A few of us have been sneaking approval voting in wherever it makes sense. Familiarize people with the process.
One example is we use approval voting for prioritorizing our motions during membership meetings. Previously, our endorsement process was a major food fight, was way too dependent on the order which motions (from the floor) were recognized (unfair, yielding results not representative of the membership) and took way too long.
Now, using approval voting to pre-rank the motions, everyone sees the process is open, fair, accountable.
I think this would have a much more dramatic benefit toward US politics than the elimination of the electoral college, adding new US states, or fighting gerrymandering would.
Does Approval Voting really pass the "favourite betrayal criterion"?
There are scenarios where if you don't vote for both Good and Ideal, bad will win, and scenarios where if you do vote for both Good and Ideal, Good will win (over Ideal)
You obviously don't know which scenario you're in when you vote, so you don't know whether you should vote for both good+ideal or just ideal - both options can cause a worse result
I suppose the idea is that you use polling to decide which scenario is more likely and vote accordingly
Yes, approval voting leaves the door open for some tactical voting.
So if candidate #1 would win a simple majority (50.01% of voters), but is absolutely hated by everyone else, but #2 is "acceptable" or second best to 100% of the voters, should #1 win?
We are used to the binary "simple majority takes all" in elections, but maybe the compromise "everybody is pretty OK with this one" _should_ win.
If it were a group of friends going to a restaurant, or playing a game, #2 sounds like a much better option.
I'm not totally convinced that particular compromise should win. The issue I have with it is whether political candidates should have to please everyone. Using that strategy would likely lead to a lot of charismatic but ineffective people getting elected. None of big issues would get any sort of resolution, because the candidates that the most people are ok with are the ones with the fewest controversial stances.
But, the alternative is the dysfunctional partisan mess we have in the US. If parties split government, nothing gets done (save a lot of finger-pointing and chest-thumping). If a single party controls government, "progress" is made, but that progress is likely to be extreme and to the displeasure of a massive portion of the electorate.
Good winning over Ideal is a desired outcome when Ideal doesn't have the votes to win. That's one of the main attractions of non-plurality voting. The point of approval voting is not to rank candidates as Ideal > Good > Bad, that's what IRV is. Approval voting says, "These candidates are equally acceptable to me, so I'll be happy if any of them win over the candidates I didn't vote for."
Sure, but what the OP is saying is that in practice there's a massive tactical decision over whether to vote for "good" or not in a tight race, because "these candidates are equally acceptable to me" is a relatively rare situation
I don't think that's a rare situation at all. There's a large population out there that doesn't identify with R/D (in US politics), but does identify more closely with a 3rd party. But they never vote for them because "tactics". That doesn't mean those people wouldn't find R or D acceptable, it just means the're ideologically closer to a 3rd party. What AV is trying to get at is the situation where the most people are accepting of the outcome.
If a candidate is not acceptable, you don't vote for them. The situation you're describing is not one of selecting from Ideal, Good, Bad; it's the situation of selecting from Ideal, Somewhat Bad, Bad. If you're unwilling to vote for a candidate, they are by definition not "Good" to you. So you would just vote for the Ideal candidate in that situation.
In the worse case (everyone just votes for a single candidate), AV simply acts as FPTP. In the average and best case, the candidate that is acceptable to the most voters wins.
Look at it from the view of the US's extreme polarization: if I'm strongly for A, and you're strongly against A, nothing will be done if A wins. But I'm strongly for A, and we're both "meh" on B, then B might not get everything done that either of us wants, but at least the process won't completely break down.
But assuming most of these people consider R and D acceptable but actively prefer the third party (if they consider them equally acceptable FPTP isn't really depriving them of their ideal electoral choice), AV is inferior to most other non FPTP votes like IRV, which accurately represent their preference for the third party first, and the Rs and Ds only if that party can't win.
AV has a pretty significant "if I tick the good candidate box, am I impairing my ideal candidate's chances of winning?" problem.
IRV actually has the opposite problem: "If I vote for my ideal candidate, am I impairing the good candidate's chance of winning?" For instance, if it were Nader/Gore/Bush with 26%/25%/49%, and (hypothetically) all Nader voters preferred Gore over Bush but not all Gore voters preferred Nader over Bush, then voting for Nader instead of dishonestly choosing the "lesser evil" Gore would end up electing Bush.
No voting method can entirely escape both of these problems. But newer methods like STAR voting balance them against each other and manage to minimize both.
> No voting method can entirely escape both of these problems. But newer methods like STAR voting balance them against each other and manage to minimize both.
First you need to demonstrate that it's a problem in reality, there are very few times in history where it has been a real issue. Alternative to IRV are usually much more complex which is a terrible tradeoff for an almost entirely imaginary problem. Star voting has the same problem as star rating systems, almost everyone will score either a 5 or a 1.
Previous comments complained that approval is bad because you cannot express detailed preferences, but you say no one wants to anyway? So then approval is ok?
The "favorite betrayal" criterion is defined mathematically. In your scenario, as long as you approve Ideal, you are not betraying your favorite, even if you also approve Good. The only betrayal would be if you didn't approve Ideal. And in approval voting, there is no way others could be voting so that betrayal would be a good idea.
