[EM] Manipulability stats for more poll methods (fixed footnotes)
Kristofer Munsterhjelm
km_elmet at t-online.de
Sun May 5 04:17:52 PDT 2024
On 2024-05-05 00:25, Michael Ossipoff wrote:
>
>
> On Sat, May 4, 2024 at 14:45 Kristofer Munsterhjelm
> <km_elmet at t-online.de <mailto:km_elmet at t-online.de>> wrote:
>
>
> Yes, that's right. But consider a voter with the following utilities:
>
> A: 0.57
> B: 0.32
> C: 0.23
> D: 0.08
>
> Normalization to two steps fixes the highest value (0.57) to 1 and the
> lowest value (0.08) to 0 and rounds off the intermediate values after
> linearly scaling them.
>
>
> Yes. So far, so good. But…
>
> This in essence says that a value is rounded off
> to 1 if it's greater than or equal to 0.325 (the midpoint between 0.08
> and 0.57)
>
>
> What? You didn’t average the normalized values. You averaged two of the
> values before normalization. The midrange isn’t usually the same as the
> mean. You used the midrange as the mean.
Yes, that's my point. What I call "normalize" uses the midrange, because
it's a generalization of the following approach for mapping utilities to
a cardinal scale:
Let min be the minimum utility (the utility of your worst candidate),
and max be the utility of your favorite. Then the normalized ratings
from 0 to k inclusive are given by
r_i = round( (x_i - min)/(max-min) * k), where x_i is the utility of the
ith candidate, and r_i is the rating of that candidate.
I thought this was the natural way to create Range ballots from
utilities. To my knowledge, there's no obvious way to generalize
above-mean utility thresholding to say, a 0-10 scale.
I'm aware that "normalize" as I've defined it above is different from
above-mean thresholding, and that the latter is more common in
treatments of approval voting in particular. For that reason, I use mean
utility cutoffs for the approval hybrids, and range-based normalization
for the Range hybrids. My point was just that one shouldn't confuse the
two and think I used the same approach for approval and Range.
-km
More information about the Election-Methods
mailing list