There has got to be a way to have a scoring based on the civility of a post, rather than just popularity

(Mofosyne) #1

We need a scoring on how civil you are in talking to others. We have ‘love/like’ scoring, but we need a ‘civility’ score as well. And those two metric needs to be separate such that you may be hated by the community, but still be given high standing as a civil speaker.

Basically the ‘love’ heart is mostly use when a person likes a post, but we need to encourage people to downvote post that they like if it does not contribute to the quality of a board. E.g. They may mention a support for an issue that you like, but then add an personal attack on another person (which you dislike from a civilised discourse point of view.)

You can then tie this score that keeps track of how civilised you are in your discourse with others. Perhaps you can make it such that if you are consistently good at being civilised in your discussion, your meta-vote would be weighted higher (aka: you are treated more as a moderator). And this would be separate from your popularity/like score.

Plus this can be tracked by forum moderators as a tool to assist in spotting those who incite incivility, even from within their own ideological side.

Is this part of a wider concept of democratic design philosophy perhaps?

On a related note on civilised discourse, I’m thinking that this sort of consideration would lie in the wider context of democratic design philosophy, of which I have a newish forum to encourage designers to think about these things at Designing Open Democracy. Currently just collating various organisations that I think would be related like Discourse. I’ve also noticed that Nation Builder is dealing with a similar potential issue of how to encourage civilised politics (at least within political parties) via software as well. Have Nation Builder and Discourse considered working more together in exchanging design knowledge as well in regards to fostering these kinds of ideals?

(Mofosyne) #2

To improve granularity, we should perhaps put tags like how slashdot does it, and require a reason tag for all civility up/downvotes.

or maybe a stackoverflow style of a short oneliner comments within a post that can be up or downvoted as well

(Michael Howell) #3

Isn’t that basically what “flagging” is supposed to be?

(Blu McCormick) #4

First of all your concept rocks. We are trying to do something similar in the US with local politics. I was reading your website thinking, “Amen to that, brother.”

Second of all, I really like where you are going with promoting civil discussion. Let’s face it there are enough personal attacks and incivility in politics as it is and I’ve learned that not everyone has the skills to conduct civil discourse.

After reading your topic, I am inspired to really brainstorm about how to do this.

You can change the heart icon for likes to a different symbol and promote it as approving the civility of a post but I suspect people might just start using it as they would a like icon as that is so ingrained.

One thought I have is we installed the Q-A (question-answer) plugin, which permits people to upvote one reply per topic. Ideally, some of us would like it if members could upvote more than one reply per topic (or at least be able to set the limit). In any case, the upvoting of replies could be promoted as specifically for the most civil, thoughtful answer.

I am educating myself on tags and don’t know much about them, but wouldn’t a problem be that if every member was allowed to use tags, the tags (where you pick the reason for the upvote) would have to be associated with the upvote mechanism?

Question Answer Plugin
Retort - a reaction-style plugin for Discourse
(Mofosyne) #5

Well I think I would rather keep the ‘like’ icon to keep it as a lightning rod for populism sentiment. Without it, people will just look for the next best thing to express their tribal feelings, its possibly unavoidable so you will have to account for that.

But what I’m proposing is that in conjunction with the “like” button/metric, you should have a separate mechanism for measuring civility. And that the civility mechanism be used as a feedback mechanism for those who is posting it.

And also I suggest that there is two known approaches for it in practice already. The slashdot tagging system as well as the stackoverflow metacomment system:

  • stackoverflow: The reason I was thinking of stackoverflow example, is that I like how the “side comments” is used as a meta-comment on how well the answer is going. So for discourse, you could have people add meta comments like “Please break up your post into paragraphs” (with maybe a diff that the original poster can then click to merge into their post). This is the most flexible, but takes up precious screen real-estate, also harder to do granular metric compared to tags.

  • slashdot: Is much simpler of placing a tag (like linkedin skills tag), where you can then have a quick feedback of pre-approved tags (e.g. (23 up-votes : Wall of text)). This is more quicker and anonymous, but it is not as detailed. However it does lend to easier application of setting reward badges, or automated warnings thresholds. (e.g. Grammar Nazi Badge, for those that use lots of spellcheck tags)

Maybe the ideal approach is to just have both, but where meta-comments is optional however must be associated with a tag. This would encourage people to just use tags if possible, but give people the flexibility to elaborate and suggest improvements to another person’s post.

(Dave McClure) #6

Maybe curating a set or civility-related emoji and using this plugin would be an option? Retort - a reaction-style plugin for Discourse

You could have one meta topic that describes more fully the intended meaning of each of the chosen emoji.

(Angus McLeod) #7

I think there are two concepts here. One is “civility” in the vernacular of different cultures (e.g. Anglo-Saxon or Chinese). The other is “compliance with local norms” (‘local’ here meaning specific forums). They are interrelated, but different in important ways. I think you’re talking about both, but I’ll deal with the latter (‘compliance with local norms’), as it is perhaps the more relevant one.

Consider the Stack Overflow example. A good question, or answer, on SO is clear, to the point, non-duplicative and uses examples well. Comments on answers will point out the presence or lack of such attributes.

But if you were talking with someone in real life, some of these attributes would be undesirable and you would want additional attributes such as an exchange of pleasantries. If someone asked or answered questions in real life as one does on SO, they would be considered a bit strange.

Likewise, if someone talked exactly as one does on SO in other online communities they may be considered, perhaps not uncivil, but not adhering well to the norms of the community.

This brings us to the crux of the issue, which is that local norms are local. Having meta-comments or meta-tags on a post may be an ok way to encourage norms in one forum, but it wouldn’t work in all forums. Some would even consider it rude.

