Community vote on banning X/Twitter links aka "Twitter in the shitter?"
Community vote on banning X/Twitter links aka "Twitter in the shitter?"
UPDATE: Proposal approved, see https://lemmy.dbzer0.com/post/36484194 for policy change announcement.
Ahoy me hearties!
We were thinking this might be a good test run topic for instance voting in our !div0_governance@lemmy.dbzer0.com community. Please be patient with us if anything breaks or isn't working properly. Feedback is welcome.
The voting topic
Given the current political backdrop and recent video of Elon Musk performing clearly identifiable Nazi salutes at the Presidential inauguration, some communities have started banning all links to X/Twitter. A couple of examples I noticed yesterday:
The vote is on whether our instance should follow suit and implement an instance-wide ban on X/Twitter links in posts and comments.
I've noticed some people suggesting allowing screenshots to still be used (e.g. for memes). Feel free to drop a comment if you have an opinion on that.
How to vote
Simply upvote or downvote this post. The /0 Bot will automatically calculate and update a tally of votes every 15 mins or so according to the voting rules (so don’t expect instant updates). An upvote is counted in favour of the resolution. A downvote is counted as against the resolution.
Note regarding crossposting: please be aware that only votes on the original post in !div0_governance@lemmy.dbzer0.com will be counted.
When to vote
Voting starts as of now. We'll close voting once the flow of votes stops - not sure exactly when that will be yet, but I'd like to keep the topic open for at least 2 or 3 days (maybe a week?) to give everyone a chance to vote.
Who can vote
TLDR here is that anyone can vote, but your votes will be weighted differently depending if you are a financial supporter, local instance member or external instance member.
As discussed in the announcement post, the initial plan was that only stakeholders can vote and open threads. That now includes everyone who is supporting us with any monthly donation amount.
Voting rights have also been extended so that votes of other local instance members who otherwise have no voting rights will be accounted at a rate of 1/100 from a random sample of up to 1000 of their votes. This means that a vote can go up to max +/- 10 from local community votes and it’s a fractional count (i.e. +1.1, or -0.7) which should make the local community sentiment an excellent tiebreaker, without overwhelming the people who are directly supporting the instance. Furthermore, I decided to display the “outsider sentiment” which is votes from non-valid-voting users from other instances. The outsider sentiment is only flavour (“Positive”, “Negative” etc) and is disregarded from the total. This is just shown for reference of the outsider sentiment which I think might be useful.
What constitutes a successful vote on a topic?
We are totally open to debate on this. I was thinking for this topic, a 2/3 majority vote would be a good target to aim for so we can be certain the community vote represents a clear majority of our users' opinions.
My thinking here is that if some topics are split close to 50/50 then achieving a 51% vote for example does not produce a clear mandate and may simply cause unnecessary division.
Having said that, I acknowledge a 2/3 majority is an arbitrary choice, but unless we implement a more complex voting system hopefully it is "good enough" to indicate a clear majority. As mentioned previously, feedback is very welcome and we will review and make adjustments where necessary.
Community participation
I strongly encourage all our instance members as well as subscribers from different instances to vote on this topic. If we only get a small handful of votes it's not going to be very representative of overall sentiment. This is a test run, so if things don't work out in terms of participation we will re-assess and perhaps revisit the topic.#
You're viewing a single thread.
1/100 from a random sample of up to 1000 of their votes.
Whaa?
So votes are 0.01 of a vote?
100 users collectively have 1 vote? Like that?
1/100 for non-supporting, non-vouched and non-mvp users, yes. The linked threads explain it.
1/100 seems extreme, I think a 1/2, 1/3 or even 1/4 of a vote, would be enough to minimize the possible effects of potential sockpuppeting.
It's not just about sockpuppeting. It's about primarily wanting people invested in this instance to have a say. To put it another way, this instance has less than 30 supporters and MVPs putting the actual effort for the 1000 other MAUs. My belief is that those users deserve the most say in the decision-making. A 1/2 from the others, would completely overwhelm any such say.