You are talking about a form of strategic voting — whether or not you approve Good. But it isn't favorite betrayal. In fact, it isn't betrayal at any level. All rational approval strategy (assuming a convex distribution over others' votes) is "semi-honest" in that you never approve a less-liked candidate but not a more-liked one. The only strategic choice is where to set your threshold, then honestly approve all candidates above it.
According to the Gibbard-Satterthwaite theorem, any voting method has some form of strategy, so in that sense the best you can possibly do is ensure that all strategy is semi-honest. Approval is optimal in this sense.
Article makes some good arguments for approval voting, but the "Ballot Spoilage" argument is a bit disingenuous.
Ballots can't be spoiled in approval voting because any set of circles can be filled in. However it's a lot easier to corrupt a ballot undetected. Just fill in more bubbles for your candidate.
Perhaps then each candidate should be marked with either a "0" or a "+". It is difficult to change either of these into the other without signs of tampering.
That's fixable by using voting machines to print the ballots. E.g. fill in the circles but also print the number of filled-in circles for each contest. Or, print only the selected candidate names.
I'm not sure I understand, nor how this would combat the problem that I can accidentally spoil my ballot by miscounting the number of choices I've made. An easy mistake to make - even easier if you have a condition like dyslexia.
It's also misleading because the simplicity of the appearance ballot paper disguises the complexity of the actual choice of how many boxes to tick in front of you.
(Am I better picking my favourite candidate only or both the candidates I don't find entirely unacceptable in a tight three way race?)
> the actual choice of how many boxes to tick in front of you
With approval voting, you should consider each candidate independently: "Am I okay with this person winning the election?" The number of circles you fill in doesn't matter.
But the number of circles filled in by large segments of the electorate matter hugely, and most people's level of "am I OK with this person winning the election?" varies hugely, especially when there's a candidate they'd be really unhappy with on the ballot paper.
This is definitely a better way to support IRV, except that it depends heavily on people writing numbers that are recognizable by either computers or people when scoring.
I had the same question, but I think they meant that you can fill in the bubbles for your candidate of choice on other people's ballots without anybody being able to detect it, and without causing the ballot to be discarded.
It's possible that it might be easier to tamper with these than other ballots, but there must already be countermeasures in place to deal with these types of issues in a large scale election.
In Australia, each candidate is able to send a scrutineer to each voting and counting location, and it's not exactly done in privacy where tampering would go undetected: https://www.youtube.com/watch?v=XevjASEgoK8
Unintentional spoilage and intentional tampering are two different things. There's nothing disingenuous about mentioning one of them but not the other.
It's true that, depending on the implementation, approval voting can be subject to tampering. But the key point there is "depending on the implementation". There are any number of anti-tampering measures that are compatible with approval. Other comments here have suggested some of them (such as using "0" vs "+" as marks). Others might rely on technological fixes (such as immediate ballot scans by multiple independent non-networked devices) or social ones (multiple witnesses during all stages of ballot collection and counting).
It may not be possible to be 100% certain of preventing every possibility of ballot-counting fraud, but it seems that >99.9% is possible, and that's basically good enough. In well-administered elections, spoilage is a much bigger problem. Thus, it's reasonable to focus on spoilage.
As a libertarian anything that increases the ability of a 3rd party go gain traction interest me but something smells very off about the recent uptick in discussion of alternatives. If these changes were actually beneficial to the underdog in the general case you'd see them being pushed by the left or the right on a state by state basis depending on the political make up of any given state. The parties aren't run by idiots. They've crunched the numbers. They know whether the alternatives would benefit them or not. Likewise it makes me very uneasy that we mostly only see people who align with one party pushing for alternate systems.
I'm not opposed to alternate voting systems but I'd like to see one actually increase the power of the underdog(s) in a one party state (the only kind of state where the powers that be feel secure enough to make this kind of change) in the US before I support one. I don't want something that sounds good but furthers the status quo. Political diversity is one of the US's strengths and anything that furthers the status quo is bad for that.
And political scientists! (No-one listens to us, of course).
More seriously, the case for standard First-Past-The-Post voting is really bad in terms of its decision-making properties. It's really only as dominant as it is, where it is, because of its historical position. Exactly what you might replace it with is a trickier question, with multiple plausible candidates, all better than FPTP and all different from each other. I lean towards Single Transferable Vote, and its special case AV/IRV in single member elections, but reasonable people can disagree.
Intriguingly, the British Parliament has just effectively used approval voting to try to break its Brexit deadlock. It failed in this case largely because many members voted highly tactically for their 'ideal' option only, not for what they could live with. AV/IRV would have been a better choice, I think, if a compromise winner was actually wanted.
The main benefit of FPTP, for third parties, is that they can hijack elections by acting as "anti-tiebreaker" (Threatening to turn their nearest ally into a lose by joining the election, unless that nearest ally adopts their platform). This is why third parties generally (and openly) advocate for FPTP.
Of course, the simple and well-tested solution is to reduce the power of the Executive, move power to Parliament, and assign power proportionally to parties.
> The main benefit of FPTP, for third parties, is that they can hijack elections by acting as "anti-tiebreaker" (Threatening to turn their nearest ally into a lose by joining the election, unless that nearest ally adopts their platform). This is why third parties generally (and openly) advocate for FPTP.
This isn’t true in the U.K., the only country where a third party recently tried electoral reform in a country with FPTP. The Liberal Democrats wanted to get rid of FPTP.
