Show-of-hands voting in meetings

[Mike Ossipoff]

In 2 separate letters, Tom mentioned the 2 methods that I
consider best for a show-of-hands vote, when pairwise show-of-hands
voting is un-feasible due to too many alternaives & voters &
too little time (& no computer).

But I'd make changes in both.

1. Inclusive 2nd Ballot:

Yes, for such a quick show-of-hands method, this is the best.
Though, from the latest that I've heard, France, in its latest
version of this, excludes from the 2nd balloting anyone who
got less than 12% of the vote in the 1st balloting, there
was at least 1 period during which France used _inclusive_
2nd balloting, where all the candidates of the 1st balloting go
into the 2nd one. That's better. 

But, if I felt that the committee using it would accept this
improvement, I'd improve Inclusive 2nd Balloting by making
the 2nd balloting an Approval vote instead of a 1-vote Plurality
vote.

This would also be good for public political elections where
people wouldn't accept rank-balloting, &/or wouldn't accept
the expense of setting up voting technology that would make
it convenient. For instance, for a state electoral law that
has to take into account that many counties don't have,
& don't want to set up the technology that would make rank-
balloting convenient (I'm not saying that's the case; I'm
just saying "if").

For public political elections, where a 1-dimensional political
spectrum is at least a good possiblity, or sometimes a fair
approximation, I'd propose the above method as-is.

But for meetings where there's no 1-dimensional issue
spectrum, I'd rather make both ballotings Approval ballotings.

The advantage of not using Approval initially is that it
gives more useful information when there's a 1-dimensional
spectrum. But when that condition doesn't obtain, at least
the use of Approval in the 1st balloting would give an idea
how the various alternatives can do in an Approval election.

Let me re-name & define those 2 methods:

Majority//Approval:

Ask people to vote for their favorite in a 1st balloting. If 
anything gets a majority it wins. If not, then hold a 2nd
balloting by Approval.


Approval-Majority//Approval:

Hold an initial Approval vote. If 1 or more alternatives have
vote total equal to at least 1/2 the number of voters, then
the one with most votes wins. Otherwise hold a 2nd balloting
by Approval.

Though I prefer Approval-Majority//Approval for meetings
without a 1-dimensional spectrum, the choice of which to
propose would, even then, depend on whether voters would
accept the slightly wordier definition (especially the
way it defines a majority).

In fact, some people don't accept Approval, mistakenly believing
that it violates a meaningful interpretation of 1-person-1-vote.
For that situation it might be easier to just propose
ordinary Inclusive 2nd Balloting (Majority//Plurality), rather
than try to explain Majority//Approval or Approval-Majority//Approval.

***

2. Repeated Ballotings:

This one takes longer than the 2-balloting methods, but would
seem to take roughly about as long as the vastly worse IRO, and
less than Pairwise show-of-hands voting (one version of which
could be a Condorcet version). Like the 2-balloting methods,
this seems the best show-of-hands method if we exclude
methods that take longer.

Tom, it seems to me, suggested a fixed number of ballotings,
for when no majority is won. I'd rather say that, if nothing
gets a majority then the number of ballotings is N-1, where
N is the number of alternatives. More than that is unnecessary.
Fewer than that is insufficient for being as free as possible
from strategy dilemma.

Of course I'd make these Approval ballotings, if the voters
would accept that. And that means that the way to win in
a balloting before the N-1th one would be that 1 or more
alternatives have vote totals of at least 1/2 the number of
voters, and then the winner is the one with the most votes.

This method is similar to the one proposed by Robert's Rules
of Order, except that it seems to me that they don't
specify any limit to the number of ballotings. As I said,
I'd limit it to N-1, since there's no need for more than that.

Another method this is similar to, in its Approval version,
is Bucklin. It's a show-of-hands Bucklin.

***

Mike



--