you are viewing a single comment's thread.

view the rest of the comments →

[–]Wanga 4 insightful - 2 fun4 insightful - 1 fun5 insightful - 2 fun -  (12 children)

or by a cultural taboo against it, like Voat

[–]magnora7 2 insightful - 1 fun2 insightful - 0 fun3 insightful - 1 fun -  (11 children)

True, cultural taboos can be powerful but hard to cultivate

[–]wizzwizz4[S] 2 insightful - 2 fun2 insightful - 1 fun3 insightful - 2 fun -  (10 children)

I'm attempting with these PSAs, but it's not working. I should really look at the psychology papers to figure out a way to do this, but I keep having ethical issues with doing so.

[–]magnora7 6 insightful - 2 fun6 insightful - 1 fun7 insightful - 2 fun -  (9 children)

Well the thing is people don't like being told how to act. So the more you go "you all have to act like X" you get 20% who agree, then 20% who go "fuck that, no one tells me what to do, I'm going to do the opposite!" so in the end it's hard to tell if any real progress is made. The latter is called the "Backfire effect" in psychology.

This is also one reason I'm against having tons of rules, because it apparently just pisses everyone off and doesn't actually better define the culture of the site, and kind of just actually acts like a wedge. I think encouraging positivity is a more useful tool to the end of creating a good site culture, than trying to define ever-narrowing rules which just anger people who then feel attacked by those rules.

It's much easier to just group everything under the pyramid of debate, and give individual warnings as specific scenarios arrive, then do a '3 strikes you're out' system. Keeping everything low-pressure like this makes a better community, then trying to play "culture cop". If you see what I mean.

Voat used to have this thing called "protect voat" and it was basically a mob that descended on anyone they didn't like. It was one of the things that was supposed to stop the cultural backslide problem, but only served to intensify it. I try and learn from their mistakes.

This stuff is super tricky. I think it's one of those things where "less is more". And "If you do everything right, they won't know you've done anything at all". Both those quotes often come to mind when administrating or moderating. I haven't found a better technique, personally. But I do appreciate the PSA, everything said. The culture has to hold itself up, at the end of the day. But I think it works best by example, rather than by telling people how to act.

Anyway, that's kind of my take on this whole thing, because it is insanely complex and subtle to make it work right. So that's what I've figured out so far, hope that gives you food for thought.

[–]wizzwizz4[S] 2 insightful - 1 fun2 insightful - 0 fun3 insightful - 1 fun -  (8 children)

Definitely food for thought.

Perhaps we could have a little poll thing, and then articles like Part I of this this, and then polls afterwards, and see what happens to people's opinions. We'd have to A|B test it, and have to throw in a tonne of groups and things that we're not targeting with the pseudojournalism to avoid tipping people off… It could be like a competition: who can be the first to spot it? We could run multiple ones at once, and have the polling being constant and ongoing (like chat: in a little box on every page, asking for people's opinions on a scale of 0 to 10 with radio buttons) and it'd encourage people to think critically about everything they read here.

That isn't quite relevant, but it's close.

[–]JasonCarswell 2 insightful - 1 fun2 insightful - 0 fun3 insightful - 1 fun -  (3 children)

Very confusing.

Very complicated.

Didn't read the article.

But there's a seed of an idea that is good on a few levels. Participatory community building, statistical analysis, and in this example, journalism quality assessment.

I don't know about the competition part or "radio buttons".

Sounds like you're going after fake news with a truth meter.

If anything this sounds more like a 3rd and maybe 4th vote. I broke "trust" into 2 for the example below, because you may not agree with someone you trust. It's still lacking because they may discuss several matters but it's better than a vague one.

  • Insightful : Y/N (0 to 10 in future?)

  • Fun : Y/N (0 to 10 in future?)

  • Source : untrusted 0 to 10 trusted (default 5)

  • Content : false 0 to 10 true (default 5)

Might be good if there was an option to participate or not in https://saidit.net/prefs/

If you create this system make it flexible to add more pref options in the future, assuming folks want it.

This may also be an opportunity to just start the voting system over from scratch. Yes there may be a lot of kickback but it may be worth it.

[–]wizzwizz4[S] 2 insightful - 1 fun2 insightful - 0 fun3 insightful - 1 fun -  (2 children)

That's also a good idea, but not what I was proposing.

