What about a meta election?

Mike Ositoff ntk at netcom.com
Tue Oct 6 19:43:40 PDT 1998

> Charles Fiterman wrote:
> > 
> > At 02:43 AM 10/3/98 EDT, you wrote:
> > >Any tiebreaker must obviously be mathematically understandable to John/Mary Q.
> > >Voter with 100 political I.Q.'s.
> > 
> > By the definition of I.Q. half the voters will have
> > over 100 and half under, its a median. As an election
> > judge I can attest that 90% of the problems in an
> > election will be due to the bottem 10% on the scale.
> > 
> > You can't disenfranchise people just because they
> > don't understand some complex system.
> IMO "consent of the governed" requires that most (by far) of a
> population understand something as fundamental as how their votes are
> being counted, how the system can be manipulated, etc.  I believe this
> is a responsibility that goes along with voting.  It is not enought that
> the 90% simply know how to physically enter their votes.
> Sending the votes through a "black box" will also tend to weaken the
> public's confidence in the outcome, especially in close races, or when
> results appear to be counter-intuitive, or event when they differ from
> pre-election polls.  Could a situation arise where the public trusts
> poll results more than the actual election?

Yes, and this strongly argues for Approval. Though Approval
doesn't meet all the standards & criteria of the best rank-methods,
it goes most of the way, and its simple count won't have
the above problem.

That's true, when a sizable minority don't like the outcome
of a rank-count, and understand the count rule barely or not
at all, there'll be trouble. In fact, after each election, the
debate would heat up all over again about how rank-ballots
should be counted. Some claim that IRO is simple, but it also
produces perverse results, and will cause dis-satisfaction.
It even violates monotonicity, so that you can defeat someone
by ranking him higher.



More information about the Election-Methods mailing list