[EM] Election methods in student government...

[Juho]

I support raphfrk's thoughts that the basic methods might be the best  
choice. Open list and PR-STV are two quite well working basic PR  
methods.

Raphfrk proposed also asset voting. Also that one is quite simple and  
natural. It may add some "asset trading", which might be a bad thing  
in this environment that I understood was quite interested in  
plotting. Well, maybe they'd really like that second negotiation  
round of asset voting and be happy to use this method. Or  
alternatively they should not be given the chance to start this kind  
of horse trading.

I have also another less common method that I have promoted on this  
mailing list. Open list methods can be extended to cover also  
hierarchical candidate settings. The lists are no longer lists, so  
"tree voting" is a better name for this method. The candidates (and  
parties and other groupings) thus agree a tree-like structure where  
candidates are leaves and branches of the tree may have descriptive  
names (party names, party branches, interest groups). Voters vote  
simply one of the candidates. The seats are allocated to the branches  
of the tree (based on the number of votes received by each branch)  
starting from the root using some suitable PR allocation method. This  
process is then repeated towards the smaller branches until all the  
seats have been allocated to the candidates in the leafs.

Tree voting gives candidates some more information on the candidates  
than traditional open list does (from party alliances to small  
interest groups). Also the division of seats within each party is  
more fair (not plurality based but proportional also within parties  
and party alliances). Tree voting doesn't allow giving her personal  
inheritance order in the ballot (in STV style). But the candidate  
given order of inheritance (in the form of the tree) is there, and it  
may reflect the true thoughts/intentions of the candidates quite  
accurately (=not much space for the candidates to tell each voter  
group that the candidate would promote especially their values). In  
elections where there are numerous candidates expecting all voters to  
know the candidates well enough to give them a STV like preference  
order may be too much. In elections where all candidates are well  
known that might be ok.

Juho Laatu


On Dec 21, 2006, at 14:34 , raphfrk at netscape.net wrote:

