[EM] Truncation

[Joe Weinstein]

(This message repeats one just sent, inadvertently, without the address 
mentioning the intended subject:  Truncation.)

TRUNCATION

Various recent postings in effect discuss special cases of the general 
problem of how to treat ‘truncation’, i.e. a voter’s omission of a mark or 
evaluation for one or more candidates.

An election method can use various devices and approaches to SCORING - i.e. 
for deriving the winner(s) from the assemblage of marked ballots.  Scoring 
can feature or incorporate summation or averaging (as in CR, including 
Approval; and as in constrained-marking approaches, such as cumulative 
voting, including usual Lone-Mark plurality).   Other scoring approaches and 
devices include pairwise comparisons (e.g. Condorcet) and vote transfer 
(e.g. IRV).

But, as to what sort of voter MARKING of ballots is allowed, there is every 
reason for the method to promote voter participation by allowing truncation.

There is no point in invalidating or penalizing ballots of voters who have 
reason to lack perfect information (especially when there is a gaggle of 
mostly unknown candidates), or who do have (for their purposes) perfect 
information but are simply trying to minimize time and effort and confusion 
in marking their ballots.   Rather, when a given candidate has not been 
explicitly marked, he can still be given a default evaluation.

A good election method will accept a truncated ballot as valid, and use a 
reasonable and useful convention to specify the default evaluation (grade or 
rank) - a convention which accords with common sense and is stated clearly 
in the ballot instructions.

I have previously argued that the best common-sense convention is just to 
award each unmarked candidate the lowest possible grade.  However, as Forest 
Simmons in effect has lately (11 Sep 2002) noted convincingly: when a 
candidate is left unmarked, the voter’s usual intention (to be realized, as 
much as possible, in the operation of the scoring mechanism) is that the 
candidate ranks below all actively favored candidates but above actively 
disfavored ones.

When a method allows three or more grade (or rank) levels, it is desirable 
that there be (by intention, for scoring purposes, insofar possible) at 
least one grade for actively favored candidates, at least one grade for 
actively disfavored candidates, and an intermediate neutral grade for 
candidates whom the voter (whether on account of ignorance or another cause) 
finds insufficient reason either to actively favor or actively disfavor.

When just three grades (or ranks) are allowed, it may be more useful to take 
the middle grade as intended for candidates perceived as favored but less 
intensely so; and the bottom grade as intended for both neutral and 
disfavored candidates.  Such is the approach used, for instance, in MCA - 
Majority Choice Approval.  However, with four or more allowed grades, 
neutral and disfavored candidates may be distinguished, while still allowing 
two or more different grades for favored candidates.

Forest notes that the usual American academic grading system - which has the 
advantage of being very familiar to almost all Americans - allows the six 
grades A-E.  Typically,  grades A-D are degrees of definite pass, grade F is 
definite fail, and grade E indicates that information is incomplete to infer 
a definite pass or definite fail.

Accordingly Forest proposes that if a given election is to be run so as to 
allow voters to use at least four grades, then a good approach would be to 
allow use precisely of the six academic grades, with ungraded candidates 
being awarded the default grade E.

This proposal makes excellent sense.  Four levels are surely enough to 
distinguish substantially distinct degrees of active approval.  One level 
suffices to express active disapproval.

Remember that the academic grading system is the most frequently used and 
personally meaningful ‘election method’ - or anyhow evaluation scheme - in 
the personal life of most Americans, certainly of most younger American 
voters.

Joe Weinstein
Long Beach CA USA







_________________________________________________________________
MSN Photos is the easiest way to share and print your photos: 
http://photos.msn.com/support/worldwide.aspx

----
For more information about this list (subscribe, unsubscribe, FAQ, etc), 
please see http://www.eskimo.com/~robla/em