IE 11 is not supported. For an optimal experience visit our site on another browser.

Beware of presidential polling two years before Election Day

Presidential polling two years before Election Day has a dubious track record when it comes to predicting results.
Confetti on the floor on the last day of the 2012 Republican National Convention.
Confetti on the floor on the last day of the 2012 Republican National Convention.

By most metrics, the 2020 presidential campaign hasn't begun in earnest. We know Donald Trump is seeking a second term -- he launched his re-election campaign last year, earlier than any president in history -- but the Democratic field, which is expected to be enormous, does not yet have any officially announced candidates.

That said, as a practical matter, the 2020 race began months ago, as likely Dem candidates started lining up support and scheduling events in states that hold early nominating contests. It was only a matter of time before polling began.

And that time is apparently upon us. CNN released the results of a national poll of Democratic voters late last week, and the Des Moines Register published a new poll of Iowa Dems over the weekend.

I'm not here to tell you the results are completely irrelevant, because they're likely to have some real-world impact, even if the results largely reflect name-recognition. If you're, say, a former vice president weighing whether to run, you might look to the results of independent polls to help guide your decision.

Likewise, if you're a sought-after Democratic campaign staffer or a major donor, and you'll soon have to make some decisions about your future plans, the polling may have a significant influence.

That said, about once every four years, I like to remind folks that presidential polling two years before Election Day has a dubious track record when it comes to predicting results.

Let's take a stroll down memory lane:

* At roughly this point in the 2000 cycle, George W. Bush was in a strong position, but in the race for the Republican nomination, Elizabeth Dole and Dan Quayle were both seen as top-tier contenders. By the time voting began in the nominating contests, neither Dole nor Quayle were in the race.

* At roughly this point in the 2004 cycle, John Kerry was in a strong position, but in the race for the Democratic nomination, Joe Lieberman led most national polls and Dick Gephardt was in the top tier. Tom Daschle was also seen as a strong contender. In reality, Daschle didn't run; Gephardt quit before New Hampshire; and Lieberman generally embarrassed himself.

* At roughly this point in the 2008 cycle, Hillary Clinton was the frontrunner among Democrats, but Al Gore and Bill Richardson were both polling in double digits. (Gore didn't run and Richardson fared very poorly.) In Iowa, in late 2006, John Edwards had stronger support than Clinton and Barack Obama combined, but it was Obama who ended up winning the Iowa caucuses.

Among Republicans in the 2008 cycle, Rudy Giuliani enjoyed big leads in nearly all national polling. He proceeded to win literally zero contests.

* At roughly this point in the 2012 cycle, Mitt Romney was fairly well positioned, but Mike Huckabee and Sarah Palin were both at or near the top in national Republican polls. Neither Huckabee nor Palin ended up running.

* At roughly this point in the 2016 cycle, Mitt Romney led in all national polling, while Donald Trump was so far from the GOP's radar, he wasn't even included in the surveys. Romney, of course, ended up not running, and I suspect you know what happened with Trump.

Maybe the latest 2020 polls offer some hints about what's to come? Sure, it's possible. But looking back over the last couple of decades, we've seen enough top-tier candidates fizzle -- or skip the race altogether -- to know that way-too-early polling should be taken with a grain of salt.