[EM] the intrinsic value of the metric of *strength* of personal preference (was: Re: Compatibility)

[Juho]

On May 5, 2010, at 8:29 PM, robert bristow-johnson wrote:

>
> On May 5, 2010, at 12:22 PM, Juho wrote:
>
>> In the pairwise comparison based methods we must also take into  
>> account the fact that we do not measure the strength of personal  
>> preferences, and therefore the decisions that we make are often  
>> simplified assumptions (i.e. one set of votes may refers to a wide  
>> range of different possible sincere preference strengths).
>
> isn't there a fundamental reason for this, Juho?  at least in the  
> political realm where disparate ideas are competing for influence  
> and power?  imagine a simple election between two candidates with  
> Range or Score Voting.  who will voluntarily dilute their vote and  
> not score their preferred candidate with the full 99 points (and the  
> other candidate with 0)?  citizens with franchise consider  
> themselves equal to each other with equal effect and the right to  
> representation in a democratic government.  this is one reason why i  
> usually say that the Range ballot requires too much information from  
> the voter (the other reason is requiring to voter to agonize over  
> whether he/she thinks some candidate is worth a grade of 46 or 51 or  
> 60, they may as well just get out their dart board).  if i like  
> Candidate A over Candidate B by just a little, and you like  
> Candidate B over Candidate A by a great amount, why would i dilute  
> my political preference to serve your political interest?  why  
> *should* i?  i want to count (even with my tepid preference) just as  
> much as you count.

Yes. We may get more distorted information if we try to collect more  
information. It may be difficult for the voter to decide whether to  
vote sincerely 61-65-77 or 61-65-78, or to exaggerate 0-25-99 or to  
vote strategically 0-0-99 or 0-99-99. And we may lose even the amount  
of information that we would have gotten with rankings. The voters may  
face a strategic dilemma that leads to unpredictable results. And in  
general the atmosphere gets worse because of widespread strategic  
voting. In competitive elections rankings offer a simpler, more  
sincere and well working alternative. In some non-competitive  
elections and polls ratings may be the best choice ((as the primary  
information to determine the outcome)).

>
> and the Approval ballot asks for too little information from the  
> voter and immediately presents the voter with a tactical dilemma if  
> he/she likes a particular candidate but approves of more.  in my own  
> State Senate district in Vermont, we *nearly* have Approval voting  
> (we vote for at most 6 out of maybe 20 candidates and the 6 highest  
> vote getters win state senate seats) and i have to consider each  
> election how many candidates i will actually vote for.  i think  
> human behavior will eventually drive Approval voting into the FPTP  
> gutter.

Yes.

>
> the Ranked-order ballot, what is used for IRV or Condorcet or Borda,  
> requires precisely the right amount of information from the voter:  
> who, between any pair of candidates, do you prefer?  it does not ask  
> how *much* you prefer one over the other, and i think that such is a  
> useless question.

Yes, in competitive elections that is the appropriate level. As I  
already noted in some non-competitive elections one could try to  
collect also more detailed information. Then we would probably make  
also the assumption that majority rule no longer applies but we would  
look at the sum of utilities, the worst utilities, nice distribution  
of utilities or something like that.

>
> sorry if i took this off-topic.

I think this is very much on-topic.

>
> BTW, Juho, i heard a BBC news story about your election tomorrow  
> regarding the desirability of FPTP vs. proportional methods in  
> electing Parliament.

James Gilmour already took the ball. I don't have strong links to UK  
although my email address is from that domain :-).

>  so it sounds like that the UK, with its new Lib-Dem party, is also  
> being confronted by the same problems.  i don't quite get it,  
> though.  aren't MPs elected out of geographic districts?  are there  
> more than one MP elected in any given district?  if it's only one MP  
> per district, how can a proportional method be used?  how can the  
> losing votes in one district be transferred to another district to  
> help elect someone there?  you would *have* to have more than one  
> candidate elected per district with all candidates running at large,  
> no?  if it's one MP per district, it's a single-winner election (and  
> then, of course, i would advocate for Condorcet, in a 3+ party  
> context).

It is also possible that withe FPTP and single-member districts we  
could have more than two stable parties since in different parts of  
the country the two leading parties (that the method favours) could be  
different.

Juho



>
> --
>
> r b-j                  rbj at audioimagination.com
>
> "Imagination is more important than knowledge."
>
>
>
>
> ----
> Election-Methods mailing list - see http://electorama.com/em for  
> list info