True, but then the Lib Dems were succeeded by UKIP as the third party and the electoral threat they posed to the Conservative Party was a significant factor in putting moves to leave the EU onto the Conservative Party's agenda despite its leadership's reluctance.
Of course, without skewed vote shares due to FPTP and with a limited choice of realistic coalition partners for a future government, that third party can achieve exactly the same thing through coalition bargaining anyway.
I'm also libertarian (leaning, in the sense I have voted for libertarian candidates in multiple elections at multiple levels, but not most of the time) but do research on this. I am not sure what to say to allay your concerns, except to say that mathematically/statistically speaking approval voting is what I would implement after a rating system, which I think is the best but also maybe the most cumbersome.
I personally don't really care about the political impetus behind changing the election format (not to say I don't care about politics, just that in the case of alternative voting systems the standard US format is almost the worst you could do). Anything that moves the US toward something better would be an improvement.
My one concern is that for whatever reason ranking systems seem to be the thing being pushed most, when approval or rating systems would probably be better.
Changing to an approval system would probably be the easiest way to increase 3rd party viability at the moment. Which of those 3rd parties would benefit is a different issue, but approval systems allow people to diversify their expressed preferences, so at the very least it wouldn't be worse for 3rd party candidates.
The basic issue is that plurality voting is a very poor measure of preference. Imagine if you were asked to measure height of 5 people in the room, 3 of whom are over 6 feet 10 inches tall. You could report how tall they are; this is most similar to a rating system. You could rank order their heights; this preserves rank ordering information but not anything about absolute difference, so the two people who are 7'2" and 7'1" look as different as the people who are 7'2" and 5'2". You could also ask people to just mark which are taller than 6'10", which at least acknowledges that more than one person could be taller than 6'10", and preserves some info about absolute height. Or, you could do as is standard, and just indicate who is tallest, which doesn't preserve much information at all about height.
It's a bit silly to argue that candidates take up too much space on a IRV ballot when it would suffice to duplicate only the box and not their name, as in
1st 2nd 3rd choice
John Adams [ ] [ ] [ ]
Ben Franklin [ ] [ ] [ ]
Tom Jefferson [ ] [ ] [ ]
Betsy Ross [ ] [ ] [ ]
>Or just put in a number in a single column boxes like in Australia or Ireland.
What do you do when some moron voter can't count and reuses the same number, or skips a number? This kind of voting scheme seems like it would much too complicated and confusing to the average American voter.
Throw the ballot out. People mis-fill ballots now. People will always mis-fill ballots. The column of numbers thing seems extremely intuitive and wouldn't baffle a significant number of American voters; certainly it wouldn't lead to hanging-chad type situations.
IRV, or ranked choice voting, can be practical even with large numbers of candidates. Ballot designs can be made that keep the voting manageable.
More importantly, ranking multiple people gets rid of the important problem first past the post has of splitting the vote with large numbers of candidates.
In Australia we have tens of MP candidates in most electorates as well as several dozen senate party candidates with hundreds of individual senator candidates.
IRV works perfectly fine here, though in the case of the senate does result in quite wide ballot papers. But if you only have a dozen candidates it's very simple -- you have a list of names and the voter specifies what preference they have (1st, 2nd, 3rd) in the box next to their name.
actually India mandates EVM - the elections in India are the biggest democratic event in the history of mankind. Its hard to to carry ballot papers over camel in the desert (that's a real thing).
Approval voting is far more suitable for straight scrolling screens.
Further, my jurisdiction's general election has 30+ contests. Ballot real estate is already scarce.
One mitigate is to split our plus-sized ballots into separate federal, state, and local, but no one is considering doing that. The primary argument against is ballot process (central count for vote-by-mail) is already complicated enough.
It seems to me that Approval Voting might have the unintended negative consequence of electing a joke candidate, since you do not rank your choices nor are you limited to how many you can choose. Unfortunately, many voters would not give a second thought to ticking "Mickey Mouse" or similar in addition to their main choice and alternates, since there is no tradeoff involved. Under IRV or Ranked Choice, few voters would likely rank the joke candidate high.
But there is a downside, by your own logic. If you vote for a candidate that you disapprove of (don’t want to elect), you run the risk of increasing their support.
Surely this disenfranchises those most willing to compromise, who are probably going to end up being those who support small parties?
Let's say you a Right and Left party, and breakout Far Left which is actually very popular but has previously lacked electoral support due to its 'outsider' status.
People's favourites might look like this:
Right: 30%
Left: 30%
Far left: 40%
However, 75% of Far Left favourites actually vote for Left too, because they prefer them to right. This ends up with the following result:
Right: 30%
Left: 60%
Far left: 40%
So we are left with 'establishment' mediocrity instead of the most popular result.
Is this OK because a majority of people don't actually want 'far left'? Maybe. Intuitively though, I think taking preferences away will freak people out, particularly in situations where there might be different levels - what if there was a Far Right party that Far Left voters really don't want in - would some end up voting for Right to avoid that happening? They might win despite having 70% unpopularity!
I'd prefer difficult counting processes over a blunting of people's democratic will.
"Left" and "Right" are nearly meaningless when people can have stances on an issue-by-issue basis. As someone who has important disagreements with both big US parties, some would call me a moderate, and I feel disenfranchised by the current system, which doesn't let me pick the person whom I agree with most.
