Thu

Dec 13
2007

Nat Torkington

Nat Torkington

Evaporative cooling of group beliefs

Over on Overcoming Bias there was a great post called "Evaporative cooling of group beliefs" where the author talks about how ejecting outliers moves the group's average position towards the other extreme. At the end he draws an interesting conclusion:

My own theory of Internet moderation is that you have to be willing to exclude trolls and spam to get a conversation going. You must even be willing to exclude kindly but technically uninformed folks from technical mailing lists if you want to get any work done. A genuinely open conversation on the Internet degenerates fast. It's the articulate trolls that you should be wary of ejecting, on this theory - they serve the hidden function of legitimizing less extreme disagreements. But you should not have so many articulate trolls that they begin arguing with each other, or begin to dominate conversations. If you have one person around who is the famous Guy Who Disagrees With Everything, anyone with a more reasonable, more moderate disagreement won't look like the sole nail sticking out. This theory of Internet moderation may not have served me too well in practice, so take it with a grain of salt.

It's interesting to compare this to the techniques Theresa Nielsen Hayden uses on her "Making Light" blog and on Boing Boing comments. There's an art to building online communities that nobody has yet well documented.


tags:   | comments: 9   | Sphere It
submit:

 
Previous  |  Next

0 TrackBacks

TrackBack URL for this entry: http://blogs.oreilly.com/cgi-bin/mt/mt-t.cgi/6118

Comments: 9

  judielaine [12.13.07 05:13 AM]

There's an art to building any community, whether online or off. I was fortunate in graduate school to have met a person who was able to bring together a community of spiritual people of different traditions. I basked in the discussion, sharing, and support of that community. Since then, participating in discussion groups in various fields and interests, i've realized that the organizer of that group in grad school was gifted as a "connector" and as moderator. I would guess her effectiveness was partly because she made an effort to get to know people as individuals and then use that to facilitate the conversation as a group. Nothing about that is linked to being offline. If a moderator or host wants to foster community, taking the time to find out more about the individuals would help the moderator to comfortably elicit participation. In the context of this theory, having outlying opinions -- and having a moderator who understands more about that person than the boundary pushing nature, might mean being able to invite discussion from that person on topics where they're more moderate. Or, it helps the moderator figure out that the troll is just someone who is uninterested in anything other than being a thorn.

  Michael H [12.13.07 05:18 AM]

I'm a fan of Slashdot's moderation system where comments are modded up or down. The moderators are randomly chosen and only given 5 mod points at a time. When those are gone, it might be awhile before they get them back.

There's more to it, including meta-moderation and user karma, but I think it's one of the best systems for sites with large, active userbases. It's a bit like checks and balances, because I think it would be very hard to game the system without incurring some sort of penalty.

  Greg Wilson [12.13.07 05:38 AM]

Nice post, but I have to take issue with your claim that no one has documented the process well: Derek Powazek's often-overlooked "Design for Community" (Waite Group, 2001, 0735710759) was, and is, a great look at the social mechanics.

  Mike Figueroa [12.13.07 07:56 AM]

Fortuito.us blog had a relevant post a while back:

http://fortuito.us/2007/05/some_community_tips_for_2007

  Tim O'Reilly [12.13.07 08:04 AM]

The role of conversation in moderating extremes and the danger of conversing only with like-minded people was the subject of Cass Sunstein's excellent book Republic.com, which explored the impact of self-selected groups (partly facilitated by the internet, as well as the explosion of choice on cable TV) on the rise of extremes in American politics.

  Sachin [12.13.07 08:35 AM]

The conclusion of Overcoming Bias is a good thumb rule, though the devil almost always lies in the details.

I agree with Greg above. Powazek's book was a good insightful book on online communities. But high time someone took another look at it.

  Thomas Lord [12.13.07 09:42 AM]

"There's an art to building online communities that nobody has yet well documented."

Well, there's all of literature. Oh, but you mean something like a short pamphlet for bloggers? A field guide to Internet wildlife? Perhaps with literary quotes under the chapter headings?

I think you are proposing something quite ugly, not an art, and not community. I'm sure I could write something like such a pamphlet -- but only as satire.

First, you divide everyone into stereotypes. It's amazing that, at this late date, you've already got words for three distinct kinds of trolls (regular, articulate, and pet ("Guy Who Knows Everything")). Your first step towards community, apparently, is to eliminate individuality and to reduce people to the simplest possible terms.

Second, you start looking for rules that will tell you who to exclude and who to goad based on which stereotype you think of them as. Your second step towards community, apparently, is to discriminate on the basis of stereotype and try to manipulate people you approve of into doing work on your behalf.

Well, sorry to be blunt but, what are you thinking?

Why not just think of people as human and deal with what comes rather than trying to put people in boxes and stack the boxes just so? What's the quest for "rules" or "art" here?

In forums I led, over the past few years, I noticed some of the very ugliness I see here: People started believing in "theories" about who was or wasn't a good contributor to an open source project, for example. Or another example: person X is accused of being a troll and posts a defense or clarification, person Y then posts a long rambling message attempting to give an objective proof of X's trollness.

This kind of thing was disasterous. If the discussion over theories of open source virtue and vice dominates, it surpresses conversation about, you know, the actual code. If the discussion is over competing theories of how to measure someone on the troll meter, otherwise fairly reasonable people quickly leap at each other's throats.

Adults. Adults did those things and they did them because if you tuned into Slashdot around that time, or some of the other blogosphere, people were starting to promote these "theories". (Tim might recall a certain late entry he put into that fight.)

It gets worse because once you've got shared theories of how best to stereotype people, now you've started labeling and those labels stick. That's another thing I've seen happen:

Person X shows up new to some forum and offers a seemingly pretty good idea. At the first sign of positive uptake, person Y chimes in with "No, ignore X. X used to be in that other group and is just a troll." Again: adults.

(Part of the reason people with influence like to latch on to these "theories" has to do with how power works -- how it is acquired and maintained. But, that's Machiavelli, not Wilder.)


Here is the art of building community:

Be honest and helpful yet reserved. Listen carefully and speak responsively. Share information -- not just what you came up with yourself but also pass things along and help the group get to know itself. The best reaction to being insulted is a belly laugh and then a relaxed puzzlement over exactly why. Exaggerate your good manners when dealing with people you don't know well. Be tolerant of unfamiliar manners. Never forget that as richly complex, operatic, and overwhelming human YOUR life is -- it's the same for the other guy.

And resist the Blue Meanie Spirit that would put us all in boxes to be sorted through by forum bosses.

-t

  Nat [12.14.07 03:46 PM]

Sometimes 'You are a troll' is a way to say "I can't find a correct answer, you may be right but I can't elaborate"

We may need a good "trollometer", an objective way to qualify a "troll" (disposing of it, then, is trivial). It may be "assertions or questions neglected by other participants".

  gnat [12.14.07 07:02 PM]

@Greg @Mike - thanks for the reading list. The mefi moderation post was great, and I'm hunting down Derek's book as we speak. Ok, so maybe it's not so undocumented after all :-)

Post A Comment:

 (please be patient, comments may take awhile to post)






Type the characters you see in the picture above.

RECOMMENDED FOR YOU

RECENT COMMENTS