Thanks for the note and provocative questions. I think assuming the lack of explicit communication disqualifies the activity as collaboration is a very narrow view of collaboration. Indeed, in the context of a breaking news event the talk pages is a poor approximation of the actual communication taking place. For example, the 2011 Tohoku earthquake and tsunami article at 8:00 UTC on March 11 covers substantially more information than is reflected on the talk page at the same time while the revision comments for this period are especially rich. My previous research suggests that power editors who dominate contributions to these article resort to more synchronous channels such as IRC to coordinate their work (paper) which bears out your argument that editors resort to other backchannels to coordinate this work.Moreover, a substantial amount of collaboration takes place through implicit work of reading revision comments and monitoring revision histories which leads to the development of develop shared mental models of the work to be done and transactive memory systems of who is doing what. Furthermore, theories of stigmergy suggests that the artifact (like the article) itself can encode consensus and coordinate subsequent action in distributed information work absent any explicit communication (link, link, link). The work that has already been done and still needs to be done is encoded within the object itself (e.g., broken template needing closing braces) without the need for any communication. I think this is absolutely collaboration.Walking down the street or asynchronously editing a document with several dozen other editors are examples of collaboration which demand collective mind and heedful interrelation. Borrowing from Weick and Roberts (1993): "actors construct their actions (don't run into other people while en route) understanding the system consists of connected actions by themselves and others (other people are trying to do the same thing) and interrelate their actions within the system (try to walk the same speed as the people around you)." This "stampede" may not require much in the way of higher cognition as with jointly writing a research paper, but it's absolutely the joint accomplishment of work towards a shared goal which is collaboration in my book.What the analysis attempts to capture at a very coarse level is not the happenstance interactions (the peripheral nodes and one-off links) but rather the emerging framework of "silent collaboration" as some editors have a consistent tendency to have many interactions, repeatedly interact, or are connected to particular kinds of other editors. Regarding the coarseness of the analysis, you would be right to point out that the fact that someone makes a change once after another user is almost certainly a spurious interaction: they could have been editing completely different parts of the article. Work by Aaron Halfaker (paper) and Jeff Rzeszotarski (paper) has each looked at content-level changes which is substantially more persuasive in this regard. However, the work editors do in the long-run is necessarily contingent on the work others have done before them and is what I believe emerges into the clusters of sustained interaction which my method captures (not just edit wars). To the (very limited) extent these interactions are the function of edit warring (the "Possibly related events" in early history of the 2011 Norway attacks is an excellent example of it), other editors intervene and also make changes. These subsequent and sustained dyadic interactions are potentially illustrative of important dynamics such as polarization, centralization, specialization, etc. which emerge from successive interactions.As you suggest, my post and much of my other research is absolutely missing a crucial qualitative component of what motivates these editors, how they make sense of others' contributions, and how they accomplish this work. Some forthcoming research of mine suggests that these articles are not one-off collaborations but there are editors who are effectively dedicated to coordinating and collaborating on breaking news articles. Further work in the domain should explore the extent to which this implicit work involves repeated repertoires of action (e.g., focused on reformatting references), emulating stylistic precedent/consensus (e.g., fill in the blanks from using previous hurricane articles), or employing specialized practices (e.g., adopting templates for death tolls). I think survey research of any defensible validity is immensely difficult to design and execute, so I avoid it at all costs, but that's my methodological bias :)I'm very happy to have people poke holes in my ideas now so that I can have snappier responses in front of conference audiences and dissertation committees!Best,BrianOn Sun, Jul 22, 2012 at 8:52 PM, Kerry Raymond <k.raymond@qut.edu.au> wrote:
I think this study of the “collaborative” dynamics is interesting, but I have some questions.
Do we have any evidence that collaboration is actually occurring? With breaking news like this, it may just be many individuals operating independently? Collaboration pretty much requires a communication channel, but internal to WP the only visible communication is the talk page (and perhaps user talk pages). We might infer that editors participating in a consensus-building thread in the talk page (or user talk pages) are acting collaboratively in relation to the issue under discussion (but not necessarily more widely. However, if editors are disagreeing in a talk page thread, it is hard to say whether their edits in relation to that issue are collaborative or “warring” (deliberating seeking to undo another) or simply independent (using their best judgement at that moment). Nor can we readily judge if editors not writing on the talk page might still be reading it and thus informing their actions based on those discussions – that is, might be acting in “silent collaboration”. Nor can we tell if any of the editors are having private conversations via email or other means . As communication takes time, in a breaking news situation editors might prefer to just “be bold” and keep the page as up-to-date as possible, using their own “best judgement” rather than “waste” time arguing on the talk page.
