Minority Opinions

Not everyone can be mainstream, after all.

Rank Your Choice

with 2 comments

Having decided that we want to change the voting system, how should we go about it?  First, consider the two main effects that we want to diminish:  First, the choice between vote-splitting or compromising; and second, the mudslinging.  We probably can’t eliminate either one completely, but we can at least make things better.

Fortunately, there are mathematicians who have been working on precisely this problem, and they have come up with a variety of alternative voting systems, along with several ways to determine which ones are, in some sense, better than others.  Unfortunately, they have also proven that there is no such thing as a perfect voting system, so we’re stuck with a bunch of arguments about which one is best.  I have some opinions on some of them, but it turns out that the precise system doesn’t really matter to the voter.  Instead, there are two things that matter:  First, how to mark the ballot; and second, how to get the best result possible, given the likely votes of other voters.

The easiest way to mark a ballot is to simply mark your preferred choice.  Such ballots are also extremely easy to count, at least in theory.  Unfortunately, such ballots also yield only enough information for the type of first-past-the-post tally system that we’re already using.  For the better voting systems, we need to let each voter say more.

The next step up is to allow voters to mark any number of candidates.  Those used to the current system can vote just like they currently do, but the option is available to those with a good second choice.  For example, if you really want the Green Party candidate to win, but would accept a Democrat, and disapprove of the Republican, then you can mark the first two but not the last.  Given long enough, the Green Party might even collect enough votes to become a major contender.  At that point, an interesting conundrum raises its head:  If three candidates seem relatively close in polls, do you vote for both your first- and second-place choices to prevent the third from winning, or do you vote for just the first lest the second overtake it?  Where do you draw the line?

To resolve that dilemma, the ballot could allow voters to cast a certain number of votes, divided as you choose among up to that many candidates.  In the example above, if the voters were allowed three votes each, you could mark your favorite candidate twice, and your second choice once.  Such a system is currently being used in some places, particularly for elections with multiple winners.  For small numbers of votes, this can be presented as a grid of radio buttons or bubbles to fill in; larger values work better as a box for each candidate in which you write a number, ensuring that the total matches your allowed count.

At that point, it would probably be worth considering a full ranked system.  Using numbers to mark relative order feels slightly unnatural, but we tend to get used to it in grade school.  Even better, when we develop secure voting machines, would be a system that lets us drag candidates above and below each other.  The benefit of such a system is that it allows you to express your relative preference for even your third- and fourth-choice candidates, without worry that doing so will impair the chances of your favorite.  Depending on the tally system, however, you might have to rank all candidates, lest you give a slight advantage to the ones left unranked.

Combining the last two, and as the final stage of information you could possibly give the tally system, would be a way to express relative weights for each candidate.  For example, you might need to rate each one on a scale of 1 to 100.  That could get unwieldy, but even more unfortunate is the tendency, in practice, for some voters to use only the ends of the scale, while others space the candidates out a bit more.  Depending on the tally system, that could mean that certain voters have more weight to their preferences.  Fortunately, relative ranking information could be extracted, allowing any ranked ballot method to be used; if so, the relative weights are still available as a tie-breaker when necessary.

So, as a voter, would you be comfortable ranking your candidates?  If so, would you want the ability to say which preferences are strongest, or would it be just extra hassle?

For my part, I’m currently in favor of ranked ballots, with or without weighting.  The unresolved question is which tally method to use.  Ideally, it should be one that encourages me to vote honestly, by giving me the best result for doing so.


Written by eswald

10 May 2011 at 8:40 am

Posted in Politics

2 Responses

Subscribe to comments with RSS.

  1. The “mark any number of candidates” system is called Approval Voting and you characterize the strategic issue with it perfectly. But after that is where you lose me, implying that that strategic issue can be rectified. It turns out that the ideal system that encourages you to vote truthfully, while reflecting the will of the electorate, doesn’t exist. It’s logically impossible. Check out Arrow’s Impossibility Theorem, to start with. Of course, that’s a recipe for endless quibbling about how close to ideal we can get in practice.

    My view is that, in practice, no system really improves on Approval Voting in terms of the strategizing it encourages. For example, ranked voting. If the race is tight between A and B and you prefer A over B (but both both A and B are middle-of-the-pack in your eyes) then you should maybe disingenuously rank A first and B last, in hopes of giving A the biggest possible nudge over B. That’s more convoluted strategery than just picking a threshold for which candidates to approve of. And since Approval Voting is least different from the status quo, I think Approval Voting is the clear winner. (The status quo, by the way, is known as Plurality Voting and is uncontroversially Awful for myriad reasons, like you talked about in your previous post.) As you point out, people who don’t care about this don’t even have to do anything differently. Continue voting for your single favorite and everything still works fine. Even the winner criterion is unchanged: whoever has the most votes (“approvals”) wins. The only change is to stop throwing away ballots with multiple candidates checked.

    It’s really a shame that all the people who agree that Plurality Voting has got to go (which is everyone who gives it any thought) spend most of their energy debating the merits of their favorite alternative (approval, instant run-off, range voting, etc etc) rather than joining forces and getting us off of the worst-of-all-worlds that we currently use.

    In conclusion: can we please agree to switch to Approval Voting?

    Daniel Reeves

    18 May 2011 at 10:25 pm

    • I hadn’t considered approval voting from the minimal change angle. That certainly makes it a great candidate for actual adoption purposes. I’m not entirely convinced that it’s the best voting method overall, but I could be persuaded. Stack Exchange sites have certainly done well with a simple approval/disapproval system, for example.


      19 May 2011 at 3:48 pm

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s