Hello all,
The way I see it, the kind of person willing to be an early-adopter is the kind of person with gumption. They’re willing to deal with uncertainty, they’re willing to stake out a claim based on principles, they’re willing to put in a little extra work where it’s needed.
Alight, maybe, maybe not.
But if that sounds like you, I want to recruit you to lead a referendum to switch Ohio or a local government in Ohio to Approval Voting.
If you haven’t heard of Approval Voting before, here it is:
- Vote for everyone you like.
- Most votes wins.
That’s it.
But OH BOY does it fix a lot of problems. Under Approval, it’s always safe to vote for your favorite candidate. With Approval, you can’t submit an invalid ballot. And best of all, Approval voting doesn’t have spoilers.
Approval Voting helps show how much support every candidate in the race actually has, since there was nothing stopping anyone from voting for them. Everyone’s final total represents their approval rating!
If this sounds like a good deal to you, let me know and we can talk about what it would take to switch your elections to approval. If you have any questions, fire away! I’m even perfectly happy to tell you what kinds of problems and limitations approval has, because all voting systems have problems and limitations!
P.S. I couldn’t find any community specific rules and I’m not sure if this counts as advertising or spam per the site rules but if it does let me know and I’ll delete the post.
P.P.S. I do actually live in Ohio, I’m not gonna go around recruiting people for projects in a state where I don’t live.
Let me preface this by saying I think anything is an improvement over “First Past the Post”.
What makes you prefer approval voting over some other methods (e.g. Ranked Choice)?
Great question!
For any non-shit voting method (damn near anything other than FPTP) the results are usually the same or comparable. This is the case for basically kind of election, be it single-winner, multi-winner, or proportional. Yes yes, there are differences, but honestly they’re kind of small.
To make this point, compare these calculated win regions for four voting methods. Approval isn’t in that video so here’s a link of the same calculations including approval. Anyway, big differences, right? Well, in real-world polling data it seems… Eh. Not so much. You can find polling data which claims big differences, but it’s unclear how good that date is. If we take that data to be valid? Well, approval does very well.
Okay so let’s at least just pretend that the results we get from our fancy new voting methods are all basically the same. What are the differences, then?
Well, now it’s an administrative and user experience question. How easy is it to run your election, how easy is it to verify the results, and how easy is it for the voter to understand what’s going on?
I think it’s worth mentioning that as far as the preceding paragraph is concerned, First Past The Post does extremely well. The only problem is that the results are total shit.
As far as these questions go, it doesn’t get better than approval. I mean, seriously, how could you possibly mess it up? Where are you going to get confused. Vote for everyone you like, most votes wins.
The ease of administration and understanding applies to the multi-winner and proportional methods too. The instructions are the same, vote for everyone you like. With party proportional, your ballot is divided up evenly amongst the parties you voted for. Vote for three parties, and they each receive 1/3 of a vote. You get the idea. With multi-winner, the winners are selected round-by-round. After every round, if you voted for the winner in that round, your ballot weight is reduced according to the harmonic series (1, ½, ⅓, ¼, ⅕, …). You can use any reduction list you like, but the harmonic series is a good compromise.
Since, as the other commenter pointed out, multi-winner elections would be a better fix to our democracy, it’s important we move to a voting system that’s easy to implement and easy to understand no matter what kind of representation system we use. That’s approval voting, it’s dead simple.
If you need convincing to recognize why simplicity is important, I think the large amounts of complaints about Mastodon/Lemmy/Kbin being too complicated is a great example. For a social network it doesn’t matter if some people are excluded because they don’t understand how it works, but for a voting system? It has to work for everyone.
But okay, you’ve read this short essay and you’re thinking, “mmm, but I’m more concerned with the system mechanics. I’m smart enough to understand voting systems and I want one that works really well.”
Great luck! Approval has great voting theory properties too. It passes the Sincere Favorite Criterion, meaning it’s always safe to vote for your favorite, which isn’t actually true under RCV. It fails Later No Harm, which means it discourages voting for candidates you don’t actually like (again, unlike RCV). Finally, spoilers are mathematically impossible under the Independence of irrelevant alternatives and, you guessed it, this isn’t true under RCV.
Is anything better than FPTP? Yes, and so we should fix it right the first time, especially since we’ll ultimately plan to move to multi-winner and/or proportional elections.
Thanks for the detailed writeup.
I think my main concern with approval is specifically the lack of ranking, as it feels weird to essentially equally weight selections when in my preference they are definitely not equally acceptable.
Yup! That’s the common complaint, and it’s also the reason it falls “later no harm.” It’s essentially a balancing act between being able to say everything you want, without encouraging voters to cast “safety votes” for candidates they don’t like at all.