Can we consider reversions and mutual reversions in a “breaking news” situation as revealing an “edit war”? With many editors simultaneously active, I think you have to consider that it is just a stampede that is taking place. It’s a bit like “walking together”. If just two people walk down a street, we can say pretty clearly if they are walking together (they will remain in close alignment most of the time). But if a crowd of people are walking down the street, it’s hard to say that two people are walking together – they might just be forced into that alignment by the crowd. I think we have the same situation with reversions with simultaneous editors operating in a breaking news situation; many individuals acting independently and reversions might not be intentional.
From a research perspective, doing a survey or interview of some of the editors on their perspective of what was going on might be informative to provide better interpretation of the data. Given the protection/semi-protection of the page means it is probably possible to contact many of them via their user talk page. It would be very interesting to know if those who appear to be involved in an edit war saw it as an edit war, and to what extent they thought they were acting collaboratively and by what means was that collaboration fostered (e.g. explicit discussions on talk page, or more implicit, e.g. adopting a consistent style established by other editors).
Kerry
From: wiki-research-l-bounces@lists.wikimedia.org [mailto:wiki-research-l-bounces@lists.wikimedia.org] On Behalf Of Taha Yasseri
Sent: Monday, 23 July 2012 3:20 AM
To: Research into Wikimedia content and communities
Subject: Re: [Wiki-research-l] Wikipedia's response to 2012 Aurora shooting
I resend my previous message that is not delivered yet. Sorry for potential duplicate receiving .
Now, after two days, there are 30 Wikipedia language editions who have covered the event (have an article on it).
Here: http://wwm.phy.bme.hu/blog.html, see the dynamics, i.e. number of covering WPs versus time, measured in minutes and counted from the event time (t=0).
For those who are familiar with spreading phenomena, the curve comes as no surprise. What is surprising, is the fast reaction of Latvian (3rd place) and rather late reaction of Japanese Wikipedia (the latter is most likely related to time zone effects).
As I did this in a very unprofessional way, errors and miscalculations are expected, please notify if find.
bests,
.tahaOn Sun, Jul 22, 2012 at 7:06 PM, Dario Taraborelli <dtaraborelli@wikimedia.org> wrote:
Nice (and timely) work as usual, Brian. I was going to enable AFTv5 on this article but decided to hold off for a number of reason (most importantly the fact that we're slowly ramping up AFTv5 to enwiki and we're mostly focused on scalability at the moment). It'd be interesting to study how enabling reader feedback affects the collaborative dynamics of breaking news articles, especially semi-protected ones on which anonymous contributors don't have a voice.
Dario
On Jul 21, 2012, at 5:06 PM, WereSpielChequers wrote:
It is currently semiprotected, there were IP edits when it was first created. But according to the logs it was fully protected for a while due to IP vandalism. However the edit history only shows it going to semi protection, but there were some moves which have complicated things
WSC
On 21 July 2012 22:46, Taha Yasseri <taha.yaseri@gmail.com> wrote:
Ok! the page is protected. Sorry!
On Sat, Jul 21, 2012 at 11:43 PM, Taha Yasseri <taha.yaseri@gmail.com> wrote:
Thank you Brian,
Could you also plot the absolute number of edits, and editors, (instead of the ratio)? Though, since the data is ready I could do it on my own too!
Surprisingly I see no IP contribution to the article (or may be only few), not in accord with my expectation for such a topic.
cheers,
.TahaOn Sat, Jul 21, 2012 at 9:05 PM, Brian Keegan <bkeegan@northwestern.edu> wrote:
My preliminary analysis of (English) Wikipedia's response to the 2012 Aurora shootings. Data is available at the bottom:
http://www.brianckeegan.com/2012/07/2012-aurora-shootings/
--
Brian C. Keegan
Ph.D. Student - Media, Technology, & Society
School of Communication, Northwestern University
Science of Networks in Communities, Laboratory for Collaborative Technology
_______________________________________________
Wiki-research-l mailing list
Wiki-research-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l
--
Taha.
--
Taha.
_______________________________________________
Wiki-research-l mailing list
Wiki-research-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l
_______________________________________________
Wiki-research-l mailing list
Wiki-research-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l
_______________________________________________
Wiki-research-l mailing list
Wiki-research-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l
--
Taha.
_______________________________________________
Wiki-research-l mailing list
Wiki-research-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l
--
Brian C. Keegan
Ph.D. Student - Media, Technology, & Society
School of Communication, Northwestern University
Science of Networks in Communities, Laboratory for Collaborative Technology
_______________________________________________
Wiki-research-l mailing list
Wiki-research-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l