Popular Science is turning off its online comments section: http://www.popsci.com/science/article/2013-09/why-were-shutting-our-comments... .
Why? Turns out, incivility ruins people's ability to think clearly and rationally:
"Uncivil comments not only polarized readers, but they often changed a participant's interpretation of the news story itself. In the civil group, those who initially did or did not support the technology — whom we identified with preliminary survey questions — continued to feel the same way after reading the comments. Those exposed to rude comments, however, ended up with a much more polarized understanding of the risks connected with the technology. Simply including an ad hominem attack in a reader comment was enough to make study participants think the downside of the reported technology was greater than they'd previously thought."
Food for thought...
(h/t to Jared Zimmerman)
On 09/24/2013 05:06 PM, Maryana Pinchuk wrote:
Popular Science is turning off its online comments section: http://www.popsci.com/science/article/2013-09/why-were-shutting-our-comments....
Why? Turns out, incivility ruins people's ability to think clearly and rationally:
That's sad to hear. I don't doubt that the effect they identified can occurred. But it's unfortunate that they couldn't find any solution to address it satisfactorily. Of course, this is an old problem, and there are many solutions that have been tried:
* Straight up moderation (e.g. by an employee/employees of the newspaper, or a small list of trusted people) before posting appears (fails, mainly due to the time required and resulting lag). * Real-name comments (my understanding is people are less likely to be rude if using their real name, though I don't think it magically solves the whole problem) * Meta-moderation - Let readers and active participants moderate, and let/require other active participants check whether they're doing a good job
The public policy problem they discussed (everyday people suddenly thinking they're expert scientists and that real scientists don't know what they're talking about) is also real and worrisome.
Matt
On Wed, Sep 25, 2013 at 6:50 PM, Matthew Flaschen mflaschen@wikimedia.orgwrote:
That's sad to hear. I don't doubt that the effect they identified can occurred. But it's unfortunate that they couldn't find any solution to address it satisfactorily. Of course, this is an old problem, and there are many solutions that have been tried:
- Straight up moderation (e.g. by an employee/employees of the newspaper,
or a small list of trusted people) before posting appears (fails, mainly due to the time required and resulting lag).
- Real-name comments (my understanding is people are less likely to be
rude if using their real name, though I don't think it magically solves the whole problem)
- Meta-moderation - Let readers and active participants moderate, and
let/require other active participants check whether they're doing a good job
The public policy problem they discussed (everyday people suddenly thinking they're expert scientists and that real scientists don't know what they're talking about) is also real and worrisome.
Yeah how does all this compare to your experience with comments on StackOverflow Matt?
(For those who don't know Matt is a big time StackOverflow contributor.)
On 09/25/2013 09:58 PM, Steven Walling wrote:
Yeah how does all this compare to your experience with comments on StackOverflow Matt?
(For those who don't know Matt is a big time StackOverflow contributor.)
Stack Overflow is pretty good about this. I don't see a lot of blatantly bad posts (ad hominem, etc.). I think this is partly because of who ends up on the site to start with, partly because of the structure the technology imposes, and partly because of good moderation):
StackOverflow is *very* structured (I think enforcement has also gotten stricter over time) in terms of what you can post.
* If you post a question that is off-topic (you'll be sent to the appropriate other Stack Exchange site, if any), a duplicate (you'll be linked to what it's a duplicate of) or just unreadable/spammy/very bad, it will be closed. Questions can also be -1'ed. * If you post a bad answer, it will be -1'ed. If it's very bad/spammy, it can be deleted, though this is not as common for just badly informed posts. An interesting case is that "Thank you!" answers are also deleted, because they're not, well, answers. * Comments should be used when you're actually commenting on someone's question or answer. "Thank you!" comments can also be deleted.
There is a review tool (http://stackoverflow.com/review) specifically to ensure adequate moderation. Since it's StackOverflow, you can get badges for reviewing (http://stackoverflow.com/help/badges)
The review tool currently has (parentheticals are my own explanation):
* Close Votes (closing questions requires a quorum, so this lets you support a close, if appropriate) * Suggested Edits (Stack Overflow is a wiki, but if you have < 2000 reputation your edits have to be approved Flagged Revisions style). * First Posts (pretty self-explanatory; first posts are more likely to need moderation) * Late Answers (these can also require moderation; "Thank you!" posts are common here) * Low Quality Posts (moderate posts the Stack Exchange software flagged as potentially problematic) * Reopen Votes (similar to Close Votes, but voting to reopen).
