Earlier this week, Maine made headlines as the state’s legislature passed a bill introducing a substantial electoral reform to the state’s presidential primaries and elections. The bill institutes ranked-choice voting, a system that allows voters to rank their preference of candidates for a position and takes all of these preferences into account in the final result. In short, ranked-choice voting is likely to lead to elections that better reflect voters’ preferences, allow for more diversity of candidates and reduce vote-splitting. Recognizing this, Democratic parties in six other states have plans to implement measures similar to Maine’s in their upcoming presidential primaries.
Also, this month, the field of possible Democratic presidential nominees was effectively cut in half. Out of the 20 candidates present on the stage of the second debate, only 10 met the requirements for polling and donations set by the Democratic National Committee to qualify for the third. With the Iowa caucus still far off, similar polling cut-offs are likely to be used by the DNC to further narrow the field in the future.
Watching these two events occur in tandem, it seems strange that the moves being made by Maine Democrats to reform actual elections haven’t been reflected in these debate qualifiers. In the weeks since the second debate, the polls used in the DNC’s cut-offs have served as a sort of pre-election, defining the candidates that voters will and will not get to choose in the future. And while the 10 candidates eliminated in this round were all unlikely to win, these polling cut-offs will become much higher stakes as the field narrows. This is especially true for the “second tier” of candidates — those who are not frontrunners but have a shot at becoming one in the future. At this point, that’s essentially everyone other than former Vice President Joe Biden, Sen. Bernie Sanders, I-Vt., and Sen. Warren, D-Mass.
These candidates may be victims of the same problem that Democrats in Maine and other states are trying to solve with ranked-choice voting. In one of the polls used by the DNC to determine the candidates for the third debate, CNN asked respondents to indicate both their first choice for a nominee and the candidate that they are next most interested in hearing from. The results show an enormous discrepancy between voters’ first and second choices, revealing some candidates to be significantly more popular than they initially seem. For example, in April polling, Sen. Kamala Harris, D-Calif., garnering only 5 percent of first-choice votes, beat out the three frontrunners by being listed as 23 percent of voters’ less preferred choice. In fact, almost all candidates other than the frontrunners appear substantially more popular once second choices are taken into account, while Biden and Sanders seem to perform worse.
Essentially, this indicates that while most voters prefer Biden, Sanders or Warren as their top choice, many of them wouldn’t choose one of the other three frontrunners as their second choice. Another one of the qualifying polls confirmed this: Morning Consult explicitly asked supporters of the top candidates to indicate their second-choice preference. About half of supporters of the top three candidates did not list one of the other three frontrunners as their second choice.
This means that the results of subsequent polling cutoffs may be significantly different between traditional and ranked-choice voting mechanisms. While I wish I could tell you how the results would be different, the problem of ignoring voters’ second-choices is too entrenched to allow that. None of the polls used by the DNC published detailed results of voters’ second choices, and most of the ones I looked at didn’t ask about second choices at all.
Ultimately, this problem extends beyond the cut-offs for the primary debates and into the way we cover and discuss elections. Polls aren’t just hurdles for candidates to jump over on their way to the debate stage. They are also one of the primary ways voters, politicians and commentators try to understand what is going on in elections. The results of polls can go viral, thrusting candidates into the spotlight and giving their prospects a boost, as happened when a single poll earlier this month contradicted all others by finding Warren tied with Sanders and Biden among Democratic voters. Furthermore, the campaigns themselves are influenced by polling data in developing their messaging and policy.
If polls ignore voters’ second choices, they can make some candidates, like Harris and Mayor Pete Buttigieg of South Bend, Ind., seem substantially less popular than they actually are (and others, like Sanders, more so). If polls only incentivize candidates to make attention-grabbing, radical statements to accumulate more first-choice followers, it may lead them to ignore building a broadly supported platform that will actually get people out to vote in the general election.
At a time when the Democratic party is increasingly fractured, and winning the general election has never seemed more important, these could be costly mistakes. The good news is that substantive change need not wait on the lengthy legislative processes faced by states like Maine attempting to institute ranked-choice policies in actual elections. The DNC could immediately require qualifying polls to ask more questions about voters’ second and third choices. Even if subsequent debate cut-offs don’t use ranked-choice mechanisms, altering the polls alone would instantly deepen our understanding of the election and the public conversation about it. Doing so now, before the field narrows further, could shape the outcome of both the primaries and the general election.
Jared Stolove can be reached at jstolove@umich.edu.