[EM] [RangeVoting] Dartmouth uses Approval Voting to elect trustees

[Abd ul-Rahman Lomax]

At 06:09 PM 5/25/2007, Jan Kok wrote:
>A total of 18,186 voters, 28 percent of alumni, cast 32,941 votes
>using the "approval" method whereby voters can vote for as many of the
>candidates as they wish. Smith received 9,984 votes. Voting took place
>both online and via paper ballot April 1-May 15, and the process was
>managed by the Association of Alumni balloting committee. The other
>candidates who ran were Richard L. "Sandy" Alderson '69, the CEO of
>the San Diego Padres; Sherri C. Oberg '82, Tuck '86, the CEO and
>cofounder of Acusphere; and Ambassador John S. Wolf '70, the president
>of the Eisenhower Fellowships.

There are a number of aspects of this report worth noting. It would 
be nice to have the complete vote counts. I wonder how many people 
actually voted for more than one?

By the number for the winner, I would guess not many.

Does this mean that Approval is a waste? Not at all. Approval costs 
nothing to implement. You just count all the votes.

Okay, there could be a small cost, if people cast a lot of extra 
votes. Depending on the counting method, counting more votes costs a 
little more. But my expectation in *most* elections is that few 
people will cast extra votes. Those extra votes, however, allow 
substantially increased freedom of the voters.

The winner only received a quarter of the vote. I would turn around 
and submit this for ratification. I'd use FAAV (Fractional Approval 
Asset Voting) to create an electoral college which would have the 
power to ratify or otherwise resolve the election until a majority 
has supported one candidate, as a minimum. It would be better to get 
a higher vote of confidence.

It's entirely possible that the winner above, would handily pass a 
ratification vote. But it's also possible that he would not. Depends.

You don't have a democratic decision until the majority has expressed 
itself, unless the majority, perhaps, has preapproved -- and I mean 
immediately preapproved, the same electorate -- some other outcome as 
equivalent. The majority can delegate its decision-making power.

(A ballot could have on it the question, "Shall the candidate with 
the most votes in this election be declared the winner? Yes/No." If 
the majority answers Yes, then the candidate with the most votes is 
democratically elected. Frankly, though, I would generally vote No, 
unless the runoff process were so cumbersome as to be impossible. 
Which is where the problem would be, hence FAAV, which uses 
candidates as proxies representing the voter, making further process, 
including negotiation -- which is an aspect of deliberation -- 
feasible. It's a reasonable assumption that simplifies matters.)

Practically speaking, in this election, if my idea were taken up, the 
candidates who received votes would be asked to participate in an 
Approval election where they may recast all the votes they received. 
(If they received N votes, they may cast up to N votes for any given 
candidate). (There are alternate methods that could be used, but I 
will point out that an Approval election where the voter may cast 
multiple votes *is* a Range election; in this case, the Range is 
variable depending on the voter's level of trust among the general 
electorate, which I find fascinating as a possibility. *Whenever* the 
majority has not clearly accepted the outcome, further process should 
ensue or the process isn't fully democratic.

One more point. The size of this community, with over 18,000 voters, 
is larger than many small towns.

We should have a list of communities that use Approval Voting. It's a 
good selling point.