Boards›Discuss the Boards

Punishment and Forum Moderation

SuicideAll·8/1/2017, 7:44:22 AM·2 votes·1,133 views

https://www.psychologytoday.com/blog/feeling-our-way/201401/punishment-doesnt-work

https://www.sciencedaily.com/releases/2017/07/170704093910.htm

I may add links. Feel free to post your own links.

Sure, I suppose you could arrange for a totalitarian state to ensure that the person is always feeling watched and thereby inhibit the behavior permanently under an umbrella of anxiety, numbness, and hate (the emotions that punishment produces). But even then, the tendency (or desire) to engage in the punished behavior will not change.

Sometimes, of course, punishment is necessary, like when you stop a child from running into a busy street. But if you want it to stick, you have to reinforce a behavior that competes with running into the street (like stopping and waiting for the light).

I think this is the meat and bones of the situation.

How do moderators reinforce a behavior that competes with "toxicity"?

First thoughts, it sounds like allowing a forum poster to edit their post instead of deleting it. With guidance of course. Not sure that is a practical idea though.

Second thought, the guidance could be the most important part. So, delete the posts, but give very clear suggestions on how the post can be edited to remain within the rules?

EDIT: I think it is important to note that I am the only one to provide third party sources. None of these moderators have been able to provide sources for their opinions.

44 Comments

Deep Terror Nami8/1/2017, 11:26:40 AM8 votes

it sounds like allowing a forum poster to edit their post instead of deleting it

While that sounds nice, it's not practical for most situations;

We do edit out some things if the post still has discussion value without the offending segments such as a name & shame post that doesn't need the player's name to discuss the behavior, or a legitimate reply where they needlessly closed with an insult. However, if the post has no discussion value anyway (or loses it without the context we have to remove) or maybe the person is harassing someone and we don't want them replying at all, we don't have a good reason to leave it up.

If we need the player to make adjustments, we can't sit around waiting on them, hoping they will do so in a timely fashion. If the post is a problem we need to action on it as soon as we can. We aren't going to spend all day telling you what your post should say, but the deletion message should definitely be clear on what shouldn't be in it; if a message is not very clear I recommend visiting us on the Moderation Discord to speak with a moderator directly (preferably the one that removed the content).

Teridax688/2/2017, 1:24:26 PM3 votes

Thoughts on this:

  • The article itself mentions that in the reward and punishment scenarios, the behavioral conditioning wore off in 100 level presses in both cases when reinforcement was abandoned. Effectively, the article defeats its own point by showing punishment to be equally effective compared to reward. In psychology and behavioral science (the latter of which is being slowly deprecated for a host of reasons), it is true that reward-based positive reinforcement is the most effective form of conditioning, and punishment-based reinforcement is the least effective, but there is a lot more nuance to this.
  • The comparison between punishment in rats and humans is very surface-level: when a rat is shocked for pressing a lever, they just internalize the fact that pressing the lever carries negative consequences. By contrast, when a human is punished for bad behavior, they are usually told why their behavior was wrong, and so are much more likely to amend their behavior permanently, unless they score exceptionally low on traits like emotional empathy, theory of mind, executive function, and so on. As Wardens, when we do punish a player (and, believe it or not, we generally don't want to punish someone unless their behavior is really egregious), we work to explain to them what they did that was wrong, and why it was wrong. If you feel like this is not being done well enough, please let us know, as proper communication is essential to how we operate.
  • Comparing board moderators to Miniluv is by no means new, but I think it's worth talking about this: as mods, it is both our job and our interest to genuinely improve the atmosphere in the space we're moderating. Our power and actions are kept in check and follow a strict protocol, which means that when mistakes and abuses of power do happen, they're immediately identified and severely reprimanded. We are moderators, but also players, so when the environment on the Boards is more positive, it benefits us too, and by contrast, when it's not, we suffer as well.
  • I agree that it is important that we very clearly explain our rules so that players don't get trapped into having their content deleted on here. We do have rules in place for this, but if they can be made clearer or more visible, please let us know, as it's important that they get to be seen and understood by everyone on here.
  • I also completely agree that it would help tremendously to incentivize positive behavior, and not just punish toxicity when it happens. Right now, though, we're operating on limited resources (we can't change the Boards framework ourselves), and are in need of input for this, so I'd also be more than happy to discuss with you how we should go about achieving rewards for good conduct.
ModKnightsKemplar8/2/2017, 4:56:13 AM2 votes

I haven't sifted through the comments yet to see the rest of the discussion, but I'll tell you that there's one key difference between this experiment and the Boards; the punishment is consistent here (at least in theory).

In the experiment discussed, the author is talking about how punishment is only effective as long as it remains in place. This is important in the context of his argument with punishing children because they are eventually out of reach from your punishments. However, in the case of moderation, the subject is never away from the punishment if it continues to "push the lever."