Some moderation can be done simply by higher-repped users (http://stackoverflow.com/help/privileges), but there are also diamond moderators with extra powers (http://stackoverflow.com/help/site-moderators). These are elected, except for new sites that don't have an electorate yet.
Matt
I think we can learn a lot from Stack Overflow about the structure and moderation of comments.
Creating and sustaining useful environments for web comments is a difficult task for anyone, but there are methods that can make the experience better.
In a related note, YouTube is trying out different approaches for their own comments sections, integrating them more closely with Google+, as detailed here:
"You'll start to see a new set of comments rise to the top: those by the video's creator, "popular personalities" (i.e., YouTube celebrities), "engaged discussions" with a long thread, and people you know and interact with — both on YouTube and Google+."
http://mashable.com/2013/09/24/youtube-comments-upgrade/
I think the last one is key -- surfacing comments from people you know and interact with seems like a sensible way to weed out the trolls.
But this suggests moving to a tiered presentation of comments, rather than a chronological one, which is a big departure. We experimented with this approach with the Article Feedback Tool, which shows 'featured' comments by default, and tucks away comments found inappropriate or requiring no action.
https://en.wikipedia.org/wiki/Wikipedia:Article_Feedback/Help/Editors
Note that tiered comments doesn't mean that you can't have chronological conversations within each discussion items. But this approach ranks conversations in the initial display to the user (unless they wish to change their filters).
In our environment, this could possibly be done by tracking who you interact with -- and eventually providing the option to 'follow' users you find interesting, which I think is long overdue.
This whole field is particularly challenging, as nobody has found a killer app yet that solves all these issues. Until we do, I think comments will remain an important component of interacting on the web, even if their crude and unsatisfying form. But there are some promising approaches which we might want to learn from and possibly adapt for Wikipedia.
-f
On Sep 25, 2013, at 8:40 PM, Matthew Flaschen wrote:
On 09/25/2013 09:58 PM, Steven Walling wrote:
Yeah how does all this compare to your experience with comments on StackOverflow Matt?
(For those who don't know Matt is a big time StackOverflow contributor.)
Stack Overflow is pretty good about this. I don't see a lot of blatantly bad posts (ad hominem, etc.). I think this is partly because of who ends up on the site to start with, partly because of the structure the technology imposes, and partly because of good moderation):
StackOverflow is *very* structured (I think enforcement has also gotten stricter over time) in terms of what you can post.
- If you post a question that is off-topic (you'll be sent to the appropriate other Stack Exchange site, if any), a duplicate (you'll be linked to what it's a duplicate of) or just unreadable/spammy/very bad, it will be closed. Questions can also be -1'ed.
- If you post a bad answer, it will be -1'ed. If it's very bad/spammy, it can be deleted, though this is not as common for just badly informed posts. An interesting case is that "Thank you!" answers are also deleted, because they're not, well, answers.
- Comments should be used when you're actually commenting on someone's question or answer. "Thank you!" comments can also be deleted.
There is a review tool (http://stackoverflow.com/review) specifically to ensure adequate moderation. Since it's StackOverflow, you can get badges for reviewing (http://stackoverflow.com/help/badges)
The review tool currently has (parentheticals are my own explanation):
- Close Votes (closing questions requires a quorum, so this lets you support a close, if appropriate)
- Suggested Edits (Stack Overflow is a wiki, but if you have < 2000 reputation your edits have to be approved Flagged Revisions style).
- First Posts (pretty self-explanatory; first posts are more likely to need moderation)
- Late Answers (these can also require moderation; "Thank you!" posts are common here)
- Low Quality Posts (moderate posts the Stack Exchange software flagged as potentially problematic)
- Reopen Votes (similar to Close Votes, but voting to reopen).
