Eating our own dog food - try out these polls
Jan Kok
the golden rule, etc. So, if we advocate inflicting some polling or
voting method on millions of people, we ought to be willing to try
them ourselves.
There are two Range Voting presidential polls on the MSNBC site, and a
third one coming up tomorrow. In these polls, you give each candidate
a "-", "0" or "+". After you submit your votes, you will see a report
on the percentage of -'s, 0's and +'s that each candidate received.
I urge everyone to take these polls seriously, i.e. check out the
candidates and vote as if you are voting in the Iowa Republican caucus
straw poll (the Democrats do things differently), or the national
nominating convention, or the general election. Then report your
experience here on this thread (or report to me and I will summarize).
After doing the Range poll, it should be easy to imagine how you would
vote with Approval Voting, Plurality Voting, and IRV. How would you
rate or rank the various methods, in terms of ease or difficulty of
voting, "expressivity" (were you able to "send a message" with your
vote), and confidence that the method would make a fair choice of
winner?
Regarding the Range Voting poll, was the -1,0,1 range about right, or
would you prefer a larger or smaller range, e.g. 0,1, or -2 to +2, or
0 to 10?
Please give these a polls a try! Here they are:
Poll in advance of the May 3 Republican debate (10 candidates):
http://www.msnbc.msn.com/id/18421356/
There will presumably be a new poll following the debate.
Poll following the Democratic debates (8 candidates):
http://www.msnbc.msn.com/id/18300340/
By the way, if you would like to know how others are voting, in order
to vote strategically, the results of the Republican poll are posted
at http://groups.yahoo.com/group/RangeVoting/message/4535 and someone
might post the results of the Democratic poll as well.
Vote early and often! :-)
Cheers,
- Jan
Abd ul-Rahman Lomax
>Regarding the Range Voting poll, was the -1,0,1 range about right, or
>would you prefer a larger or smaller range, e.g. 0,1, or -2 to +2, or
>0 to 10?
I really like the simplicity of -1, 0, +1. I assume that partial
abstentions (where a voter does not rate all candidates) are counted
as 0, and that the reporting was on all ballots cast. It might even
include totally blank ballots. I wish they'd describe the procedure somewhere!
-2, -1, 0, 1, 2 would be very nice.
Balanced Range (negative N to positive N) has the nice feature that
abstentions are equivalent to midrange votes. This system biases
relatively unknown candidates to midrange. I've contended on the
Range Voting list that this is better than more complex systems which
allow explicit abstentions and which then need a quorum rule to
prevent a write-in from winning from one max vote!
There were two candidates with positive net votes in the Dem poll I
looked at, after the debate. Edwards and Obama, with Obama having a
clear lead. Clinton was in third place, with one more percent of
negative than of positive. I find these polls far more informative
than "What's your favorite?"
Now, if the same kind of poll were taken for *all* candidates, what
would it look like?
Michael Ossipoff
Ahhh...imagining that irresistible "new car" smell?
Check out new cars at Yahoo! Autos.
Abd ul-Rahman Lomax
>Your link to the Democrat -1,0,1 poll actually only led to the
>Republian poll. If you or anyone knows where to vote on the -1,0,1
>poll among the Democrat candidates, then post a link to it, or post its URL.
I followed his link and it went to the Dem poll. But I voted in that
poll, maybe there is some anomaly due to this. They obviously place a
cookie, which prevents revoting. I think to see the poll results, you
must vote. They may have closed the voting.... don't know and can't
tell without taking more trouble than its worth.
Mr. Ossipoff should look again, is what I'd suggest to him.
Abd ul-Rahman Lomax
voting. I got the dem poll. Here is the URL again:
http://www.msnbc.msn.com/id/18300340/
I think it may be a Javascript page, so if Mr. Ossipoff is accessing
it through a library computer, there could be some problems.... Among
other things, cookies set by other users? Or do they work that way?
What I get on the other computer is not the results. It's the poll
itself. The defaults are 0 ratings. If I Submit without voting, I
think it increases the zero percentage for all candidates. Which is
neutral in net rating result, but does skew the results in a
different way. In this poll, zero ratings are either abstentions or
are explicit neutral ratings.
