In this crowded primaries race, the candidates with the highest percentages in the polls and state primaries are those gathering the most attention. Sooner or later, however, candidates will drop out, and we’ll be down to a final race between two candidates. Who those candidates will be can be better predicted when voters’ second choices are known. A recent (Feb. 10th) Quinnipiac poll gives us an idea. In the following table, taken from Quinnipiac’s survey, each column corresponds to the 6 highest polling candidates. For each one, the table shows percentage of those who prefer that person, and below are the second choices if that candidate doesn’t win. So, for example, we see that 14% of those polled chose Warren, and of those 37% would take Sanders as their second choice.
Primaries: first and second choices
|
Biden |
Sanders |
Warren |
Klobuchar |
Bloomberg |
Buttigieg |
| First Choice |
17% |
25% |
14% |
4% |
15% |
10% |
| Biden |
|
19% |
13% |
10% |
21% |
17% |
| Sanders |
20% |
|
37% |
0% |
4% |
7% |
| Warren |
8% |
33% |
|
7% |
3% |
25% |
| Bloomberg |
33% |
4% |
7% |
5% |
|
21% |
| Buttigieg |
19% |
11% |
26% |
26% |
9% |
|
|
Note that Klobuchar was not offered as a second choice, which slightly warps the results.
General impressions from the table: Sanders and Warren are a clear block, each of whose voters clearly favor the other as a second choice. Bloomberg’s voters clearly prefer Biden, and indeed polls indicate that most Bloomberg voters are former Biden voters, accounting for Biden’s falling as Bloomberg rose. Biden’s voters are surprisingly open to Sanders, but Bloomberg’s are not. Buttigieg’s voters prefer Warren to Sanders. Klobuchar’s voters really like the other Midwesterner, Buttigieg, and don’t care much for the others, especially not Sanders. Buttigieg’s voters are open to anyone except Sanders.
Now, let’s estimate what happens for various contests, in which one centrist and one progressive remain, and each gains his or her own voters, plus the second choices of those whose first choices dropped out, based on the above table. These pairs don’t add up to 100%, because only not all polled gave a first choice or a second choice among those listed, and because I don’t account for the votes of people whose first two choices aren’t in the match.
Biden-Sanders 24% : 31% (+7% : +6%)
Biden-Warren 28% : 25% (+11% : +11%)
Buttigieg-Sanders 19% : 34% (+9% : +9%)
Buttigieg-Warren 18% : 24% (+8% : +10%)
Bloomberg-Sanders 24% : 34% (+9% : +9%)
Bloomberg-Warren 24% : 26% (+9% : +12%)
I’m pretty sure it’ll come down to a progressive vs. a centrist, but just for fun here are two progressives and two centrists against each other:
Sanders-Warren 30% : 19% (+5% : +5%)
Biden-Bloomberg 26% : 19% (+9% : +4%)
At a glance: Sanders wins against all three centrists; Warren wins against Buttigieg, and breaks even with Bloomberg and Bloomberg. Biden and Bloomberg are comparable in strength, and Buttigieg is weaker. Of course, all these numbers are very uncertain. The main winner in each of these matches is “other/don’t know”, and the numbers will change as voters’ opinions and perceptions of the candidates evolve.
The numbers in parentheses after each match show how much each candidate gains the votes of dropped candidates. The numbers are in general surprisingly even, showing that even without second-choice numbers, the first-choice numbers give a good picture of how various pairs of candidates will fare against each other.
My main takeaway is that it’s too early to count Biden out, as some have done. His numbers are down because of Bloomberg, but if Bloomberg ends up being a flash in the pan and fades out, Biden will gain his votes and will come out looking stronger. He is still the leading centrist. In contrast, Buttigieg is weaker in the national polls than his recent success in Iowa and New Hampshire might indicate.
