Defects in various election systems

classic Classic list List threaded Threaded
2 messages Options
Reply | Threaded
Open this post in threaded view
|

Defects in various election systems

Craig Carey-2
I quite like the idea of having voters select one candidate as first
choice, then mark "approvals" for as many others as they wish.  I
especially like it for primaries.  How would the winner be determined,
though?  And how would it work in delegate allocation in presidential
primaries?


At 8:22 PM 2/22/96, Rob Lanphier wrote:

>On Thu, 22 Feb 1996 [hidden email] wrote:
>> Condorcet
>>
>> Based on relative comparisons and not absolute approval or disapproval
>> (especially for second and later choices).
>
>>From a pragmatic point of view, how useful is information about the
>absolute preferences of the voters?  If we need to pick one and only one
>President, we only need to know which one (and only one) they like best.
>Since voters seldom agree, we *then* need mechanisms for determining
>compromises.
>
>Hence the ranked ballot.  Voters pick *one* top choice, and then can list
>other candidates the can "live with", in order of preference.
>
>Absolute approval/disapproval is irrelevent when we want people to
>*choose* a candidate.  If giving each candidate a grade is what we care
>about, then we should consider approval.
>
>> Circular result ties possible in
>> pair comparisons.
>
>A system that measures voter ambiguity is far superior to one that
>arbitrarily chooses a winner in spite of ambiguity.  Ties are highly
>unlikely in any election method with a sufficient number of voters,
>including Condorcet.
>
>> Approval Voting (AV)
>>
>> Based on absolute approval or disapproval. No relative comparisons among
>> multiple majority winners.
>
>It also gives disproportionate representation to voters who can't
>distiguish the difference between several candidates.  It is riddled with
>very real strategy problems.
>
>Consider the 1992 election, had it been an approval election.  Voters who
>preferred Clinton, but would take Bush in a last ditch effort to keep
>Perot out of office stood the danger of helping Bush get elected if they
>voted for Bush and Clinton.  Yet they also stood the danger of losing to
>Perot if they voted for only Clinton, since there was probably a large
>number of people who would have voted for Clinton and Perot to run Bush
>out of office.  Perot voters who really didn't like Clinton or Bush would
>be forced to vote for neither if Perot is who they truly wanted.  Bush
>voters would have had the same problem.
>
>Ultimately, voters should be expected to choose a candidate, and give
>fallback info if their first choice is not acceptable.  They shouldn't be
>able to choose a wad of candidates and leave it up to everyone else to
>decide.  Nor should voters be forced to use this as a method of
>expressing their preference.
>
>Rob Lanphier
>[hidden email]
>http://www.eskimo.com/~robla

*****************************************
Matthew Shugart
Associate Professor of Political Science

Address:
Graduate School of International Relations and Pacific Studies
University of California, San Diego
9500 Gilman Drive
La Jolla, CA 92093-0519

Phone:  619-534-5016
Fax:     619-534-3939
E-mail:  [hidden email]
*****************************************



Reply | Threaded
Open this post in threaded view
|

Defects in various election systems

Rob Lanphier-2
On Tue, 27 Feb 1996, Matthew Shugart wrote:
> I quite like the idea of having voters select one candidate as first
> choice, then mark "approvals" for as many others as they wish.  I
> especially like it for primaries.

Hmm, it does have a ballot simplicity.  But Condorcet's is far superior
in eliminating strategy.

>  How would the winner be determined,
> though?  And how would it work in delegate allocation in presidential
> primaries?

Got me.  A Condorcet-like calculation could be used (favorite is ranked
#1, approved are tied for #2, the rest unranked), but my feeling is that
as long as we are going to use something Condorcet-like, we might as well
use genuine articles.

Rob Lanphier
[hidden email]
http://www.eskimo.com/~robla