Some moderation can be done simply by higher-repped users (http://stackoverflow.com/help/privileges), but there are also diamond moderators with extra powers (http://stackoverflow.com/help/site-moderators). These are elected, except for new sites that don't have an electorate yet.
Matt
EE mailing list EE@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/ee
_______________________________
Fabrice Florin Product Manager Wikimedia Foundation
On 26 September 2013 17:18, Fabrice Florin fflorin@wikimedia.org wrote:
This whole field is particularly challenging, as nobody has found a killer app yet that solves all these issues.
That's because the problem is AI-complete, because people are a problem.
"I've found an application which will detect which of your friends are small-minded racists! It's called Facebook."
- d.
On 09/26/2013 12:18 PM, Fabrice Florin wrote:
I think we can learn a lot from Stack Overflow about the structure and moderation of comments.
Creating and sustaining useful environments for web comments is a difficult task for anyone, but there are methods that can make the experience better.
I would add that the best approach depends hugely on the community who will comment, and the goals of the developers.
In a related note, YouTube is trying out different approaches for their own comments sections, integrating them more closely with Google+, as detailed here:
"You'll start to see a new set of comments rise to the top: those by the video's creator, "popular personalities" (i.e., YouTube celebrities), "engaged discussions" with a long thread, and people you know and interact with — both on YouTube and Google+."
http://mashable.com/2013/09/24/youtube-comments-upgrade/
I think the last one is key -- surfacing comments from people you know and interact with seems like a sensible way to weed out the trolls.
It is worth being aware of this. However, YouTube is acting under very different constraints from us.
In particular, they have a huge number of commenters, who often have no actual constructive feedback on the video, and simply want to vent (often on something not at all related, or tangentially related). In other words, they simply have to deal with a firehose.
Our goal is to enable productive collaboration. We don't have much of a firehose, at least on actual talk pages. Even on our busiest articles, it's far more manageable than on even a somewhat popular YouTube video.
Google is also still trying hard to catch up with Facebook, which is an obvious secondary motivation for integrating Google+ into everything they do.
In our environment, this could possibly be done by tracking who you interact with -- and eventually providing the option to 'follow' users you find interesting, which I think is long overdue.
We should be careful about taking the filter bubble (http://en.wikipedia.org/wiki/Filter_bubble) idea too far, where the app mysteriously chooses what you see, for reasons unbeknownst to you.
If we look into following users, we need to be sure that will actually promote productivity, rather than simply endless clicking around (without actually *doing* anything) Facebook-style or wiki-hounding (https://en.wikipedia.org/wiki/Wikipedia:Harassment#Wikihounding)
This whole field is particularly challenging, as nobody has found a killer app yet that solves all these issues. Until we do, I think comments will remain an important component of interacting on the web, even if their crude and unsatisfying form. But there are some promising approaches which we might want to learn from and possibly adapt for Wikipedia.
There is no one killer app, because we all have different goals and trade-offs.
Matt Flaschen
On Thu, Sep 26, 2013 at 3:42 PM, Matthew Flaschen mflaschen@wikimedia.orgwrote:
We should be careful about taking the filter bubble (http://en.wikipedia.org/wiki/**Filter_bubblehttp://en.wikipedia.org/wiki/Filter_bubble) idea too far, where the app mysteriously chooses what you see, for reasons unbeknownst to you.
If we look into following users, we need to be sure that will actually promote productivity, rather than simply endless clicking around (without actually *doing* anything) Facebook-style or wiki-hounding (https://en.wikipedia.org/**wiki/Wikipedia:Harassment#**Wikihoundinghttps://en.wikipedia.org/wiki/Wikipedia:Harassment#Wikihounding )
These are both extremely astute observations. Thank you Matt.
On Wed, Sep 25, 2013 at 6:50 PM, Matthew Flaschen mflaschen@wikimedia.orgwrote:
But it's unfortunate that they couldn't find any solution to address it satisfactorily.
I'm sure they could find a solution if they wanted to. But all possible solutions require effort - software design/expense, raw manpower, or usually both. If you're not us, StackOverflow, or another site that actually depends on functional community, then that is all cost and no discernable benefit.
