1. The Tea Party used a variation of range voting that did not utilize an abstain option. Instead, blanks were treated as a score of 0. What’s interesting, is that doing this allows a more familiar readout, as follows:

   Points Awarded; Expressed in Percentages of Total

   29.05% Dan Eichenbaum
   22.55% Kenny West
   17.29% Greg Newman
   16.28% Jeff Miller
   8.66% Ed Krause
   6.17% Jake Howard

   The group gave the usual range averages as well as shown in the link. The ballot used can be found here:

   http://ashevilleteaparty.wordpress.com/2010/03/06/results-of-atp-nc-11-candidates-debate-vote/

2. Thanks to Aaron for explaining how they got every voter to rate all the candidates. The answer is that they didn’t. This makes the published result potentially misleading.

   Dealing with blanks is an important problem in the design of a range voting system. Other voting methods have analogous problems.

   In general, range can work better in opinion polls such as this than it would in elections. That’s because the incentives to vote strategically are both different and much weaker. Range works best of all in contexts where the voters are impartial judges. Olympic events scored on a point system are one example.

3. “This makes the published result potentially misleading.”

   Specifically in what way?

   Non votes are tallied as zero.

4. “In general, range can work better in opinion polls such as this than it would in elections. That’s because the incentives to vote strategically are both different and much weaker.”

   I would have to disagree. First, any voting system works better when complete honesty is used, not just range. Second, (if it’s being insinuated that range is inappropriate in voting with elections) this is simply not the case. In analysis, range performed better than any other system using either honest or strategic voters (tied with approval for strategic). See here: http://rangevoting.org/UniqBest.html

   At worst, we have a regression to approval voting, which still performs better than the other systems when considering strategy. In an exit poll using range in the US 2004 election, numerous voters used more than just the ends of the range score (http://math.temple.edu/~wds/homepage/votdata.txt) Further, voters may wish to distinguish among candidates they like or dislike and express the degree to which they prefer one to another. Range is the only voting system that allows the option to express degree rather than just ranking.

5. Pingback: Cannondale Mountain Bikes-One Of The Biggest & Most Successful Manufacturers Of Mountain Bikes

6. Bob Richard,

   As I have mentioned to you several times, Score Voting works quite well when there are many strategic voters. This is actually one of its greatest strengths. It doesn’t degrade into plurality voting the way alternatives such as IRV tend to.

   The logic is pretty simple. Say I’m a tactical voter, and my sincere scores are:

   Krause = 10
   Howard = 9
   Eichenbaum = 7
   Miller = 6
   West = 4
   Newman = 0

   I would want to start by giving Eichenbaum a “10”, and West a “0” — since they appear to be the front-runners, and I want to strategically maximize the effect of my vote.

   But THEN, I would have no fear of supporting Howard and Krause — or anyone else I preferred to the front-runners. I might even give Howard and Krause both a 10, since that would just increase the odds that one of them wins instead of Eichenbaum or West.

   With IRV on the other hand, I would want to strategically put Eichenbaum in first place, which would prevent me from supporting the candidate I sincerely prefer to Eichenbaum. I have to throw Howard and Krause under the bus in this case. If most voters use that strategy, then IRV behaves almost identically to plurality voting.

   Here’s proof that most voters DO use that strategy in countries that have used IRV for a long time:
   http://scorevoting.net/AusAboveTheLine07.html

   And looking at the last mayoral election in Burlington VT, we can see exactly the kind of scenario I describe here, in which a bloc of voters who preferred Republican over Democrat over Progressive could have gotten the Democrat instead of the Progressive if they had strategically top-ranked the Democrat instead of the Republican. (And presumably they will wise up and vote more strategically if Burlington continues to use IRV.)
   http://scorevoting.net/Burlington.html

   Here’s further explanation of this issue:
   http://www.electology.org/later-no-harm

   The argument you’re alluding to is that voters will just “bullet vote” with Score Voting, by giving a maximum score to their favorite candidate, and a minimum score to all other candidates. But that’s clearly unrealistic, since that same argument says that plurality voting is virtually free from strategy (which IRV supporters themselves often argue is not the case).

   And actually, in this data set from the Asheville straw poll, there was a bit of strategic voting. But it was mostly REAL strategic voting, where you give multiple candidates the maximum score, not just your favorite.

   So Bob, I think you owe us a retraction here. Your claim is misinformed, and seems disingenuous.

7. That’s nonsense. IRV doesn’t degrade into plurality voting.

8. Tim (#3), unless the instructions to the voter said explicitly that blanks were going to be counted as zeroes, blanks mean “don’t know” or possibly “don’t care”. Zeroes mean “less desirable than all the candidates rated above zero”. These are not the same thing.

   My understanding is that the standard procedure is to exclude blanks from the results, using the average score rather than the total score to determine the winner. A minimum percentage of non-blank ratings (I think I’ve seen it called a “quorum”) is required. Candidates not rated by at least that percentage of the voters can’t win. Clay will no doubt correct this if it’s wrong.

9. Longer term —

   YES or NO on each party hack candidate (ABSOLUTE support)

   AND

   1, 2, 3, etc. on each candidate (RELATIVE support).

   Condorcet head to head math — with the YES/NO tiebreaker.

10. At greenchange.org, we’ve been using the range voting app at the following website to decide on topics for our “blog action days:”

    http://www.youtube.com/watch?v=vNc5o9TU0t0

    It’s not perfect, but it works for our purposes.

11. #8 Bob Richard:

    yes, we clearly indicated on the ballot that no score equals zero. you can view the ballot here:

    http://ashevilleteaparty.wordpress.com/2010/03/06/results-of-atp-nc-11-candidates-debate-vote/

    for various reasons, some of which i do not fully understand as i’m not mathematically gifted, we opted to not include a “No Opinion” option.

    but in general, we didn’t feel a quorum was required. we just wanted a simple result where all candidates were scored without any chance of confusing anyone.

    also, particularly in this straw poll, we felt strongly that a “No Opinion” option was entirely moot, as the voters had the opportunity to hear all the candidates.

    i do understand there are valid reasons to consider including a “No Opinion” option if this were an official election. however, without better understanding the weighting (or non-weighting) of the “No Opinion” vote with regard to the method advanced on RangeVoting.org, i cannot fully support it either.

    also, i might be mathematically off base, but it seems to me that if someone has no real opinion about a candidate one way or the other for whatever reason, perhaps scoring a “so-so” value is in order? (on a 0-9 scale, a 4 or 5?) i mean, if 0 is the worst score and 9 is the best, then isn’t a 4 or 5 “so-so”?

    of course, i understand that this is only really applicable if someone KNOWS about a candidate and thinks they are “average” or “so-so”. this doesn’t take into account for NOT knowing a candidate.

    on the other hand, perhaps NOT including “No Opinion” might encourage voters to GET ONE! 😎

    regardless, i’m very happy that there are fellow Americans learning more about, while also implementing, Score/Range Voting.

This site uses Akismet to reduce spam. Learn how your comment data is processed.