Why parties shouldn’t use polls to vet candidates
Polling is not suited for determining democratic outcomes
The takeaway: With all their uncertainty, can we really be confident that polls are measuring the “pulse of democracy”? Maybe, but they are not up to the task of determining who gets to run for president.
Editor’s note:
Thanks for subscribing! Your membership adds up and makes all this newslettering possible (reminder: I do all this work independently). Please consider sharing online or with a friend; the more readers, the merrier. Remember that apart from getting special articles subscribers can also comment below each post and participate in exclusive threads.
As always, send me your tips about what you’d like to read about next. Or what you don’t want to read. Or your feedback otherwise. Also, cat pictures are nice, so please send them to me! I’m elliott@thecrosstab.com, or @gelliottmorris on Twitter.
Thanks all,
—Elliott
I don’t think parties should use polls to determine who gets on the debate stage. There, I said it.
This idea is a bit controversial; on the one hand, polls allow us to incorporate the “pulse of democracy” into the political process. But on the other hand, it remains a question whether parties are using polls to measure the public’s attitudes or determine them.
Jill Lepore had a lot to say about polls and primaries in an article for The New Yorker in 2015. I encourage you to read it. But here’s where I come down on the subject: there are just too many questions with polls to use them, effectively, to determine who gets to run for president. Here are those questions, and some answers that I think tell us “no, we shouldn’t use polls to determine who gets to run for president”.
Are polls really precise enough? The predictive margin of error for a poll is actually about double what a pollster reports. If a party is using a threshold of 4% to determine which candidates to invite to an upcoming debate, there’s a good chance they could be including or excluding candidates who fall above or below the threshold simply by chance. Requiring candidates to meet the threshold in multiply polls helps decrease the chance that chance determines outcomes, but should it still have any role in the process at all?
Should we use polls that ask the horse race of all adults, registered voters, or likely voters? And if the latter, who is a “likely voter”? Parties should listen to all of their voters. Or should they listen to all of their members? Using polls requires parties to choose between the two groups. If parties choose to prefer voters over all their members, they then are relying on how pollsters define “likely voters”. Those categorizations often contain assumptions about voter behavior that don’t always pan out. Assumptions have no role in the democratic process.
Which polls do we include or exclude? Why? What if the polls we exclude are “better” than the ones we don’t? The DNC has decided that some polls count for meeting thresholds while others do not. But different candidates qualify depending on which polls we include. This is another way that arbitrary decisions influence the process.
When polls determine candidacy, what are we valuing in our candidates? The theory behind using polls to determine candidacy is appealing. Candidates with support from “the people” deserve to be included in the process, and those not backed by public opinion don’t. But are polls a quality barometer for public opinion? Only 5% of Democratic voters are even considering voting for Tom Steyer, for example, but he bought enough ad time in key states to rank highly in certain polls and qualify for the debates. Polls might not enable parties to value popularity, but instead place a premium on fundraising ability and social media popularity.
Some people might argue that polls are an imperfect way to determine who gets on the debate stage, but that they’re still the best worst option. I submit that imperfect tools have no place in determining the leader of our country.
Why parties shouldn’t use polls to vet candidates
Good topic. I share Craig Stokes question. Also, which "polls?" I'm interested in what content is being used -- horserace data are unfair and misleading early on. If, however, someone was exposing a representative sample to the promotions and objections that the candidates were making about themselves and each other, I'd be interested in using those responses to filter.
So, if polls should not be the screening mechanism, what measure, if any, should the DNC use?