[EM] electing a variable number of seats

[Andy Jennings]

As far as iterative methods go, I like RRV.  It seems to be the natural way
to extend the divisor methods to work with range-style inputs.

But I feel like the combinatorial methods will give better proportionality
than iterative methods.  If there are lots of candidates, the best
four-winner set and the best five-winner set might be totally disjoint.
 Consider a worst-case, two-dimensional scenario:
- Voters are equally spaced around a circle.
- There are nine candidates, all of them on the same circle as the voters.
- The first four candidates form a perfect square and the last five form a
regular pentagon.
Obviously the first four candidates are the most representative 4-set and
the last five are the most representative 5-set.

It is clear that an apportionment method should be iterative to avoid the
Alabama paradox.  The number of representatives a state is assigned in the
legislature should not decrease if the size of the legislature increases.

But I don't think this paradox is that problematic when choosing winners in
a PR method.  Don't you think that people could be made to understand that
candidate A is a good representative if we're choosing four winners, but if
we're choosing five winners then we can choose an even more representative
set of winners that doesn't include A?

Andy Jennings



On Sat, Feb 19, 2011 at 8:55 AM, Brandon Wiley <brandon at blanu.net> wrote:

>
>
> On Sat, Feb 19, 2011 at 5:21 AM, Kristofer Munsterhjelm <
> km-elmet at broadpark.no> wrote:
>
>> Brandon Wiley wrote:
>>
>>> While I think Range Voting would work great here, if for some reason it
>>> doesn't go over (sometimes people think it seems complicated) then Approval
>>> Voting would also be very easy to use. Again just rank candidates by number
>>> of approvals and take the top X.
>>>
>>
>> Both bloc Range and Approval will let a majority determine the composition
>> of the entire board. Therefore, I don't think that would be very
>> proportional.
>>
>
> Well RRV <http://rangevoting.org/RRV.html> is the PR-compatible method for
> using Range Voting. The debate of PR vs. non-PR multi-winner voting methods
> is an issue in itself, but either way you can use Range Voting, either RRV
> or simply RV, respectively.
>
> ----
> Election-Methods mailing list - see http://electorama.com/em for list info
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.electorama.com/pipermail/election-methods-electorama.com/attachments/20110219/ef87e47e/attachment-0004.htm>