Luis
On 09/26/2013 12:55 PM, Luis Villa wrote:
On Wed, Sep 25, 2013 at 6:50 PM, Matthew Flaschen <mflaschen@wikimedia.org mailto:mflaschen@wikimedia.org> wrote:
But it's unfortunate that they couldn't find any solution to address it satisfactorily.I'm sure they could find a solution if they wanted to. But all possible solutions require effort - software design/expense, raw manpower, or usually both. If you're not us, StackOverflow, or another site that actually depends on functional community, then that is all cost and no discernable benefit.
Luis
It's not applicable for us, but for PopSci, there are out-of-the-box tools they could have used. I'm not sure how much they considered them.
Matt Flaschen
<quote name="Matthew Flaschen" date="2013-09-25" time="21:50:19 -0400">
- Real-name comments (my understanding is people are less likely to
be rude if using their real name, though I don't think it magically solves the whole problem)
I thought this was actually a bunk theory? For my non-science reasoning: see facebook. Full of hateful/stupid/racist/etc/etc comments.
Greg
On 26 September 2013 18:07, Greg Grossmeier greg@wikimedia.org wrote:
<quote name="Matthew Flaschen" date="2013-09-25" time="21:50:19 -0400">
- Real-name comments (my understanding is people are less likely to
be rude if using their real name, though I don't think it magically solves the whole problem)
I thought this was actually a bunk theory? For my non-science reasoning: see facebook. Full of hateful/stupid/racist/etc/etc comments.
Yes. (a) there's no evidence for this theory (b) it's seriously discriminatory:
http://geekfeminism.wikia.com/wiki/Who_is_harmed_by_a_%22Real_Names%22_polic...
- d.
On Thu, Sep 26, 2013 at 10:08 AM, David Gerard dgerard@gmail.com wrote:
I thought this was actually a bunk theory? For my non-science reasoning: see facebook. Full of hateful/stupid/racist/etc/etc comments.
Yes. (a) there's no evidence for this theory (b) it's seriously discriminatory:
http://geekfeminism.wikia.com/wiki/Who_is_harmed_by_a_%22Real_Names%22_polic...
Yep, agreed. Requiring real names is never going to happen on Wikipedia.
Jared has, in the past, suggested that for users that choose to set their real name (in preferences or during registration, if that field is an option at that point), we could display Flow comments or similar under someone's real name. I thought this was interesting, since the real name field is specifically intended for public attribution purposes AFAIK?
Interesting example of failure of real name requirements: for a time South Korea mandated that all websites over 100K visitors had to use real names. It decreased abusive comments from drive-by users (1-2) comments, but did nothing to encourage civility in longtime participants.[1]
As I told Jared and some of you, I am worried that we're entering a slippery slope where we start designing incentives and functionality around real names and we end up treating pseudonym users as second-class citizens. There are many intermediate options between real names and "cheap" pseudonyms. Paul Resnick and Kaliya Hamlin gave excellent presentations at the Reputation workshop last year [1] on many alternatives that can be explored. I asked them to share a couple of readings on this issue and I'll forward them to the list when I hear back from them.
Dario
[1] https://en.wikipedia.org/wiki/User:DarTar/Hypothes.is_Reputation_Workshop
On Sep 26, 2013, at 10:40 AM, Steven Walling swalling@wikimedia.org wrote:
On Thu, Sep 26, 2013 at 10:08 AM, David Gerard dgerard@gmail.com wrote:
I thought this was actually a bunk theory? For my non-science reasoning: see facebook. Full of hateful/stupid/racist/etc/etc comments.
Yes. (a) there's no evidence for this theory (b) it's seriously discriminatory:
http://geekfeminism.wikia.com/wiki/Who_is_harmed_by_a_%22Real_Names%22_polic...
Yep, agreed. Requiring real names is never going to happen on Wikipedia.
Jared has, in the past, suggested that for users that choose to set their real name (in preferences or during registration, if that field is an option at that point), we could display Flow comments or similar under someone's real name. I thought this was interesting, since the real name field is specifically intended for public attribution purposes AFAIK?