They're merely labels. You can imagine 'far left' as 'centrist'. I was going to name them A, B and Z but wanted to get across that Far Left voters might prefer to compromise on Left than Right, but might prefer Right to Far Right. Clearly this is going to be a more nuanced decision in real life.
> So we are left with 'establishment' mediocrity instead of the most popular result.
In this case, Left is the most democratic winner of the race, also known as the Condorcet winner (assuming Right supporters would support Left over Far Left in a two-person race between them). However, not all elections have a Condorcet winner. Also however, approval voting doesn't guarantee the Condorcet winner wins the race (if one exists), but it's an improvement over FPTP.
Arrow’s Impossibility Theorem, as I understand it, applies quite broadly to voting systems. To put it simply, you can’t avoid it. You just decide what trade offs you want to live with.
Arrow's theorem applies only to ordinal voting methods. Approval is not actually subject to Arrow's theorem; it's not even well-defined whether it passes the criteria.
However, there are various voting impossibility theorems inspired by Arrow which do apply to approval. Most notably, Gibbard-Satterthwaite.
No voting system is perfect but Approval Voting is my favorite. Still, most schemes that people have put forward are far better than the First Past the Post and there's been some momentum for IRV in Mass, where I live, so I've been helping with canvassing and it looks like our efforts might be paying off.
Sortition (that is, random selection) is a system for roughly-proportional representation; that is, for choosing multiple winners for a legislature so as to represent a population.
Approval voting is a way of picking a single winner.
They solve entirely separate problems.
(Sortition is nearly optimal if your only goal is to match the distribution of voters by a subset of equally-weighted representatives. Technically, it's possible to improve on it slightly by allowing voters to cluster themselves into a hierarchical structure of parties and factions, then performing sortition under the constraint that each party and/or faction must come within 1 seat of its correct proportional representation. This kind of technique is called "reduced-variance sampling" when used in sequential monte carlo sampling.
But sortition, or party-clustered sortition, pays no attention to candidate quality. Most people would agree that if there are two candidates that hold the same ideology, but the only difference is that one of the two has a medical problem that only allows them to work for 1 hour a week, the other one would make a better representative. Sortition cannot be fixed to take this kind of factor into account, which is a big part of why many people prefer more traditional voting methods.)
One can most certainly pick even a single winner at random.
Equal chance to represent, there is no fairer way than that.
Voting inherently leads to a money/campaign/popularity contest (and parties, more often than not 2), it is very good at keeping the status quo by feeding back to itself.
I'm not aware of any large scale studies of election method preferences, but keep in mind 1. people are used to voting 2. even if they are aware fairer methods exist their self interest is to keep voting if they think they will benefit from the resulting unfair system.
What the article misses is that there are alternatives better than BOTH approval or IRV.
I will always find "Approval" to be unsatisfying because I want to say I like A more than I like B.
I will always find IRV to be dumb because it has bad outcomes and for no additional cost or complexity those ranked ballots could have been counted by a Condorcet method that results in better democracy.
https://bolson.org/irv/
>because it has bad outcomes and for no additional cost or complexity those ranked ballots could have been counted by a Condorcet method that results in better democracy.
Can you show how the differences in outcomes is better using the Condorcet method?
Yes, there was a real world election run using IRV that got the wrong answer. Fortunately the full (anonymized) vote set was released and we can see how it would have gone using other election methods.
https://bolson.org/irv/
I've also done simulations of millions of elections with various election algorithms and there are trends about what works and what doesn't:
https://bolson.org/voting/twographs.html
The behavior graphs seem to use a linear "happiness" metric to judge the quality of outcome. I question this choice. It is not intuitive to me that some unit of happiness should cancel out a unit of misery when it comes to determining election outcomes.
The IRV page extrapolates the majority principle for two choices to more than two choices, but its not obvious that this is correct. There are arguments that need to be made before we can start comparing the quality of outcomes between voting methods.
You can convey this by only submitting a single choice on your ballot (a powerful tactic in AV). Conversely, you can vote everyone BUT the one you wish to harm. And by default you can vote NOTA. While it may be that Approval Voting is not the "perfect" answer, it is far better than most and comes with the added benefit of easy adoption/implementation and simple explanation.
>You can convey this by only submitting a single choice on your ballot (a powerful tactic in AV)
Yeah, but then you run the risk of a Trump getting elected in an election between Trump, Hillary, and Bernie, where you really want Bernie, and you really don't like Hillary at all, but you'll still take her any day over the utter disaster that is Trump.
This is why AV isn't that great: all they have to do is put up a truly revolting candidate to scare voters into voting for the not-as-revolting candidate, and she gets elected because there's no way for voters to express a clear preference.
While I would certainly be more happy in a world with either IRV or Condorcet method than FPTP, I think this argument against IRV is too simplistic.
The situation is candidates D, R with the largest numbers of voters who prefer them first, and consensus candidate I with smaller first preference than D, R but lots of 2nd preference. Under any Condorcet method, the D and R voters can hurt their candidate by ranking I as second (later-no-harm criterion). This leads to a negative partisanship spiral, say, as D/R voters feel like they must only rank 1 candidate in order to counter the threat from the other party doing the same, and thus the consensus candidate I, which would be the Condorcet winner under voter's true preferences, does not become the winner under the actual votes. IRV does a better job of inducing everyone to rank choices.