If the net effect is to make toxic people leave the Boards (avoiding the source, as it were), then I'd say the moderation was successful. It isn't the moderator's job to make sure those people are never mean to anyone ever again; our job is to make sure they aren't mean on the Boards. As such, I'm not sure this experiment, and the ensuing evaluation, is the best analogy to internet forum moderation. The goals are different.

EDIT: To answer your question directly, I think us just being active and calling out posts that are good is about the limit that is useful. We tried to do a "Featured Post' sort of thing on the Gameplay sub-board, but I'm not sure how useful it was. We're still exploring some ideas in that arena, such as the Rengar mega-thread we did recently, and I think that was useful. But I'm not sure it falls into the category of "positive reinforcement," either.

SuicideAll8/2/2017, 11:30:16 AM1 votes

http://www.gdcvault.com/play/1020389/Enhancing-Sportsmanship-in-Online

OMG, Team builder confirmed the best que ever made.

http://www.gdcvault.com/play/1022160/More-Science-Behind-Shaping-Player

Clear feedback is everything Be humble when you eventually screw up

Moderators could improve on those aspects.

SuicideAll8/2/2017, 12:07:06 PM1 votes

If anyone gets time to help, I'd like to see the % of toxic people in a free speech system. So people that use all the nasty words.

I'm under the impression that such data might not exist.

SuicideAll8/5/2017, 12:45:55 AM1 votes

https://www.sciencedaily.com/releases/2017/07/170704093910.htm

More evidence against permanent bans.

The brain uses behavioural consequences to trigger an action more easily even if the consequences are disagreeable for us.

SuicideAll8/5/2017, 12:52:16 AM1 votes

http://www.apa.org/news/press/releases/2002/06/spanking.aspx

Gershoff found "strong associations" between corporal punishment and all eleven child behaviors and experiences. Ten of the associations were negative such as with increased child aggression and antisocial behavior.** The single desirable association was between corporal punishment and increased immediate compliance on the part of the child.**

SuicideAll8/5/2017, 1:02:52 AM1 votes

http://www.sciencedirect.com/science/article/pii/S0010027715000475

I hesitate to link this one. I don't have an account to view all the details.

It was found that the magnitude of reward dictates the rate of choice repetition.

In contrast, penalties cause universal avoidance regardless of the penalty magnitude.

Thus, reinforcement and punishment act as fundamentally distinct behavioral factors.

The choice repetition effect of a reward scaled with the magnitude of the reward. In a marked contrast,** the avoidance effect of a penalty was flat, not influenced by the magnitude of the penalty. **

SuicideAll8/5/2017, 1:16:12 AM1 votes

https://webcache.googleusercontent.com/search?q=cache:zwIHmmFkOxsJ:www.sas.upenn.edu/psych/PLEEP/pdfs/Unpublished%2520manuscript%2520Kurzban%2520DeScioli%2520Obrien.pdf+&cd=43&hl=en&ct=clnk&gl=us

Varying conditions of anonymity, we find that the presence of an audience causes a threefold increase in moralistic punishment.

SuicideAll8/1/2017, 7:01:45 PM1 votes

I'm still here when I have time.

My issue at the moment is with Riot not having a forums engineer. Pretty much means all suggestions have to be about policy because code can not change.

SuicideAll8/2/2017, 9:46:19 AM1 votes

http://oce.leagueoflegends.com/en/news/game-updates/player-behavior/player-behavior-design-values-reform

Reform becomes increasingly likely the shorter the feedback loop (the time between the negative behavior and receiving feedback that the behavior was not acceptable).

I'd like to point something out here. Insults can be feedback. When insults are a retaliation against an unaccepted behavior, they are faster feedback loops than Riot is able to provide.

Insults can be an useful communication tool. Insults can be used to show levels of disagreement. For example, someone that is quick to anger about a subject is assumed to be passionate about that subject.

http://oce.leagueoflegends.com/en/news/game-updates/player-behavior/player-behavior-design-values-punishment

We only outright punish players when we’re really, really sure (false positive rates below one percent) a punishable offense occurred. In every other case, we believe punishment should be structured to encourage reform.

Kicking someone out then letting them walk right back in is not what I consider "removal". Locking someone in a prison cell is "removal" from society. If Riot can not practically remove players, then they need to stop trying. Refocus on reform.

We’ve found the approach of surrounding bad behavior with good influences tends to lead to the best results for improving the experience and helping the community actively reject negative behavior.

Being nice to someone works better than banning them. Toxic people have feelings too.

When we compare the old ban-first, ban-later strategy regarding negative language to the new combo of chat restrictions and bans, we’ve seen 40% more players reform in the new system.

Where is the testing of chat restrictions alone?

So far I'm not very impressed. I read everything except the GDC seminars. There were an unacceptable amount of typos. Math was inconsistent at times. So far nothing proving punishments are effective. It's always with encouragement of positive behavior that the systems become more effective.