You can get around this problem by going to Score or Star but now you’re adding complexity for not very much gain, if any. Both are looked at, here, which I think I already shared but I’m just pointing to it again. Since even the most disinterested person needs to be able to understand the voting system, I think the extra complexity isn’t worth the extra expression.
The biggest need for everyone to understand the voting system is not so that they, as individuals, know what they’re doing, but so that candidates can’t claim the system was rigged somehow. Yes, bad actors will make that claim anyway, but if the system is too complex for some people to understand, that argument becomes a lot more persuasive.
“It’s rigged!”
That voting system is so complicated even i don’t understand it, it really could be rigged!
You would probably prefer STAR Voting, or a Condorcet variant of RCV, like Ranked Robin or Total Vote Runoff.
They have the more expressive ballots of Hare RCV without carrying over the flaws of FPTP.
I do like Approval Voting, but only for single-winner executive situations, or maybe if the entire legislature were on a single ticket. But for regionalized representation, I feel like Approval Voting would result in an excessively moderate result, not only for the ultimate legislative results, but also for the specific individuals, and many very passionate but less moderate individuals would get the feeling they are not represented in the representative body.
And this, I believe, is what leads to candidates whose sole virtue is a promise of political revenge to people who cannot achieve political representation.
Not that our current situation is any better. But if we’re looking for improvements to representation, I think the first step would be to have larger districts sending multiple representatives who were chosen by different, non-geographic subgroups.
I totally agree with you on the multi-winner districts.
Part of the reason I like approval voting is that if can be very naturally expanded to multi-winner elections using Sequential Proportional Approval Voting.
Basically, the voting is exactly the same, vote for everyone you like, then the first winner is selected as normal. Then, everyone who voted for that winner has their ballots reduced to ½ weight and we tally the votes again to find the next winner. In the third round, if you voted for both winners your ballot counts for ⅓ and if you voted for only 1 of the 2, your ballot counts for ½. Basically, for every winner you have on your ballot, your ballot weight goes down in the next round.
Doing this helps ensure that very popular candidates still get into office while making it possible for candidates who have enthusiastic supporters (who don’t like anyone else) to make it into office in the later rounds.
That’s kind of neat. My main concern with something like that is that the math involved might make it easy for people to undermine the election by sowing confusion about the results.
Also, I’d like to see a mathematical analysis showing the degree to which halving the votes of everyone who put a popular candidate on their “approved” list ends up amplifying the votes of the minority. I do want more groups represented, but I feel like this system may end up going too far in the other direction. Also, it would encourage a weird kind of strategic voting: people not voting for someone they would actually like because they are too popular, and they don’t want their votes cut by supporting a popular candidate.
I’ll probably put more thought into it. I’ve never encountered this idea before, so I’m just spitballing for now.
EDIT: Here’s an idea that might address my second question, but potentially make my first concern even worse. Instead of cutting the value by 1/2, cut it by the ratio between the number of people who supported a winning candidate, and the total number of votes. So, for example, if someone is so great that literally everyone votes for them, this would have no effect on the weight of their vote with regard to other candidates (multiply value by 1/1). If a candidate is so popular 3/4 of voters supported them, multiply their supporters’ votes by 3/4. This would reduce the problem of a vote for a super popular candidate effectively nerfing the votes of those who support them beyond what would be appropriate.
Of course, this makes the math even more complicated, making it more difficult for people to verify the vote, easier to muddy the water.
Yeah that’s a valid concern! The problem of incredibly popular candidates is somewhat self-correcting no matter what decay method you use. To illustrate this, we can take it to the extreme like you did, and use the 1, ½, ⅓, … decay chain for simplicity’s sake.
Suppose literally everyone voted for the most popular candidate. Well, okay, in the next round everyone’s votes count for ½ so the relative popularity of each candidate doesn’t change.
As you lower the popularity of the first winning candidate the few people who didn’t vote for that candidate have twice as much influence in the second round as everyone else. But they’re a small minority, so they’re still going to be fighting against large numbers of people. They might be able to influence who the winner of the second round is, but they’re really going to have to pick between two candidates everyone else likes. The further you lower the popularity of the first winner, well, the less the rest of the population can really be considered a fringe minority.
The reason for the decay chain is fairly simple. If one of the people you voted for got into office, you’ve got some amount of satisfaction already from the results. Your opinion isn’t invalid, but the amount of joy you get from a second representative is less than the amount of joy other people would get from their first representative.
why this vs ranked choice? Ranked choice would let you prioritizes your votes so you can identify the order of the candidates you would prefer, but still allow you to vote for multiple people.