True, but I think it's good feedback. We should probably consider bumping up the weighting of the non-supporting instance users from 1% of total votes to a larger percentage. I support the idea that donating users should have more of a say, but we could manage it something like the following to ensure there needs to be an alignment between the views of the donating/vouched users and the non-donating home instance users.
Example below using arbitrary weights (for consideration):
- Votes from vouched users or donating users is weighted to 50% of overall tally (this proportion still provides a big boost to voting power given the number of people who donate).
- Overall votes from non-donating home instance users is weighted to 45% of overall tally.
- Overall votes from external instance users is weighted to 5% of overall tally.
The final math would be as follows:
- Donating/vouched members = Count of Yes votes / total votes from this group, e.g. 55% in favour x 50% weighting
- Non-donating home instance sentiment = Count of Yes votes / total votes from this group, e.g. 61% in favour x 45% weighting
- External sentiment = Count of Yes votes / total votes from this group, e.g. 80% in favour x 5% weighting
- Total voting tally: Donating/vouched members + Non-donating home instance sentiment + External sentiment, so for this example, we would get: 27.50% + 27.45% + 4.00% = weighted average of 59% in favour, so it would not pass the 66.6% threshold.
I'll think about it. However I don't think external users should have any say whatsoever in our internal decision making
Yeah fair enough, just added it in for completeness to help explain the math.
Edit: On reflection I'd recommend a 50%/50% split between donating/vouched users and non-donating home instance users. That will help ensure that people can't game the system by making a few small donations in order to force an outcome on the majority of the non-donating instance users, and to ensure there is a general consensus, not just among donating users.
It would also mean 2 donating users would tie with 2 non-donating user which going away from the extra weight I wanted. Anyway I think we should discuss this in its own thread.
If we had voting numbers that low then it won't be a good indicator of sentiment anyway. Ideally we would want a sample size of around ~289 home instance users (95% confidence, 5% margin of error on 1149 monthly active users) to give us confidence in the results. I just think maybe the weighting is too heavy in one direction at present. But yeah, there's a lot of different approaches that might work here so it might need its own thread.
Donating users should naturally also count for non-donating votes (lest you wanna encourage donors having to make secondary accounts to receive 'non-donor privileges').
So it would be 2:1 in favor of the donors in that case.Not that I think it would be common having as many donors as non-donors partake.
Not really. The supporter's votes are already massively weighted comparatively. It's really not worth it making a sockpuppet just to vote.
I would prefer this system to the current one. A census method will have less error than a SRS in determining what the community wants, and this is clearer in how much say the different groups get. With the current system you would have to know the relative sizes of contributing vs non-contributing members to have an idea of how much say the non-contributing members get.
If we did implement this, would we use the vouch system to determine which external members get to vote? It seems vulnerable to brigading if not.
No I wouldn't give any voting to externals either way specifically for brigading concerns. I also don't see why outsiders should get a say in internal instance decisions, yanno?
Hmm, I get where you are coming from db0, but tbh I have a couple of folks who mod communities on this instance who I vouched for, but who mainly use external accounts. I think they both have local accounts too though, so in this particular case it wouldn't be a problem to switch the vouch over to their local accounts.
In any case these folks are, to my mind at least, important instance stakeholders because they each spend time and effort posting content and moderating directly on our instance. I think in the fediverse the idea of who is an "outsider" and who is an "insider" is perhaps a bit more nuanced than just where they registered their account. i.e., some "outsiders" contribute a lot more time and effort to the instance than casual local users who barely ever post anything.
Just some food for thought as we sail along :)
I'm OK with vouching for external accounts so long as we believe they're stakeholders and will have our instance's ethos and best interests in mind. When I was talking about externals, I meant non vouched random users
Awesome :)
If they're a regular on the instance I can see it. They're a part of the community as much as people who register for the instance. But yeah, managing brigading can make that unviable even for a limited number of external users.
I assume that's why there's the vouching process, so we can agree someone deserves a voice in our government regardless of home instance or payment.
Exactly. And mvp as well