The Later-no-harm criterion is basically a nonsensical idea invented to make IRV look better than it is. See a Princeton math PhD dissect it here.
http://scorevoting.net/LNH
tl;dr for non-experts (I co-founded the Center for Election Science and have studied voting methods since 2006).
The idea behind LNH is that if I vote for the Green in 1st place, it's "safe" to then rank the Democrat 2nd, because that can't cause the Democrat to defeat the Green. See this explained visually in this video by Andy Jennings, another CES co-founder who did his math PhD thesis on voting methods.
The Later-no-harm criterion is highly misleading, because "safe" here means safe for the CANDIDATE. But tactical voting is about what's good for the VOTER, not the candidate. With IRV, the Green voter's best strategy is to actually rank the Democrat in 1st place. So the whole notion that he can safely rank the Democrat #2 after he's ranked the Green #1 starts with a false premise. If he's rational, he WILL NOT rank the Green #1. It's the same reasoning that Democratic primary voters have about far left candidates. Even if you win the nomination and advance to the general, you're so far left that you'll probably lose against the Republican, by not capturing enough of the center of the electorate.
Approval Voting actually makes it safe to ALWAYS vote your favorite. And, in general, to vote for everyone you prefer to your favorite of the frontrunners.
Doesn't later-no-harm only fail on condorcet methods if there is no Condorcet winner?
From what I've seen of most Condorcet methods, the best way to tactically vote differs if there is or is not a Condorcet winner, and often become a major mistake in the other circumstance.
Basically it means unless you know what the results will be tactical voting will have a rather good chance of backfiring on you, making it not worthwhile, which seems like a good thing to me.
Total information throughput for a voting system is a function of THREE things:
1. Amount of information you can express on a ballot.
2. Accuracy of that information (which is distorted if the system gives you an incentive to dishonestly/tactically vote) .
3. Tabulation efficiency — how effectively does the system utilize the info on the ballots (for example there are numerous ranked voting systems from IRV, to Borda, to the entire Condorcet family; and IRV is extremely lossy) .
When you factor these all together, Approval Voting does surprisingly well, especially because it's so good at #2 and #3. See this graph of performance as measured by Bayesian regret.
http://scorevoting.net/BayRegsFig.html
Essentially what's happening is that the people who's approval threshold sits between candidates A and B (i.e. those people who only approve one of them) will, in a statistical sense, have about the same A-vs-B preference as the people who voted for both or neither of them. So the aggregate effect doesn't really suffer from your inability to express a ranked preference for all candidates. Voting is statistical.
To really drive that home, note that you could switch to a better voting method and STOP VOTING, and STILL get better election outcomes. The tiny impact your vote has on the process is statistically negligible. The properties of the voting system matter vastly more.
"Approval Voting does surprisingly well" and I agree that it's a mathematically valid and efficient solution that does _surprisingly_ well, but other systems do better (Cordorcet, etc).
Also in the messy human emotional sense, it's still _unsatisfying_ that I might Approval vote for Bernie and Hillary but I still want to say I like one better, and I still think that difference matters.
The article focuses on approval versus IRV, and the advantages it gives for approval are valid.
As you know, there are various other methods (Condorcet, STAR, 3-2-1, etc.) that keep most of approval's advantages, but sacrifice simplicity for better expressiveness. It would be as unreasonable for the advocates of such methods to try to undermine real-world campaigns for approval, as it would be for approval advocates to undermine those other methods.
Approval Voting is the best balance between fairness and simplicity.
I switched from IRV (RCV) to Approval Voting once I better understood the challenges of election administration.
That said, FPTP is just about the worst. For election administration, it's the most brittle, where the margin of error too often exceeds the margin for winning. So I'd happily support IRV, RCV, score voting, approval voting as a replacement.
I appreciate your opinion, but I want to point out that our definitions of fairness and simplicity may vary. Much of scholarly work around elections systems strives to ground itself on particular formal criteria.
I have focused on election integrity. For me, arguing the relative merits ranked choice voting strategies is (unforgivable) bikeshedding and ultimately results in no action. Like getting stuck arguing 4.5x vs 4.75x better than FPTP while the world burns.
For election integrity, other real world concerns also must be factored. Like voter education, verifying the hardware & software, ease of tabulation & auditing, feasibility of doing a manual recount, etc.
In conclusion, Approval Voting is almost as fair as the ideal Score Voting but much easier to implement.
It can't be mentioned enough that Approval Voting works with existing voting machines. One of the unfortunate difficulties with IRV is that it often requires a wholesale replacement of the voting infrastructure.
The Center for Election Science is a big supporter of IRV as a better system than our current "Choose One" voting method, but we've found that Approval Voting is both more descriptive and less prone to strategy, as well as being significantly easier to implement and use.
NB: I'm the Chair of the Board for The Center for Election Science. AMA.
95 comments
[ 4.6 ms ] story [ 184 ms ] threadA few of us have been sneaking approval voting in wherever it makes sense. Familiarize people with the process.
One example is we use approval voting for prioritorizing our motions during membership meetings. Previously, our endorsement process was a major food fight, was way too dependent on the order which motions (from the floor) were recognized (unfair, yielding results not representative of the membership) and took way too long.
Now, using approval voting to pre-rank the motions, everyone sees the process is open, fair, accountable.
