We ran an experiment to see how their support would shake out if we got rid of everyone else still in the race.
The three have roughly the same overall support, but Sanders and Warren would both surge while Biden has the least to gain. Biden is the one who currently benefits the most from such a large field.
It's clear, though, that no one has a clear path to victory without taking down the others.
18% of Democratic voters wouldn't be happy with any of the them, which shows there's still room for another candidate to pull ahead.
Over the past several weeks, three candidates vying for the 2020 Democratic presidential nomination have distinguished themselves as frontrunners: Former Vice President Joe Biden, Sen. Elizabeth Warren, and Sen. Bernie Sanders.
Yet, at the close of the September primary debates, there are still a total of 18 candidates in the race, most of whom have virtually no shot at becoming president. Biden is polling at an average of roughly 30%, Warren at 19%, and Sanders at 18%, according to an aggregation from RealClearPolitics. The next-closest candidate, Kamala Harris, is stuck at 7%.
The hanger-on candidates are basically real-life statistical noise, making it harder to discern who's actually likely to win the primary race by detracting polling points from their opponents in the lead.
So: What if we just got rid of them? How would things shake out between Biden, Bernie, and Warren?
Thanks to the unique way Insider conducts its polls, we can test exactly that. Unlike most polling organizations, Insider/Surveymonkey's poll asks would-be voters which candidates they would be happy with if they were to become president, rather than naming just one candidate. Using that data, we ran a mathematical model that analyzed what it would look like if we eliminated every candidate except for the top three and re-allocated those voters to Biden, Sanders, and Warren.
If we Thanos-snapped Harris, Buttigieg, Booker, O'Rourke, Yang, Klobuchar, Castro, Bennet, Bullock, Delaney, Gabbard, Ryan, Williamson, Steyer, and De Blasio out of existence, here's what the field would look like, according to the polls we've conducted this summer:
For Biden, we'd have a situation where 30% of his fans would like him and him alone, 37.4% percent deciding between him and Sanders or him and Warren, and a third of his supporters fine with all three.
That's compared to a current situation where 16.1% of those satisfied with him are satisfied with him alone, 15% have him as one of two choices, and 18% have him as one of three candidates they're satisfied with.
For Sanders, we'd have 24% of his backers liking him alone, 42% supporting either Warren or Biden, and a third liking all three.
Right now, 15% of his supporters like him alone, 18% of them like him and just one other person, and 19% like him and two other people, with the remaining supporters deciding between four or more. Those numbers are better than Biden's.
For Warren, 15% of her fans would approve of her and her alone, which is three times her current level but still half of Biden's lockdown rate. 51% of them would be satisfied with her and one other person, and a third would approve of all three.
As of now, only 5% like her and her alone, 12% like her and one other person, and 19% like her and two others.
It's also important to note that some things don't change. As of now, overall, Warren is known and satisfactory for 45% of Democrats, about 46% of respondents know of Sanders and would be satisfied with him, and 47% would be fine with Biden. Those numbers are the same whether you've got three contenders or 19.
In that sense, the three of them are in a near-tie. But the differences from what the race looks like now compared to what it would be like if it were just the three of them are telling.
Warren and Sanders would catch up to Biden
Biden is the clear beneficiary from such a large group. While he stands to gain some voters from other people dropping out, our polling shows that his supporters are remarkably loyal. Compared to other candidates, few voters would be satisfied by having anyone other than him as the nominee. His numbers roughly double, but in our Thanos'd universe the increase isn't enormous.
Sanders and Warren, though, both surge. Most of the remaining voters would be happy with either of them as the Democratic nominee. The two of them seem to have the biggest opportunity to pull supporters in and grow their base.
The sharp increase in Warren's numbers, in particular, is revealing.
One reason she's perceived to be the beneficiary of a lot of dropouts is because a lot of voters like Warren and someone else. The thinking goes, that makes her the main beneficiary of a lot of other people dropping out of the race.
While this is true, that alone won't put her in front of Biden and Sanders. Each of the top three needs to eliminate at least one of the others if they want a clear shot at winning. The only way to win is to slay a giant.
This whole experiment entirely unfair, of course. While we can be certain that DeBlasio or Delaney won't be president, Harris, Booker, Yang, and Buttigieg are still hanging in there. Even if they're struggling to pull in two digits, it's still early, and history shows that candidates from the back of the pack can make it to the finish line.
And as it stands, 6% of people who said they were Democratic primary voters who are not satisfied with anyone in the field. In our Thanocracy, where we reduce the options to three, that figure triples to 18%. There's still room for the lower-polling candidates to persuade voters.
But if Warren and Sanders want to catch up to Biden, perhaps their strategy shouldn't be to try to take him down in debates. Instead, it may be a better move for either of them to sweep away all the hanger-ons who seem less likely every day to make it to the White House. If they're itching to give Rep. Tim Ryan a flogging, it can only help them.
And if Biden wants to coast to victory, he should be in no hurry to pick off his lesser opponents who don't pose a clear threat to him anyway. Chances are, if they drop out of the race, their supporters would go to another candidate anyway.
SurveyMonkey Audience polls from a national sample balanced by census data of age and gender. Respondents are incentivized to complete surveys through charitable contributions. Generally speaking, digital polling tends to skew toward people with access to the internet. Total 1,090 respondents collected September 6 to September 7, 2019, a margin of error plus or minus 3.07 percentage points with a 95% confidence level.