[EM] Testing 1 2 3

[Dgamble997 at aol.com]

James Gilmour wrote:


>I think there is another very real problem here.  It is not clear from your 
original post >whether your model will modify the voting patterns under the 11 
different voting >systems, but my reading would be that it would not.  We 
know from practical >experience that real voters do change their voting patterns 
when you change the >voting system.

I thought somebody was going to make this point, it is of course true that 
voters change how they vote according to the system they are voting under. I 
have assumed that they vote sincerely and the same regardless of the system 
because this is the only real way to make a comparison.

>> STV was the most problematic system to model and the one I'm 
>> least happy with.

>You don't say what the problems were, but STV-PR is conceptually very 
different >from all the other
>PR systems and so there are special issues in modelling voter preferences.

The major problem was that the vote set for a 3 party 3 seat  election with 
party preferences:

41 A>B 
10 B>A 
9 B>C  
40 C>B

Could be the  1st count vote distribution:

A1 27
A2 14
B 19
C1 26
C2 14

Which gives parties A, B and C one seat each. 

Or

A1  21
A2  20
B    19
C1  20
C2  20

Which gives party A two seats and C one seat.

The model takes two extreme situations one in which the first preference vote 
of a party is equally divided between all the candidates of that party, 
another in which one candidate of each party has all the first preference votes. It 
then calculates the result under both sets of assumptions and if they are 
different decides the result using a random number.

David Gamble







-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.electorama.com/pipermail/election-methods-electorama.com/attachments/20040103/d822dce8/attachment-0003.htm>