Interesting example of failure of real name requirements: for a time South Korea mandated that all websites over 100K visitors had to use real names. It decreased abusive comments from drive-by users (1-2) comments, but did nothing to encourage civility in longtime participants.[1]
-- Steven Walling, Product Manager https://wikimediafoundation.org/
EE mailing list EE@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/ee
On Thu, Sep 26, 2013 at 10:50 AM, Dario Taraborelli < dtaraborelli@wikimedia.org> wrote:
As I told Jared and some of you, I am worried that we're entering a slippery slope where we start designing incentives and functionality around real names and we end up treating pseudonym users as second-class citizens. There are many intermediate options between real names and "cheap" pseudonyms. Paul Resnick and Kaliya Hamlin gave excellent presentations at the Reputation workshop last year [1] on many alternatives that can be explored. I asked them to share a couple of readings on this issue and I'll forward them to the list when I hear back from them.
I think that's a totally legitimate concern. I don't think we should stumble blindly toward real names anywhere without considering the negative impact that might have on pseudonymous users, whether it confuses newbies ("why are some real names and others screen names?"), and more. I think we all know that longstanding pseudonyms can be just as real as a traditional name, and I agree that when it comes to identity, we should treat them as first-class citizens in design solutions.
I feel that we'll be better off encouraging strong pseudonymity over real identity. I feel that trying to force people into using their real names works against some of our core values.
I'd suggest that we *stop* talking about "real names" and maybe begin thinking about "display names". The terminology is different but the end result is the same. (I also feel that adding a "real name" field to the site will go over like a lead balloon).
On Sep 26, 2013, at 11:38 AM, Steven Walling swalling@wikimedia.org wrote:
On Thu, Sep 26, 2013 at 10:50 AM, Dario Taraborelli dtaraborelli@wikimedia.org wrote: As I told Jared and some of you, I am worried that we're entering a slippery slope where we start designing incentives and functionality around real names and we end up treating pseudonym users as second-class citizens. There are many intermediate options between real names and "cheap" pseudonyms. Paul Resnick and Kaliya Hamlin gave excellent presentations at the Reputation workshop last year [1] on many alternatives that can be explored. I asked them to share a couple of readings on this issue and I'll forward them to the list when I hear back from them.
I think that's a totally legitimate concern. I don't think we should stumble blindly toward real names anywhere without considering the negative impact that might have on pseudonymous users, whether it confuses newbies ("why are some real names and others screen names?"), and more. I think we all know that longstanding pseudonyms can be just as real as a traditional name, and I agree that when it comes to identity, we should treat them as first-class citizens in design solutions.
-- Steven Walling, Product Manager https://wikimediafoundation.org/ _______________________________________________ EE mailing list EE@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/ee
--- Brandon Harris, Senior Designer, Wikimedia Foundation
Support Free Knowledge: http://wikimediafoundation.org/wiki/Donate
I think the most important requirement for successful social interactions online is 'persistent identity' -- so the same user name is used consistently to represent the same person.
The quality of the interactions tend to break down is when a person changes their user name all the time, or uses multiple accounts simultaneously, which can be very confusing to other users.
Once 'persistent identity' has been established, the next most important requirement, IMHO, is the ability to 'follow' users whose contributions you find interesting, so you can surface their comments or activity in your feeds.
'Real names' can also add value to social interactions, by identifying who you are interacting with and encouraging people to be more accountable for their words and actions. For that reason, I support offering a 'real name' field. But this should be an optional feature, not a requirement, in my view.
The key for me is 'persistent identity'. Without it, online interactions tend to break down quickly, in my experience. So I would recommend focusing on that requirement first.
My 2 cents,
-f
On Sep 26, 2013, at 11:49 AM, Brandon Harris wrote:
I feel that we'll be better off encouraging strong pseudonymity over real identity. I feel that trying to force people into using their real names works against some of our core values.
I'd suggest that we *stop* talking about "real names" and maybe begin thinking about "display names". The terminology is different but the end result is the same. (I also feel that adding a "real name" field to the site will go over like a lead balloon).
