Ranked Choice Poll Mess/NYC Mayoral Race
We are thinking about “lanes” – blame modern polling pundits for that – but voters apparently aren’t. It’s the easy way to break down politics, especially if you are a numbers guy, nNuAmTbTeSrIsI LgVuEyR… and don’t know much about politics.
Ranked Choice Polling (asking what voters will do in a ranked choice elections) gives away the mistake.
Here’s the last ranked choice poll that I can find on wikipedia:
Notice what happens when Morales is eliminated? Adams and Yang don’t gain votes. Stringer and Garcia gain 1% each. And Wiley gains 5%. That’s like 70% of Morales’ vote going to Wiley. That’s a lane, isn’t it?
But watch what happens when Stringer goes out. His 13% goes to Adams (3%), Garcia (4%), Wiley (5%), and Yang (3%). Pretty evenly split. No lane.
And previous polls? A May 17 poll has Morales’ vote going 1% to Adams, 2% to Garcia, 1% to McGuire (!), 3% to Wiley. No lane. And Stringer’s going 3% Adams, 4% Garcia, 3% Wiley, 3% Yang. No lane at all. More of his votes go to Garcia than to Wiley?
Where do Stringer’s votes shift?
| Date | Adams | Donovan | Garcia | McGuire | Morales | Wiley | Yang |
| May 24 | 3% | 4% | 5% | 3% | |||
| May 17 | 3% | 4% | 3% | 3% | |||
| *May 15 | 17% | 14% | |||||
| *May 12 | 6% | 5% | 6% | ||||
| *May 12 | 12% | 10% | |||||
| *April 21 | 13% | 13% | |||||
| **April 15 | |||||||
| *Mar 18 | 13% | 13% |
*Wiley and Morales were already eliminated
**Runner-up, votes did not shift
Where do Wiley’s votes shift?
| Date | Adams | Donovan | Garcia | McGuire | Morales | Stringer | Yang |
| ***May 24 | 6% | 9% | 4% | ||||
| ***May 17 | 16% | 14% | |||||
| May 15 | 1% | 0% | 2% | 1% | 1% | 0% | 1% |
| May 12 | 3% | 3% | 3% | 3% | |||
| May 12 | 7% | 5% | 6% | ||||
| April 21 | 4% | 6% | 7% | ||||
| April 15 | 4% | 5% | 7% | ||||
| Mar 18 | 4% | 4% | 7% |
***Stringer and Morales were already eliminated
Where do Morales’ votes shift?
| Date | Adams | Donovan | Garcia | McGuire | Stringer | Wiley | Yang |
| May 24 | 0% | 1% | 1% | 5% | 0% | ||
| May 17 | 1% | 2% | 1% | 0% | 3% | 0% | |
| May 15 | 1% | 0% | 2% | 1% | 1% | 0% | 1% |
| May 12 | 0% | 0% | 2% | 0% | 1% | 2% | 1% |
| May 12 | 1% | 0% | 0% | 0% | 1% | 3% | 2% |
| April 21 | 1% | 1% | 2% | 3% | 2% | ||
| April 15 | 2% | 1% | 1% | 3% | 2% | ||
| Mar 18 | 1% | 0% | 1% | 0% | 0% | 1% | 0% |
JD2718 
Interesting. As a (very novice) student of voting theory, I’m strongly in favor of shifting US elections to an approval voting system (relatively easy to implement with existing voting mechanics, easy for voters to adopt, reduces a lot of the problems with strategic voting, should eliminate the path DJT followed to the GOP nomination in 2016….)
That said, it may be that ranked choice is the right structure for polling questions like this one, especially if the dynamics of the election mean that many current candidates will drop out before voting starts.
Approval voting has massive strategic voting problems, much worse than any ranked-choice systems (where strategic voting is possible in theory but not generally in practice).
The theory behind elections, and behind primaries, is different. I think it is very weird that ranked choice is being used in a primary.