The closest you can get to a universal feature that signals norm compliance is the like button. As you pointed out, the like button is not a perfect proxy for norm compliance, but it has a broad utility as a feature in discussion software being used in a variety of communities with a variety of norms.

That said, as others have pointed out, you could extend the architecture of norm compliance/recognition in particular communities by using plugins.


I was originally going to add that local norms are epiphenomenal emergent properties of individual forums and that robust norm recognition structures are apt to overly prefigure their emergence.

But that sounds a bit wanky :wink:

(Erlend Sogge Heggen) #8

As @notriddle pointed out, Flagging already covers this.

Every community draws their own cultural line for what is acceptable and what isn’t. Flagging enforces a very straight forward zero-tolerance policy on bad behaviour: No matter how useful your content may be to a lot of users, if it insults/disrupts/aggravates/… a large amount of users at the same time, the content loses its publishing rights.

Most communities don’t need to distinguish between “good content” and “good content with poor civility” because the latter should be rewritten to be civil or not be posted at all.

What are "Likes"?
(Blu McCormick) #9

I am using the retort emoji plugin:

I have set up a separate category, “constructive discourse” and have emojis representing things like:

Great research, needs facts, personal attack, bridging viewpoints, good brainstorming, I learned something, receptive to constructive criticism, etc. This category of emojis is the only one available for posting an emoji similar to how you click on like.

We are in the world of politics and grass roots change and are shaking off the toxic misinformed culture of next door and facebook groups and extra training is essential. We don’t mind turning off people who aren’t on board with that.

When you hover over the emojis their titles come up and just looking at them gives you a sense of the respectful discourse we expect and outline in guidelines and enforce through moderation. Using these emoticons members get to be a part of the moderation without actually making decisions about the consequences.

At least that is the idea. The proof is in the pudding, right? So time will tell. I whipped these up today just to play with them.

(Christoph) #10

This is a fantastic idea! I love it. I seen much use for the retort plugin so far, but this changes everything!

So you added those to Admin -> Customize -> Emoji? And then how do you make sure that only those emojis can be used in the retort plugin? Or don’t you mind allowing users to retort with any available emoji?

Also, it would be interesting to hear your experiences at some point (or anyone else’s) of whether the retort plugin leads to lower usage of the like button (which would mean that it would be somewhat counter productive with regards to the OP’s aim for more rather than less scores.

Finally, out of curiosity: how do you use :tag_this:, :pls_pm_this: or :link_this_:?

(Blu McCormick) #11

Hey @tophee,

I am waiting for discoursehosting to install the plugin by tomorrow then I will have a lot of fun playing with it and would be happy to share videos of it in play or snapshots of the back end.

I feel like this changes everything, too! Gives our forum a less policed feel, recruits members into leading by example, and also serves to set up expectations (honestly, some of these emojis aren’t gonna be used – just their existence for people to see, might be enough of a deterrent for bad behavior.)

This topic gives step by step instructions for adding your set of emojis:

You’ll find them by clicking on emojis in the reply edit window and scrolling to the uncategorized category at the bottom. Go into settings—> customize text and change this to give your uncategorized a different title:

I am presuming functionality of the plugin before trying it from the topic on the plugin. For example, it shows counts of how many people clicked a particular emoji:
You can set in which categories the emojis can be used:

Looks like up go 25 emojis works nicely:

You can limit to a small subset of reactions:

It looks like you can limit what emojis can be used for the retorts:

We haven’t launched yet, but my thought behind those three emojis was something like:

“link this” to the source you are referring to (we want as much evidence to back up statements as possible)
“tag this” is for tagging blog ideas to particular authors assigned to a blog
“pls PM this” is for members who end up in a discussion about things not directly relevant to the topic

(Carson) #12

Please let us know how it goes!

To me, this looks way too complicated. I think having a much simpler schema of 1-4 emojis will be easier. It is a lot of work to think, ‘hmmmm, this post isn’t quite right, but in what specific way? Let me review 25 emojis and pick just the right one.’ Even in conversation, to have that kind of emotional awareness, is rare.

That said, I do think supplementing the ‘like’ button with the retort plugin does offer another way of encouraging good behavior! The like button can function as feedback on value and the retort emoji can function as feedback on the tone of the post.

(Dave McClure) #13

This. I definitely would start smaller if I were going to start experimenting with this approach.

(Blu McCormick) #14

Will try and remember to report in when have data. It’s great for our moderators because we can post an icon representing some particular discourse aspect we are praising, correcting members for and members who are curious can hover over the icon to see what it is. That way we educate everyone at their leisure in a more fun way. Some icons I imagine people will get into and there will probably be a few popular top three icons – like, “Power to People,” or “I learned something.” Members don’t need to use them or “learn” them. Make sense? All you do is hover over an icon and it tells you what it is for. That in itself in a politically charged environment is priceless. I can’t speak for @mofosyne who started this topic but I do observe that he and I are looking to make changes in the political arena including bringing about respectful constructive discourse so perhaps we prioritize this aspect more than most.

(Mofosyne) #15

That’s correct, I do hope to see improvement in the political arena by making it easier for respectful discourse to occur.

While there is already the all/nothing flagging method of enforcing such discourse. It is often heavy handed and labour intensive. Plus I think it could help to take a page off the idea of restorative justice rather than retributive justice of rehabilitating offenders rather than relying on the threat of punishment.

The idea is that receiving granular feedback on their message and why it is respectful or not, would at least encourage people to self regulate over time.

In addition this is not purely restorative justice, but it also provides metric that allows for more targeted monitoring of those with lower civility metric score. This would allow moderators to more effectively focus on those that are known problem makers.

(Blu McCormick) #16

Good point. Right now using the retort plugin I’d have to manually take notes about retorts members receive from everyone in the forum. Having that info would be useful to capture more nuanced data.