There are scenarios where if you don't vote for both Good and Ideal, bad will win, and scenarios where if you do vote for both Good and Ideal, Good will win (over Ideal)
You obviously don't know which scenario you're in when you vote, so you don't know whether you should vote for both good+ideal or just ideal - both options can cause a worse result
I suppose the idea is that you use polling to decide which scenario is more likely and vote accordingly
So if candidate #1 would win a simple majority (50.01% of voters), but is absolutely hated by everyone else, but #2 is "acceptable" or second best to 100% of the voters, should #1 win?
We are used to the binary "simple majority takes all" in elections, but maybe the compromise "everybody is pretty OK with this one" _should_ win.
If it were a group of friends going to a restaurant, or playing a game, #2 sounds like a much better option.
If a candidate is not acceptable, you don't vote for them. The situation you're describing is not one of selecting from Ideal, Good, Bad; it's the situation of selecting from Ideal, Somewhat Bad, Bad. If you're unwilling to vote for a candidate, they are by definition not "Good" to you. So you would just vote for the Ideal candidate in that situation.
In the worse case (everyone just votes for a single candidate), AV simply acts as FPTP. In the average and best case, the candidate that is acceptable to the most voters wins.
Look at it from the view of the US's extreme polarization: if I'm strongly for A, and you're strongly against A, nothing will be done if A wins. But I'm strongly for A, and we're both "meh" on B, then B might not get everything done that either of us wants, but at least the process won't completely break down.
AV has a pretty significant "if I tick the good candidate box, am I impairing my ideal candidate's chances of winning?" problem.
No voting method can entirely escape both of these problems. But newer methods like STAR voting balance them against each other and manage to minimize both.
First you need to demonstrate that it's a problem in reality, there are very few times in history where it has been a real issue. Alternative to IRV are usually much more complex which is a terrible tradeoff for an almost entirely imaginary problem. Star voting has the same problem as star rating systems, almost everyone will score either a 5 or a 1.
Then it will be similar to approval voting.
Previous comments complained that approval is bad because you cannot express detailed preferences, but you say no one wants to anyway? So then approval is ok?
You are talking about a form of strategic voting — whether or not you approve Good. But it isn't favorite betrayal. In fact, it isn't betrayal at any level. All rational approval strategy (assuming a convex distribution over others' votes) is "semi-honest" in that you never approve a less-liked candidate but not a more-liked one. The only strategic choice is where to set your threshold, then honestly approve all candidates above it.
According to the Gibbard-Satterthwaite theorem, any voting method has some form of strategy, so in that sense the best you can possibly do is ensure that all strategy is semi-honest. Approval is optimal in this sense.
Ballots can't be spoiled in approval voting because any set of circles can be filled in. However it's a lot easier to corrupt a ballot undetected. Just fill in more bubbles for your candidate.
They seem to have tried to intentionally tried to create a much more complicated ballot paper for IRV
Here's a sample of an Australian IRV voting form:
https://upload.wikimedia.org/wikipedia/commons/thumb/4/48/20...
With approval voting, you should consider each candidate independently: "Am I okay with this person winning the election?" The number of circles you fill in doesn't matter.
How can someone “fill in more bubbles for (their) candidate”?
In Australia, each candidate is able to send a scrutineer to each voting and counting location, and it's not exactly done in privacy where tampering would go undetected: https://www.youtube.com/watch?v=XevjASEgoK8
It's true that, depending on the implementation, approval voting can be subject to tampering. But the key point there is "depending on the implementation". There are any number of anti-tampering measures that are compatible with approval. Other comments here have suggested some of them (such as using "0" vs "+" as marks). Others might rely on technological fixes (such as immediate ballot scans by multiple independent non-networked devices) or social ones (multiple witnesses during all stages of ballot collection and counting).
It may not be possible to be 100% certain of preventing every possibility of ballot-counting fraud, but it seems that >99.9% is possible, and that's basically good enough. In well-administered elections, spoilage is a much bigger problem. Thus, it's reasonable to focus on spoilage.
I'm not opposed to alternate voting systems but I'd like to see one actually increase the power of the underdog(s) in a one party state (the only kind of state where the powers that be feel secure enough to make this kind of change) in the US before I support one. I don't want something that sounds good but furthers the status quo. Political diversity is one of the US's strengths and anything that furthers the status quo is bad for that.
Edit: why is this opinion unacceptable?
And political scientists! (No-one listens to us, of course).
More seriously, the case for standard First-Past-The-Post voting is really bad in terms of its decision-making properties. It's really only as dominant as it is, where it is, because of its historical position. Exactly what you might replace it with is a trickier question, with multiple plausible candidates, all better than FPTP and all different from each other. I lean towards Single Transferable Vote, and its special case AV/IRV in single member elections, but reasonable people can disagree.
Intriguingly, the British Parliament has just effectively used approval voting to try to break its Brexit deadlock. It failed in this case largely because many members voted highly tactically for their 'ideal' option only, not for what they could live with. AV/IRV would have been a better choice, I think, if a compromise winner was actually wanted.
Of course, the simple and well-tested solution is to reduce the power of the Executive, move power to Parliament, and assign power proportionally to parties.
This isn’t true in the U.K., the only country where a third party recently tried electoral reform in a country with FPTP. The Liberal Democrats wanted to get rid of FPTP.
