How the INSIDER 2020 Democratic primary tracker works

Advertisement
How the INSIDER 2020 Democratic primary tracker works

2020 Democratic presidential candidates are seen in a combination of file photos (L-R top row): U.S. Senators Kirsten Gillibrand, Amy Klobuchar, Elizabeth Warren, Bernie Sanders, (L-R bottom row): U.S.Senator Kamala Harris, Former Texas congressman Beto O'Rourke, U.S. Senator Cory Booker and Democratic presidential candidate Andrew Yang. REUTERS/Files

Reuters

Combination photo of 2020 Democratic presidential candidates

Advertisement
  • We're tracking the Democratic primary contenders and want to find out their overlaps in support.
  • This helps us figure out exactly who most stands to gain when candidates falter, and who is most directly competing for support.
  • It's early in the race and votes won't be cast for months, so this allows us to get inside the changing minds of voters.

INSIDER is interested in Democratic voters' frequently overlapping perceptions of different candidates jockeying for the presidency.

There are plenty of places that are actively polling the Democratic primary, and for people interested in the topline numbers - the percentage of people who say a given candidate is their first choice at a given time - those numbers are better for their needs. But we're interested in a different approach. We want to know how satisfied a voter would be in the event that each particular contender would be the nominee.

Moreover, we choose not to focus on the overall percentage of Democratic voters that would be satisfied in the event a candidate became the nominee. Instead, we're looking for conditional probability: given that a voter likes Candidate A, how often do they like Candidate B, and how is that different than Candidate B's overall performance.

insider 2020 banner

Samantha Lee/Business Insider

Advertisement

This lets us answer some pretty compelling questions. Some of the things we're looking at are:

  • Who stands to gain if Candidate B drops out?"
  • "Who don't Candidate A's supporters want to win?"
  • "Is Candidate A a good fit for vice president for Candidate B?
  • Which two candidates have lots of overlapping support to the point where if Candidate A dropped out Candidate B might thrive?

We'll be tracking this information over the next year and into the January Democratic primaries and throughout 2020's debates, gaffes, tweets, and October surprises as Democrats move closer to choosing their nominee.

The primary today is not about the Bernie wing vs. the Biden wing; that's a Spring 2020 question. Our polls are tracking which candidate is best-positioned in voters' minds to receive support after a more popular rival implodes or drops out, and which frontrunner is best poised to scoop up an endorsement when a lesser-known rival runs out of money.

To do this, INSIDER has been conducting a recurring SurveyMonkey Audience national poll. You can download every single poll here, down the individual respondent data. Down below we'll indicate which polls are currently rolled into the average, as after some time we drop out polls that have aged.

Here's what we asking:

Advertisement
  • Are you a registered voter?
    • Yes, No, Unsure / Don't know
  • Do you think you will participate in the 2020 presidential primary?
    • Yes, I will likely vote in my state's Democratic primary or caucus
    • Yes, I will likely vote in my state's Republican primary or caucus
    • No, I don't currently plan to vote in a major party's 2020 presidential primary or caucus
    • I don't know or I'd rather not say
  • Which of the following potential Democratic primary contenders have you ever heard of?
    • We present them with a list of contenders who have not yet removed themselves from the race and allow them to check as many as they wish to. Only the checked contenders are carried forward to the next questions.
  • Which of the following candidates would you be satisfied with in the event they became the Democratic nominee?
  • Which of the following candidates would you not be satisfied with in the event they became the Democratic nominee?
    • Both of these questions have the carried forward responses. If someone checks the first but not the second, we count them as satisfied with the contender, if someone checks the second but not the first we count them as unsatisfied with the contender, and if someone selects the candidate for both questions or neither questions we code that as neutral.

Not every time, but semi-regularly we also ask:

  • Which of the following candidates do you think would likely win an election against President Donald Trump? Check any that apply.
  • Which of the following candidates do you think would likely lose an election against President Donald Trump? Check any that apply.

For the conditional support questions, we only looked at respondents who said that they are registered to vote and will likely participate in their state's Democratic caucus or primary. It's hardly perfect but is our best approximation of what the mindset of likely voters looks like. For the win/lose supplement, we look at both the entire set as well as the subset of Democratic voters.

Here are the polls that are active in the model right now:

  • Poll 01: 1,109 respondents, 12/21/18 - 12/22/18
  • Poll 02: 1,233 respondents, 01/26/19 - 01/27/19
  • Poll 03: 1,093 respondents, 02/01/19 - 02/03/19
  • Poll 04: 1,117 respondents, 02/15/19 - 02/16/19*
  • Poll 05: 1,055 respondents, 02/24/19 - 02/26/19
  • Poll 06: 1,102 respondents, 03/01/19 - 03/02/19*

Asterisk denotes the poll also included the supplemental general election win/lose questions

Some caveats! As we mentioned before, we're a little skeptical of some of the topline numbers, and really only use them as the baseline to figure out whether a candidate has higher or lower than typical support among a rival's base.

Advertisement

Realistically, online polling balances for age and gender but not race. In a Democratic primary, particularly one where demography will potentially have a large impact on the outcome, it's good to be concerned about polls undersampling African-American and Latinx respondents. As always, when looking for a general sense of the election it's crucial to look at an amalgamation of polling to work around deficiencies in one single survey or another. All this is to say: we're not trying to get a general sense of the election, other places are way better resources for that, all we're really interested in here is shifts in conditional affinity, and as such our toplines should be taken with the typical degree of skepticism of any poll.

Second, it's early. We wanted to start early in the process to establish baselines and capture the brief moment before the ferocity and competition of the campaign really sets in. Voter preferences are still highly malleable, and many candidates are new to a national platform. It's important to keep in mind that this race is only just beginning, and these are not in any way intended to forecast the outcomes of the primaries. Yet, at least.

Either way, this is mainly an exercise to find out why candidates who are broadly ideologically similar are motivated to go after rivals that are politically most similar to themselves. We've got quite some time until the primaries begin properly, until the votes are cast. Everything else is just positioning.

Read More

Advertisement
{{}}