Promoting Plain Score
-
@gregw With a 5, 4... range, the step from 5 to 4 is a whopping 20% of the range. In other words, if the numbers were normalized to go from 0 to 1, the 5-0 range would be 1, .8, .6, .4, .2, and 0. So the difference between 1 and .8 at the top is .2. I feel that this would require me to use probability in voting. If I prefer Bush to Gore say, but think Nader is just off most voters' radar, I might want to flip a coin to decide whether to give Bush the highest score along with Nader or the next score, 20% down. Giving voters a finer-grained range near the top relives them from having to think about probability. They can just apply the exaggeration strategy directly in their scoring.
-
We have interesting choices, I rank them by my guess of initial voter acceptance: (Over time, I think voters would accept any of them.)
5, 4, 3, 2, 1, 0
10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0
100, 99, 90, 50, 10, 1, 0 (even though I rather like it)
8 + 4 + 2 + 1 (0 - 15)I am guessing that 8 + 4 + 2 + 1 (0 - 15) would become more annoying as the number of races on the ballot increases, but by the end of a long ballot the voters should be more used to it.
Is performance improvement worth voter reluctance?
Lately, I have been pondering the relative merits of {5, 4, 3, 2, 1, 0} and {10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0}.
Jonathan Quinn's VSE ratings show more improvement going from Score 0-2 and Score 0-10 than going from Score 0-10 to Score 0-1000. The exception is 1-sides strategy, where there is little difference between 0-2, 0-10 and 0-1000.
Quinn, suggests the VSE ratings are a measure of overall performance:
"In the field of voting theory, there are many desirable criteria a given voting method may or may not pass. Basically, most criteria define a certain kind of undesirable outcome, and say that good voting methods should make such outcomes impossible. But it’s been shown mathematically that it’s impossible for a method to pass all desirable criteria (see: Gibbard-Satterthwaite theorem, Arrow’s theorem, etc.), so tradeoffs are necessary. VSE measures how well a method makes those tradeoffs by using outcomes. Basically, instead of asking “can a certain kind of problem ever happen?”, VSE is asking “how rarely do problems of all kinds happen?”.
If the voter model, media model, and strategy model are realistic for a particular context, then VSE is probably a good metric for comparing voting methods. If you find a method which robustly gets a relatively high VSE, across a broad range of voter, media, and strategy models, then you can be confident that it reflects the will of the voters, no matter what that will is. That’s democracy."
Do you put much stock in the ratings?
-
@jack-waugh I don't think anybody wants to be adding up powers of 2 to work out which boxes to fill in. It's a non-starter. If the range covers more than a handful of scores, just have the voter write the score in.
-
@toby-pereira said in Promoting Plain Score:
If the range covers more than a handful of scores, just have the voter write the score in.
Not sure I want to trust voter penmanship. I already have misgivings concerning voter signatures on mail-in ballots.
Is 11 ovals is too many ovals?
-
@gregw said in Promoting Plain Score:
@toby-pereira said in Promoting Plain Score:
If the range covers more than a handful of scores, just have the voter write the score in.
Not sure I want to trust voter penmanship. I already have misgivings concerning voter signatures on mail-in ballots.
Is 11 ovals is too many ovals?
11 might be just about OK. For higher scores, you could have voters fill in two ovals to indicate any two-digit score. So two lots of 10 ovals.
-
@gregw said in Promoting Plain Score:
Do you put much stock in the ratings?
I think VSE is a pretty good clue. Maybe it could be refined by attention on the strategies it assumes voters use. I'm not saying there is necessarily anything wrong with Jameson Quinn's assumptions in that regard, but it could be interesting to examine that area for possible improvement.
-
This post is deleted! -
How about labeling the ovals 100, 80, 20, 0 and saying that filling in more than one averages them? That gives access to 90, 10, 50, 33 1/3, and 66 2/3. For hand counting, nine piles would be required for each candidate-oriented run-through of the ballots.
-
@toby-pereira said in Promoting Plain Score:
@gregw said in Promoting Plain Score:
@toby-pereira said in Promoting Plain Score:
If the range covers more than a handful of scores, just have the voter write the score in.
Not sure I want to trust voter penmanship. I already have misgivings concerning voter signatures on mail-in ballots.
Is 11 ovals is too many ovals?
11 might be just about OK. For higher scores, you could have voters fill in two ovals to indicate any two-digit score. So two lots of 10 ovals.
My own suggestion on this has been to use 0-5 and then add a bubble for 0.5; 0-5 is too few options but I think half-stars are enough.
Past research on this suggests people like this scale best, which is why so many companies use it.
-
@jack-waugh said in Promoting Plain Score:
How about labeling the ovals 100, 80, 20, 0 and saying that filling in more than one averages them? That gives access to 90, 10, 50, 33 1/3, and 66 2/3. For hand counting, nine piles would be required for each candidate-oriented run-through of the ballots.
This just seems overly complex.
@lime said in Promoting Plain Score:
My own suggestion on this has been to use 0-5 and then add a bubble for 0.5; 0-5 is too few options but I think half-stars are enough.
Past research on this suggests people like this scale best, which is why so many companies use it.
I don't see how adding half marks is simpler than doubling the range of integers. In any case, what do companies use it for? Are we talking about random surveys? In this post Sara Wolk posted this link about surveys as a justification for the limited options in STAR voting. But voting in an election is very different from answering previously unseen questions on some company's survey where motivation also also likely to be much lower. So we're potentially comparing apples with oranges. But I could probably live with half marks if they really are perceived as easier by the voter.
Finally, ranked-ballot methods have people write in the ranks. So I don't think having the write in scores rather than fill in ovals would be the end of the world. All things being equal, ovals might be better, but I don't see it as a deal-breaker.
-
@toby-pereira said in Promoting Plain Score:
I don't see how adding half marks is simpler than doubling the range of integers.
If you have individual bubbles for 0-10, fitting all of them on one piece of paper gets hard. In addition, finding the bubble you want to use is hard.
The main distinction is between ≤6 bubbles (subitization range) vs. >6 bubbles. For more than 6 bubbles, finding the bubble you want requires a "search", which is mentally costly and discourages intermediate ratings (which are more complex).
Usually this is handled by breaking the problem down into two subitization steps. This makes finding the best bubble easy, and also makes it easy for voters who don't want that extra precision to ignore it.