Of course, without skewed vote shares due to FPTP and with a limited choice of realistic coalition partners for a future government, that third party can achieve exactly the same thing through coalition bargaining anyway.
Though there is also some precedent in Canada with previous Conservative party mergers.
I personally don't really care about the political impetus behind changing the election format (not to say I don't care about politics, just that in the case of alternative voting systems the standard US format is almost the worst you could do). Anything that moves the US toward something better would be an improvement.
My one concern is that for whatever reason ranking systems seem to be the thing being pushed most, when approval or rating systems would probably be better.
Changing to an approval system would probably be the easiest way to increase 3rd party viability at the moment. Which of those 3rd parties would benefit is a different issue, but approval systems allow people to diversify their expressed preferences, so at the very least it wouldn't be worse for 3rd party candidates.
The basic issue is that plurality voting is a very poor measure of preference. Imagine if you were asked to measure height of 5 people in the room, 3 of whom are over 6 feet 10 inches tall. You could report how tall they are; this is most similar to a rating system. You could rank order their heights; this preserves rank ordering information but not anything about absolute difference, so the two people who are 7'2" and 7'1" look as different as the people who are 7'2" and 5'2". You could also ask people to just mark which are taller than 6'10", which at least acknowledges that more than one person could be taller than 6'10", and preserves some info about absolute height. Or, you could do as is standard, and just indicate who is tallest, which doesn't preserve much information at all about height.
That combo will reform our duopoly. Plus other benefits, like improve campaign civility and reduce polarization.
(FWIW, I upvoted you for your end goal.)
What do you do when some moron voter can't count and reuses the same number, or skips a number? This kind of voting scheme seems like it would much too complicated and confusing to the average American voter.
Skipping is no problem. If you want, it's fairly trivial to modify IRV to allow equal preferences.
More importantly, ranking multiple people gets rid of the important problem first past the post has of splitting the vote with large numbers of candidates.
IRV works perfectly fine here, though in the case of the senate does result in quite wide ballot papers. But if you only have a dozen candidates it's very simple -- you have a list of names and the voter specifies what preference they have (1st, 2nd, 3rd) in the box next to their name.
Approval voting is far more suitable for straight scrolling screens.
One mitigate is to split our plus-sized ballots into separate federal, state, and local, but no one is considering doing that. The primary argument against is ballot process (central count for vote-by-mail) is already complicated enough.
I think we've already crossed that particular bridge.
Let's say you a Right and Left party, and breakout Far Left which is actually very popular but has previously lacked electoral support due to its 'outsider' status.
People's favourites might look like this:
Right: 30% Left: 30% Far left: 40%
However, 75% of Far Left favourites actually vote for Left too, because they prefer them to right. This ends up with the following result:
Right: 30% Left: 60% Far left: 40%
So we are left with 'establishment' mediocrity instead of the most popular result.
Is this OK because a majority of people don't actually want 'far left'? Maybe. Intuitively though, I think taking preferences away will freak people out, particularly in situations where there might be different levels - what if there was a Far Right party that Far Left voters really don't want in - would some end up voting for Right to avoid that happening? They might win despite having 70% unpopularity!
I'd prefer difficult counting processes over a blunting of people's democratic will.
Nothing in your example means that the "Left" has to be "establishment" or "mediocrity".
In this case, Left is the most democratic winner of the race, also known as the Condorcet winner (assuming Right supporters would support Left over Far Left in a two-person race between them). However, not all elections have a Condorcet winner. Also however, approval voting doesn't guarantee the Condorcet winner wins the race (if one exists), but it's an improvement over FPTP.
https://en.wikipedia.org/wiki/Condorcet_criterion
However, there are various voting impossibility theorems inspired by Arrow which do apply to approval. Most notably, Gibbard-Satterthwaite.
https://en.wikipedia.org/wiki/Comparison_of_electoral_system...
No voting system is perfect but Approval Voting is my favorite. Still, most schemes that people have put forward are far better than the First Past the Post and there's been some momentum for IRV in Mass, where I live, so I've been helping with canvassing and it looks like our efforts might be paying off.
Approval voting is a way of picking a single winner.
They solve entirely separate problems.
(Sortition is nearly optimal if your only goal is to match the distribution of voters by a subset of equally-weighted representatives. Technically, it's possible to improve on it slightly by allowing voters to cluster themselves into a hierarchical structure of parties and factions, then performing sortition under the constraint that each party and/or faction must come within 1 seat of its correct proportional representation. This kind of technique is called "reduced-variance sampling" when used in sequential monte carlo sampling.
But sortition, or party-clustered sortition, pays no attention to candidate quality. Most people would agree that if there are two candidates that hold the same ideology, but the only difference is that one of the two has a medical problem that only allows them to work for 1 hour a week, the other one would make a better representative. Sortition cannot be fixed to take this kind of factor into account, which is a big part of why many people prefer more traditional voting methods.)
Equal chance to represent, there is no fairer way than that.
Voting inherently leads to a money/campaign/popularity contest (and parties, more often than not 2), it is very good at keeping the status quo by feeding back to itself.
I'm not aware of any large scale studies of election method preferences, but keep in mind 1. people are used to voting 2. even if they are aware fairer methods exist their self interest is to keep voting if they think they will benefit from the resulting unfair system.
Can you show how the differences in outcomes is better using the Condorcet method?
