On Sat, 11 Apr 1998, DEMOREP1 wrote:
> Example
> 0 to 100 rating range on each candidate
>
> 49 voters vote A 100 and B O
> 50 voters vote B 1 and A 0
>
> A median = 0
> B median = 1
>
> Should B be elected ?
----
Mr. Lanphier wrote on Sat, Apr 11, 1998 7:43 PM EDT--
Yup. A majority prefer B over A. Furthermore, there's no way of knowing
whether those who prefer A are overstating their preference.
----
D- I raise the acceptability question again. Neither A or B is acceptable to
a majority of the voters. The 50 B 1 voters are not exactly showing strong
support. A and B just might be Hitler and Stalin in 1932.
However---
For making 1 choice of 2 candidates there could be the following scale
values--
3
2
1
0
Median 1.5
A 3 or 2 vote indicates acceptable (above the median)
A 1 or 0 vote indicates unacceptable (below the median)
With N candidates, the general case would be having 2N-1 scale values (if 0 is
to made a value) or simply 2N (if 0 is not to be used) (e.g. possible scale
values of 4, 3, 2 or 1- median 2.5).
The median idea can be regarding as a balance beam
0 ___|__|____|____|_|_ 2N-1
/\
Unacceptable Acceptable
Only if a candidate had a median greater than half of 2N-1 could he/she be
elected.
If the 0 to 100 scale is used, then only if a candidate had a median greater
than 50 could he/she be elected.
Does the above thus combine Approval Voting and Condorcet (the Lanphier-
Demorep1 method) ?