>  > From: thully at umich.edu
> >
> > Anyway, I have been investigating alternate systems for single- 
> winner
> > elections and (especially) multi-winner elections.  Party list is  
> out
> > - the less rigidly-defined party structure makes it even less fair
> > that it would be in a national election.
>
> There is a system called open party list.  Basically, the voter votes
> for 1 candidate.  The party that that candidate represents gets the  
> vote
> and the seats are split on that basist between the parties.  However,
> the party member that actually fills each seat is determined by how  
> many
> votes the candidate obtained for the party, e.g. the party's highest
> vote getter gets the first party seat and so on down until all the  
> party
> seats are filled.
>
> This allows the voters to choose the parties but also choose which
> party member gets elected.
>
>
> > I have also investigated STV (and IRV for single-winner)   
> However, the
> > lack of monotonicity is quite troubling - the fact that you can help
> > elect a candidate by ranking them LOWER seems almost undemocratic.
>
> In the multi seat case, I don't think PR-STV is that bad.
>
> > This, coupled with the fact that the current system *replaced* STV
> > some 20-odd years ago, dampens my enthusiasm for that method a bit.
>
> That doesn't mean that STV is bad, it just means that it is bad for
> the majority.  Any PR system is bad for the majority.
>
> Look at what the new system actually does ... it give most of the
> seats to the majority party/faction.  Who decided the current system?
> Presumably, it was the majority party/faction.
>
> > I have also seen plenty of other election methods that look  
> interesting
> > - Concordet methods especially.  However, these methods are quite  
> complex
> > and don't have any good multi-winner variant (there is CPO-STV,  
> but it is
> > extremely complex and is still non-monotonic).
>
> I would suggest asset voting as a really simple way to get PR.
>
> In its most simple form, each voter votes for 1 candidate.  Any  
> candidate
> who gets the quota is elected  and can transfer his excess to any  
> other candidate.
> All the other candidates can also transfer votes in order to bring  
> one of
> them to the quota.
>
> > Right now, I'm kind of at a loss as to what the best voting  
> system would be.
>
> It depends on your objectives.  There is no 'best' system.
>
> > It's obvious that the current system isn't it
>
> Yeah, it does seem pretty bad.
>
> > I also don't like the idea of using a system that is so complex that
> > it can't be reasonably explained to non-technical types.
>
> > I also don't like the idea of using a system that is so complex that
> > it can't be reasonably explained to non-technical types.
>
> There is also the problem that if it is that complex, explaining it to
> the vote counters could even  be a problem.  Also, some of them  
> require a
> computer to count the votes.  IMO, that is to complex for a student
> government.
>
> > regular STV is about at the maximum complexity I would want
>
> This is one way to give the rules for asset.  It is not the full
> transfer method though, but it gives clear rules for the transfers.
>
> You could maybe describe it as PR-STV, but that the candidates decide
> the transfers.  However, maybe that would be a bad idea if PR-STV has
> bad press.
>
> (Anyway, here's my suggestion for the rules)
>
> The candidates, equal in number to the number of seats to be filled,
> holding the most votes, shall be given the option of transferring
> some of their votes to other uneliminated candidates.  Each of the
> candidates shall decide how many of their votes and to whom to
> transfer them to.  After, the transfer has been completed, a candidate
> shall be eliminated.  If no candidate volunteers, the candidate with
> the fewest votes shall be eliminated.  The eliminated candidate shall
> be given the option to transfer any votes held to other candidates.
> This two step process shall be repeated until the number of  
> uneliminated
> candidates is less than or equal to the number of seats to be filled.
> The remaining uneliminated candidates shall be deemed elected.
>
>
> It also allows a candidate to volunteer to stand down.  This helps
> with monotonicity issues.
>
> > Even standard STV is almost too complex - part of the reason it was
> > originally eliminated is due to its complexity!  Due to these  
> reasons,
> > ease of use, understanding, and transparency is paramount.
>
> Maybe I am being cynical, but I would say that was the excuse rather
> than the reason (or at least only part of the reason).
>
> > So far, all I have came up with which seems to potentially be a good
> > method is a variant of sequential proportional approval voting.   
> Under
> > the system, single winner elections would be simple approval voting.
> > However, for multi-winner elections each student would begin with a
> > set number of "points" equal to the number of seats to be elected.
> > Votes would be counted as in normal SPAV, and each weighted  
> according
> > to the number of points each student has remaining.  Every time a  
> voter
> > elects one of their choices, they would "use up" one of their  
> points.
> > This seems a little more understandable than standard SPAV, and it
> > hurts groups that share some preferences with the majority less.
> >
> > Is there something that would potentially be better while not  
> becoming too complex?
>
> You are basically weighting each ballot as
>
> E = number of elected candidates on ballot.
> N = number of seats
>
> Your way:
>
> (N-E)/N
>
> (the divide by N is to just rescale it, so the 2 match for E=0)
>
> instead of
>
> Normal method:
>
> 1/(1+E)
>
> Assuming, 5 seats,
>
> E (normal) (your way)
> 0: 1 1
> 1: 0.5 0.8
> 2: 0.33 0.6
> 3: 0.25 0.4
> 4: 0.20 0.20
>
> Your system greatly decreases the penalty for electing a candidate.
> This will be a benefit to larger parties.  The normal way has the  
> advantage
> that a faction of a Droop quota always gets a seat, while under your
> system it doesn't.
>
> I think in a small number of seats case, it may not be quite as big  
> a problem.
>
> However, if there were 5 seats and 3 parties with
>
> A: 77%
> B: 22%
> C: 21%
>
> A got 3.85 seats
> B got 1.1 seats
> C got 1.05 seats
>
> A will get the first 3 directly.
>
> Round 4:
> A: 78 * ( 5-3) = 154
> B: 22 * ( 5-0 ) = 110
> C: 21 * (5-0) = 105
>
> A wins again
>
> Round 5:
> A: 78 * ( 5-4) = 78
> B: 22 * ( 5-0 ) = 110
> C: 21 * (5-1) = 105
>
> B wins
>
> A wins 4 seats and B wins 1 seat.  However, C had more than 1/5 of  
> the votes
> so should have won at least 1 seat in a 5 seater election.
>
> > ... and not to complex
>
> Asset has the advantage that the voting stage is very easy for
> the voters .... just pick your favourite candidate.
>
> The candidates then move around the votes after the election in  
> order to
> give PR.
>
> You can't get much less complex than that.
>
>
> Raphfrk
> --------------------
> Interesting site
> "what if anyone could modify the laws"
>
> www.wikocracy.com
>
>
> Check Out the new free AIM(R) Mail -- 2 GB of storage and industry- 
> leading spam and email virus protection.
> ----
> election-methods mailing list - see http://electorama.com/em for  
> list info

Send instant messages to your online friends http://uk.messenger.yahoo.com