On Sep 26, 2013, at 11:38 AM, Steven Walling swalling@wikimedia.org wrote:
On Thu, Sep 26, 2013 at 10:50 AM, Dario Taraborelli dtaraborelli@wikimedia.org wrote: As I told Jared and some of you, I am worried that we're entering a slippery slope where we start designing incentives and functionality around real names and we end up treating pseudonym users as second-class citizens. There are many intermediate options between real names and "cheap" pseudonyms. Paul Resnick and Kaliya Hamlin gave excellent presentations at the Reputation workshop last year [1] on many alternatives that can be explored. I asked them to share a couple of readings on this issue and I'll forward them to the list when I hear back from them.
I think that's a totally legitimate concern. I don't think we should stumble blindly toward real names anywhere without considering the negative impact that might have on pseudonymous users, whether it confuses newbies ("why are some real names and others screen names?"), and more. I think we all know that longstanding pseudonyms can be just as real as a traditional name, and I agree that when it comes to identity, we should treat them as first-class citizens in design solutions.
-- Steven Walling, Product Manager https://wikimediafoundation.org/ _______________________________________________ EE mailing list EE@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/ee
Brandon Harris, Senior Designer, Wikimedia Foundation
Support Free Knowledge: http://wikimediafoundation.org/wiki/Donate
EE mailing list EE@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/ee
_______________________________
Fabrice Florin Product Manager Wikimedia Foundation
On 09/26/2013 01:40 PM, Steven Walling wrote:
Yep, agreed. Requiring real names is never going to happen on Wikipedia.
Yeah. Just be clear, I was not advocating it for Wikimedia sites.
Jared has, in the past, suggested that for users that choose to set their real name (in preferences or during registration, if that field is an option at that point), we could display Flow comments or similar under someone's real name. I thought this was interesting, since the real name field is specifically intended for public attribution purposes AFAIK?
Yes, officially the main point of the real name field in MediaWiki is "giving you attribution for your work". Note that Wikimedia specifically overrides this setting to make the field hidden (not just optional, but sot so no one can fill it out).
If people did choose to fill in a real name field (which would have to be clearly marked as public), then it's worth considering displaying it in discussions.
Interesting example of failure of real name requirements: for a time South Korea mandated that all websites over 100K visitors had to use real names. It decreased abusive comments from drive-by users (1-2) comments, but did nothing to encourage civility in longtime participants.[1]
Very interesting, particularly the distinction between drive-bys and long-time users.
Matt
At the very least, let's make it default behavior that the username has two editable fields, display name and ID. Tilde-signature templates would show Display Name, and link to ID.
That allows for lots of flexibility, without any commitment that one or the other is a Certifiable personality or whatever. I guess users are welcome to put some kind of authoritarian "lock" icon next to their name, as you wish ;)
-Adam
On Thu, Sep 26, 2013 at 4:01 PM, Matthew Flaschen mflaschen@wikimedia.orgwrote:
On 09/26/2013 01:40 PM, Steven Walling wrote:
Yep, agreed. Requiring real names is never going to happen on Wikipedia.
Yeah. Just be clear, I was not advocating it for Wikimedia sites.
Jared has, in the past, suggested that for users that choose to set
their real name (in preferences or during registration, if that field is an option at that point), we could display Flow comments or similar under someone's real name. I thought this was interesting, since the real name field is specifically intended for public attribution purposes AFAIK?
Yes, officially the main point of the real name field in MediaWiki is "giving you attribution for your work". Note that Wikimedia specifically overrides this setting to make the field hidden (not just optional, but sot so no one can fill it out).
If people did choose to fill in a real name field (which would have to be clearly marked as public), then it's worth considering displaying it in discussions.
Interesting example of failure of real name requirements: for a time
South Korea mandated that all websites over 100K visitors had to use real names. It decreased abusive comments from drive-by users (1-2) comments, but did nothing to encourage civility in longtime participants.[1]
Very interesting, particularly the distinction between drive-bys and long-time users.
Matt
______________________________**_________________ EE mailing list EE@lists.wikimedia.org https://lists.wikimedia.org/**mailman/listinfo/eehttps://lists.wikimedia.org/mailman/listinfo/ee