[EM] Voting system "branding" poll results

[Jameson Quinn]

The results are in for the voting system branding
poll<http://betterpolls.com/v/1189>,
in which I polled new names for using with the general public to advocate
the main classes of voting systems. Just as "IRV" is more descriptive and
user-friendly than "Hare", there must be less-opaque names than "Bucklin" or
"Condorcet" which we could use.

The site where I ran the polls, betterpolls.com, does a really thorough job.
It uses a generalized range ballot to infer the winner under a number of
voting systems, does a good job of putting the results in exhaustive tables.
So if you're interested in that, follow the link above. Here's a quick
summary, followed by my comments.

There were only 16 votes. I said before the poll that I was hoping it would
reach 25 votes, so 16, while not horrible, is not ideal. Still, I think that
the results are valid.

*First question: Name for Condorcet*. (I know that there are of course many
variants of Condorcet, but since they mostly agree for practical cases where
the Smith set is 1-3 candidates, for the purposes of advocacy they can be
lumped together.)

Winners by system:
SystemWinning nameInstant Runoff Normalized Ratings Condorcet VotingVirtual
Round Robin*Round Robin Voting (RRV)* -- nice contrast with IRV. Could be
Instant Round Robin Voting (IRRV)Raw Rating SummationRound Robin Voting
(RRV) -- nice contrast with IRV. Could be Instant Round Robin Voting (IRRV)
ApprovalCondorcet VotingInstant Runoff VotingPairwise Champion Voting (PCV)
Essentially, "Condorcet" had the broadest consensus, but "RRV" (an early
write-in) was most favored, with "PCV" a close second.

*Second question: Name for Bucklin* (again, lumping all Bucklin variants
together for the purpose of advocacy)
The winner across all methods was "*Majority Choice Approval (MCA)*".

Third question: Name for Range
This was the old question of Range or Score. Cardinal methods picked Range;
ordinal ones picked Score. In this case, it is significant that some voters
were strongly strategic, rating all options at +10 or -10, while others were
much more neutral, in some cases with no ratings below 0.

SystemWinner(s)Instant Runoff Normalized RatingsRange Voting (RV)Virtual
Round RobinScore Voting (SV)Raw Rating Summation Range Voting (RV)ApprovalRange
Voting (RV)Instant Runoff VotingScore Voting (SV)


*Fourth question: Name for IRV/Hare*
The winner across all methods was "*IRV*".

*Fifth question: Best system* (this was included only for interest, as the
sample was not at all scientific)
The winner across all methods was *Range*. Counting the results by Approval,
it was tied with Approval and Condorcet, because all three systems got
nearly unanimously non-negative ratings. Interestingly, there was a
Condorcet tie for second place, with Approval >> Condorcet > Bucklin >
Approval. That may have been a statistical fluke, although it actually makes
some sense to me.

-----------------

In my opinion, these results are good. They're not my own top choices, but
for branding, I trust the wisdom of crowds far more than my own guesses. The
old Range/Score debate is not resolved, but I have already started to use
the new names in my advocacy, and they work. For Condorcet, I say "Round
Robin Voting" for a generic Condorcet-compliant system, but "pairwise
champion" for the CW. For Bucklin, I understand that "MCA" already has a
technical meaning as a specific Bucklin variant, but I think it can still be
used to advocate for the general class of (equality-allowed) Bucklin
systems.

I'd be interested to hear if people are willing to join me in starting this
usage when advocating for these systems with the general public. Honestly, I
think that it would increase our success. I know that I'm one of the few
strong advocates for MCA, but there are a number of people on these lists
who favor some form of RRV as the best system for practical implementation.

What say you?

Cheers,
JQ
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.electorama.com/pipermail/election-methods-electorama.com/attachments/20101014/191f3700/attachment-0002.htm>