[EM] [Election-Methods] [english 94%] PR favoringracialminorities

[Juho]

On Aug 15, 2008, at 20:05 , James Gilmour wrote:

> Juho > Sent: Friday, August 15, 2008 5:24 PM
>> If you have some issue X, wouldn't it also be natural to have one
>> list "for X" and one list "against X"? I.e. lists but not "party
>> lists". You may need to arrange the candidates anyway according to
>> their opinions in some "lists" to make it clear to the voters who are
>> "against" and who are "for".
>
> Such formal "issue lists" are both undesirable and unnecessary.   
> Some issues divide on party lines, some issues divide within
> parties, and some issues run across some or all parties.  In fact,  
> in deciding their reactions to the candidates who have offered
> themselves for election, voters can and do take many more factors  
> into account:  party, issue X, sex, race, language, locality,
> issue Y, etc, etc.    Different voters will weight these factors  
> and these combinations of factors differently.  Of course, that
> depends on having a voting system that is sufficiently sensitive to  
> allow the voters to express their views in this way.  But
> preferential voting provides a workable compromise, where each  
> voter consciously or unconsciously condenses his or her n-dimensional
> assessment of the candidates into one dimension to put the  
> candidates into a personal order of preference.

I understood that in this case the parties were irrelevant and  
therefore basic lists may be sufficient to put in place a structure  
that covers all the relevant questions.

In theory, if there are more than one issue, e.g. X and Y, then there  
could be more lists, e.g. for "X and Y", "X and not Y", "not X and Y"  
and "not X and not Y".

I mentioned also the tree option to cover more complex structures.  
(And it is possible to extend from that, but that will get more  
complex then, maybe deleting the basic simplicity of voting in list  
and tree based methods.)

One thing worth noting when comparing list based and STV style  
methods is that although the voters have all kind of opinions the  
candidates represent a more limited set of opinions. One candidate  
could be said to have opinions X and Y in that order of importance.  
As a result that candidate could represent "section Y of party X".  
Probably also the voter has some order of importance in her opinions,  
and could therefore vote either "section Y of party X" or "section X  
of party Y". Very complex preferences will be lost in the rounding  
errors in any method.

Lists and trees are not quite as flexible as STV but on the other  
hand in them voting is much simpler.

There can be also some benefits in forcing the candidates to declare  
their association and priorities. If they are not declared that opens  
some doors to more vague marketing, promises in all directions and  
possibility that the candidate will promote different things after  
being elected than what the voter expected.

>> STV-PR gives the voters some  flexibility
>> that the list (or tree) based methods do not give but here I didn't
>> see anything special that would speak against the use of lists.
>
> STV-PR does not provide "some" flexibility  -  STV-PR provides  
> complete flexibility for each voter to express her or his personal
> preferences among all the candidates on whatever basis that  
> individual voter chooses.  Your preferences and mine may be identical,
> but it is probably that we have placed the candidates in that  
> common order for quite different reasons.

STV-PR is more flexible than list based methods ("some" or more than  
"some"). Sometimes that flexibility may also bring problems like long  
votes and need to analyze all the candidates. If there are very many  
candidates it could be useful to allow also inheritance by default  
(for short votes) (to parties or to candidate's own favourites  
(includes also risks like in Fiji)) or group names in votes, e.g.  
MyCandidate>MyParty. (Hybrid methods between lists and STV are thus  
also possible.)

>> (Lists may also be more practical in some cases, e.g. if the number
>> of candidates is high.)
>
> The largest STV-PR election I know of had 450 candidates for 120  
> places, but I would not recommend such a high district magnitude!

What was the number of districts here? Was this a single district  
with 120 seats or are the 120 places a sum of smaller districts?

> To allow for diversity in representation you also need a reasonable  
> minimum, so that the larger parties will be "forced" to nominate
> at least two candidates each, so that the voters get some choice  
> within parties as well as between them.

Are there problems with having too few candidates somewhere? At least  
in list based elections it seems that parties prefer nominating many  
candidates rather than one or few. In STV there might be some risk of  
vote splitting when parties nominate multiple candidates (since  
voters could either forget to rank their own party candidates or rank  
candidates of other parties instead of them). Maybe also the single- 
member tradition has some psychological influence here.

>   For public elections,
> however, there is practical trade-off because electors, especially  
> those brought up with decades of single-member districts (UK,
> USA), will want a guaranteed level of local representation.  Where  
> you can strike the balance in that trade-off will almost
> certainly vary from country to country.

Yes, there is a trade-off between locality of the representatives and  
accuracy of proportional representation (it is also possible to have  
both, but that means some other "rounding errors" like not electing  
the most popular candidates in each district). If the number of seats  
in each district is higher than one then the idea of one  
representative who knows that she represents all the local voters is  
already gone (there will be uncertainty on who represents whom) => it  
is then easier to go for higher seat numbers per district too.

Juho


>
> James Gilmour
>
> No virus found in this outgoing message.
> Checked by AVG.
> Version: 7.5.526 / Virus Database: 270.6.3/1610 - Release Date:  
> 13/08/2008 16:14
>
>
> ----
> Election-Methods mailing list - see http://electorama.com/em for  
> list info


		
___________________________________________________________ 
All New Yahoo! Mail – Tired of Vi at gr@! come-ons? Let our SpamGuard protect you. http://uk.docs.yahoo.com/nowyoucan.html