It's quite interesting to see a real Range poll. And quite
informative. When you look at Plurality polls, they create a totally
different picture. You can also interpret this poll as Approval. Same
result, which is no surprise. That is what is expected to happen most
of the time. The theoretical advantage of Range is marginal, not
fundamental, since Approval *is* a Range method (Mr. Ossipoff likes
to reserve "Range" for Range 3 or greater, but Range advocates almost
universally consider Range 2, Approval, a Range method, and are
largely united in favoring Approval implementations even where they
believe Range is better. The good is not necessarily the enemy of the
best if it is a step in the right direction.
And Approval also does not preclude moving from there to ranked
systems, including, if those making the decision want it -- I
wouldn't -- IRV. IRV designed to allow equal ratings is quite
feasible, and, even if this is somehow considered inferior, allowing
multiple rankings supplies some of the benefits of Approval; I just
dislike tossing ballots because the voter has the temerity to
conclude that two candidates are both acceptible in context. It is,
to me, fundamentally offensive, if there is a way of reasonably
interpreting a vote and using it without giving the voter some unfair
advantage, to toss it. And it has resulted, clearly, in some poor
election results.
We really should all unite behind Approval as a very reasonable,
cheap, biggest-bang-for-the-buck reform.
We think -- they aren't talking -- that the major IRV proponents are
really aiming at Proportional Representation, a reform that most of
us also support. The most prevalent form of PR uses Single
Transferable Vote, which is quite similar to IRV. But it's different.
Single-winner IRV has potential problems that don't crop up nearly as
badly with STV. And there are also non-ranked PR implementations that
are possible; there is reweighted Range Voting, there is Asset
Voting, and I think that there is an Approval method.
To my mind, these are all further reforms that can properly be
debated down the road, maybe even quite soon. But Approval should be
brought in now, and all it takes is to stop tossing overvotes, just
Count All the Votes!
Personally, it is my hope that all the major and reasonable reform
proposals be tried in at least a few places. Some of these have never
been tried, but look very promising. But Approval is a very simple
reform, already used for conflicting initiatives (if two initiatives
both pass, the one with the most Yes votes wins. That is Approval!),
and what is a multicandidate single-winner election but a series of
conflicting initiatives. Approval is very unlikely to cause harm. It
is very unlikely that it will *normally* shift results, the most
immediate effect that we consider salutary is that third party
supporters will be able to begin to vote honestly without harming
themselves. Major party supporters may also be able to indicate
support for third parties, thus helping the major parties to
understand the direction that they would like the party to move.
Approval does not require complicated strategy, does not require new
ballot forms and new equipment. All existing equipment can count it
without any trouble or fuss. (It has to be able to allow multiple
votes because of multiwinner elections). It actually should very
slightly lower vote processing costs, because there no longer need be
any special handling for overvotes. Just count all the votes.
This support for Approval should not be confused with an opinion that
Approval is ultimately the best voting method. Some experts think
that, to be sure. But supporting Approval is not being against IRV or
Range Voting or Condorcet methods. It is just recognizing how simple
and easy it is to improve voting by simple cutting out a few lines of
the election code, those that require discarding overvotes. I've
never been able to figure out why those lines are there in the first
place. In Robert's Rules, the mention of overvotes assumes that they
are mistakes. Which is circular. They are mistakes if they are
prohibited, and they are prohibited because they are mistakes.
At 02:11 PM 5/6/2007, Abd ul-Rahman Lomax wrote:
>At 12:45 PM 5/6/2007, Michael Ossipoff wrote:
>>Your link to the Democrat -1,0,1 poll actually only led to the
>>Republian poll. If you or anyone knows where to vote on the -1,0,1
>>poll among the Democrat candidates, then post a link to it, or post its URL.
>
Jan Kok
for voting in the MSNBC polls,
http://www.msnbc.msn.com/id/18421356/
takes me to the Republican poll,
http://www.msnbc.msn.com/id/18300340/
takes me to the Democratic poll.
In each case, below the big box at the top of the page that says "RATE
THE CANDIDATES", there is a smaller box in the center of the page that
also says "RATE THE CANDIDATES". Clicking that small box takes you to
the other-major-party poll.
By the way, Mike Ossipoff has been promoting better voting methods for
more than 18 years. He was my tutor and mentor regarding voting
methods for several years soon after I got interested in the subject.
Welcome, Mike!
Cheers,
- Jan