There's a N.H. presidential poll for 2020. Yes, already.

Sen. Bernie Sanders, I-Vt., speaks at the 2017 Convention of the California Nurses Association/National Nurses Organizing Committee in San Francisco, Friday, Sept. 22, 2017.
Sen. Bernie Sanders, I-Vt., speaks at the 2017 Convention of the California Nurses Association/National Nurses Organizing Committee in San Francisco, Friday, Sept. 22, 2017.(Jeff Chiu / AP)

Oh, my: There's a New Hampshire poll out. As in the 2020 New Hampshire primary.

Yes, it's way too early for the poll to mean much. Indeed, in previous cycles I would have just said, "Ignore those polls!" and moved on.


But this time around, I have to point out that I was dead wrong about the Republican nomination in 2016, perhaps in part because I dismissed early — albeit not this early — polls showing Donald Trump on top. Now, I don't think that Trump's flukish nomination demonstrated that everything political scientists believed about presidential nominations was wrong.

But I do think a little more caution is in order.

Generally, the problem with early polling is that voters — most of whom won't really engage in nomination politics until very close to the election — simply don't have strongly held opinions about most politicians. So early polling tends to be little more than an exercise in name recognition. That's no doubt why on the Democratic side Sen. Bernie Sanders and former Vice President Joe Biden are the leaders in this University of New Hampshire survey.

Does that tell us anything useful? Well, sure.

For one thing, while it's true that advantages in name recognition will dissipate by Election Day as the other candidates run their campaigns, it's also true that initial name recognition is, in fact, a resource some candidates have and some don't. From the point of view of the candidates, it's a resource like any other — endorsements, campaign money, electioneering skills, and on and on.

Again, it's hardly the most valuable, but it is something.

And from the point of view of party actors, name recognition is one of the many things they may take into account when assessing the candidates. In my view, it's one that should receive zero weight: Whatever it's worth in the primaries and caucuses, it's certain that after the convention any nominee will be equally well-known. But the issue isn't what I'd do; it's what actual real-world party actors value, and it's certainly possible some of them do believe that a candidate already known to the public before the campaign has some advantages.

In other words, party actors can value anything they want when choosing which candidate to support, and analysis that ignores what they value if it appears to be foolish would not be very good analysis.


Speaking of what party actors value: It's hardly good news for Sanders that he reached only 31 percent in the new poll, after winning 60 percent of the Granite State vote in the 2016 primary.

Sure, some of that had to do with the two-candidate field then compared with the 12 potential candidates tested here. But that's just another way of saying that we shouldn't think of Sanders as commanding solid support anywhere close to his 2016 primary and caucus results.

And polls such as this one tend to convince party actors that the strong vote Sanders received during the 2016 process was more a function of the small candidate field rather than an indication of a surge of solid Bernie people.


Jonathan Bernstein is a Bloomberg View columnist. He taught political science at the University of Texas at San Antonio and DePauw University and wrote A Plain Blog About Politics.