Look. Erm… I'm bad at explaining stuff. But…

Imagine the site looks like this:

+--------+---------------------------------
| Saidit |    subname
+--------+-----------------------------------
| Title           | chat
| content content | chat
|                +-------------
| content content | How do you feel about THING?
| content         | (not shared publicly)
|                | 1 o o o o o o o o o o 10
|                | [submit]

And then we do that for a lot of THINGs, but one of the THINGs (that we only ask half of the people about, so we have an A group and a B group, one of which is possibly primed to be extra cautious about all of the THINGs they've seen and the other of which isn't) is the TARGET. We have people attacking that TARGET with completely cherry-picked / false evidence posted to select areas of the site and voted insightful enough to get it onto the front page, and then afterwards post a thing saying "TARGET isn't actually bad; this was made up" with the results.

The results would be found by analysing the data from before the TARGET's campaign started, as the campaign progressed, and after the big reveal to see what effect on the Saidit populace this has. Then, eventually, it should stop having so much of an effect because people start going in and analysing what they're being shown.

Yes, the poll thing would be opt-in in the preferences. And I'd try to make it as flexible as possible.


You found that link confusing. That's my fault, sorry; I expected a short inferential distance again. That would also explain why the not misrepresented one I posted got very few Insightful votes; nobody could understand a word of what it was saying.

[–]JasonCarswell 1 insightful - 1 fun1 insightful - 0 fun2 insightful - 1 fun -  (1 child)

Yours seemed too complex.

I still don't see why folks need their votes private. Just be men (and women, and whatev) and own your voice and vote. So this not-shared-publicly thing is stupid IMO.

It still seems too complex. How are you going to limit folks from seeing the other groups content? How do you know they don't already have access or experience with it? This sounds more like something for a University study, not for social media. Too much authoring and prep for so little payoff. And if you're writing the questions/surveys/fakesenarios count me out. This also sounds like some of Facebooks human experiments.

IMO this is a bad idea. At least the group A and B and faked news thing. Too elaborate for naught.

Yet I do like the idea of more stuff in the side box, interactive or not.

d3rr and I had discussed an "admin notice board" between submit buttons and chat. In this box on every page would be pinned whatever they like - with changing links - announcements, plans, top post, whatever they like. The footer links are permanent. These would change with the times. And maybe they could add your 1 to 10 thing with a simple question or survey. For more detailed questions or surveys there'd always be /s/SaidItSurveys which they could of course link to.

[–]wizzwizz4[S] 1 insightful - 2 fun1 insightful - 1 fun2 insightful - 2 fun -  (0 children)

It's not a vote.

The point is to see how easily influenced we are by false but well-written information, so we know that and can do something about it. (Well, I already think I know, but it's always good to check.) Your point about the Facebook thing is good, but so long as it's opt-in— wait, we'd be sharing the false news with everyone…

Urgh. Stupid ethics. :-p Yeah, you're probably right.

[–]magnora7 1 insightful - 1 fun1 insightful - 0 fun2 insightful - 1 fun -  (3 children)

That's not a bad idea, my hesitation would be that over-focusing on the rules as a site culture is going to lead to a rule-obsessed and legalistic site culture, and I'm not sure that's desirable. But that's probably one of the better approaches I've heard, I'll think it over some more

[–]wizzwizz4[S] 2 insightful - 1 fun2 insightful - 0 fun3 insightful - 1 fun -  (2 children)

over-focusing on the rules as a site culture is going to lead to a rule-obsessed and legalistic site culture,

If the rules we're religiously sticking to are the rules about how to have civil debates that are more likely to come up with the correct answers, I see no problem with having a rule-obsessed culture; it'll be a culture obsessed with coming up with the correct answers. However, this only follows for a pretty small subset of rules, so we mustn't get ahead of ourselves here.

[–]magnora7 2 insightful - 1 fun2 insightful - 0 fun3 insightful - 1 fun -  (1 child)

You're right. I guess I mean it more in the sense of having a million rules, and the only people who care enough to keep track of them all are the anal-retentive people trying to create an authoritarian echo-chamber. That's what I saw on reddit again and again, so the less rules the better. Not that we can't have new rules, but if something is covered by the existing rules in any way we should work hard to not add more rules as long as the site culture isn't actually decaying. That's how I see it.

[–]wizzwizz4[S] 2 insightful - 2 fun2 insightful - 1 fun3 insightful - 2 fun -  (0 children)

Oh… Sorry, I must've misread your argument; this one is much stronger than the one I thought you were saying. I guess that's why you're in charge.