The IRV page extrapolates the majority principle for two choices to more than two choices, but its not obvious that this is correct. There are arguments that need to be made before we can start comparing the quality of outcomes between voting methods.
You can convey this by only submitting a single choice on your ballot (a powerful tactic in AV). Conversely, you can vote everyone BUT the one you wish to harm. And by default you can vote NOTA. While it may be that Approval Voting is not the "perfect" answer, it is far better than most and comes with the added benefit of easy adoption/implementation and simple explanation.
Yeah, but then you run the risk of a Trump getting elected in an election between Trump, Hillary, and Bernie, where you really want Bernie, and you really don't like Hillary at all, but you'll still take her any day over the utter disaster that is Trump.
This is why AV isn't that great: all they have to do is put up a truly revolting candidate to scare voters into voting for the not-as-revolting candidate, and she gets elected because there's no way for voters to express a clear preference.
tl;dr for non-experts (I co-founded the Center for Election Science and have studied voting methods since 2006).
The idea behind LNH is that if I vote for the Green in 1st place, it's "safe" to then rank the Democrat 2nd, because that can't cause the Democrat to defeat the Green. See this explained visually in this video by Andy Jennings, another CES co-founder who did his math PhD thesis on voting methods.
http://www.youtube.com/watch?v=JtKAScORevQ
The Later-no-harm criterion is highly misleading, because "safe" here means safe for the CANDIDATE. But tactical voting is about what's good for the VOTER, not the candidate. With IRV, the Green voter's best strategy is to actually rank the Democrat in 1st place. So the whole notion that he can safely rank the Democrat #2 after he's ranked the Green #1 starts with a false premise. If he's rational, he WILL NOT rank the Green #1. It's the same reasoning that Democratic primary voters have about far left candidates. Even if you win the nomination and advance to the general, you're so far left that you'll probably lose against the Republican, by not capturing enough of the center of the electorate.
Approval Voting actually makes it safe to ALWAYS vote your favorite. And, in general, to vote for everyone you prefer to your favorite of the frontrunners.
From what I've seen of most Condorcet methods, the best way to tactically vote differs if there is or is not a Condorcet winner, and often become a major mistake in the other circumstance.
Basically it means unless you know what the results will be tactical voting will have a rather good chance of backfiring on you, making it not worthwhile, which seems like a good thing to me.
This is the "expressiveness fallacy". https://www.electionscience.org/library/expressiveness-in-ap...
tl;dr
Total information throughput for a voting system is a function of THREE things:
1. Amount of information you can express on a ballot.
2. Accuracy of that information (which is distorted if the system gives you an incentive to dishonestly/tactically vote) .
3. Tabulation efficiency — how effectively does the system utilize the info on the ballots (for example there are numerous ranked voting systems from IRV, to Borda, to the entire Condorcet family; and IRV is extremely lossy) .
When you factor these all together, Approval Voting does surprisingly well, especially because it's so good at #2 and #3. See this graph of performance as measured by Bayesian regret. http://scorevoting.net/BayRegsFig.html
Essentially what's happening is that the people who's approval threshold sits between candidates A and B (i.e. those people who only approve one of them) will, in a statistical sense, have about the same A-vs-B preference as the people who voted for both or neither of them. So the aggregate effect doesn't really suffer from your inability to express a ranked preference for all candidates. Voting is statistical.
To really drive that home, note that you could switch to a better voting method and STOP VOTING, and STILL get better election outcomes. The tiny impact your vote has on the process is statistically negligible. The properties of the voting system matter vastly more.
Also in the messy human emotional sense, it's still _unsatisfying_ that I might Approval vote for Bernie and Hillary but I still want to say I like one better, and I still think that difference matters.
As you know, there are various other methods (Condorcet, STAR, 3-2-1, etc.) that keep most of approval's advantages, but sacrifice simplicity for better expressiveness. It would be as unreasonable for the advocates of such methods to try to undermine real-world campaigns for approval, as it would be for approval advocates to undermine those other methods.
I switched from IRV (RCV) to Approval Voting once I better understood the challenges of election administration.
That said, FPTP is just about the worst. For election administration, it's the most brittle, where the margin of error too often exceeds the margin for winning. So I'd happily support IRV, RCV, score voting, approval voting as a replacement.
Comparative fairness is well documented. This graph is a good appromixation:
https://www.electionscience.org/wp-content/uploads/2018/12/c...
https://www.electionscience.org/library/approval-voting-vers...
Simplicity also accounts for election integrity concerns (ease of tabulation and auditing).
Do you have a formal definition for how you are using ‘fairness’?
Start with Arrow's Impossibility Theorem. You have the maths to better understand this stuff than me.
https://en.wikipedia.org/wiki/Arrow's_impossibility_theorem
https://www.math.wisc.edu/~meyer/math141/voting2.html
I have focused on election integrity. For me, arguing the relative merits ranked choice voting strategies is (unforgivable) bikeshedding and ultimately results in no action. Like getting stuck arguing 4.5x vs 4.75x better than FPTP while the world burns.
For election integrity, other real world concerns also must be factored. Like voter education, verifying the hardware & software, ease of tabulation & auditing, feasibility of doing a manual recount, etc.
In conclusion, Approval Voting is almost as fair as the ideal Score Voting but much easier to implement.
NB: I'm the Chair of the Board for The Center for Election Science. AMA.