaning.
Since the 2016 and 2020 polls were dramatically off the mark, no one should put any credibility in the 2024 election predictions.
In 2016, Donald J. Trump’s victory shocked many Americans, especially pollsters who showed his opponent, Hillary Clinton, leading the race right up to Election Day. All data they were looking at seemed to predict her victory. Clinton’s campaign, confident she would win, had the champagne ready to pop. But Trump, who disdained data gathering, carried swing states Wisconsin, Michigan and Pennsylvania, which Democrats thought were in the bag.
After the ballots were counted, Trump had won 306 electoral votes, compared to Clinton’s 232, securing him the presidency. The pollsters offered weak excuses for their embarrassing failures including a farfetched claim that the results were skewed by whether a male or female picked up the phone.
The 2016 misfire was supposed to serve as a wake-up call for pollsters, but it did not. The 2020 election would be, according to the polling, an easy Joe Biden victory. But Biden won by only three points versus his projected margin of eight — another humbling for the touted polling industry.
Pollsters have spent the years since 2020 experimenting with ways to induce hard-to-reach voters to participate in surveys and testing statistical techniques to improve accuracy. But expert opinion is mixed on whether polling outcomes are due for a repeat of 2020, which a professional association of pollsters called the most inaccurate in 40 years. New developments, such as the shift of Black and Latino voters away from Democrats and toward Republicans and the increase of online surveys that use unproven sampling methods create additional potential for error.
Referring to 2024’s polling reliability, Stanford University political scientist Jon Krosnick said, “We are headed for more disaster.”
Joe Guzzardi is an Institute for Sound Public Policy analyst who has written about immigration for more than 30 years. Contact him at jguzzardi@pfirdc.org.