Get your chatfilter at Answerbag July 9, 2010 7:52 AM   Subscribe

AskMe is one of three Q&A sites studied in this new academic article.

I actually saw a version of this paper presented at the Association of Internet Researchers conference last fall. It's not really in this article, but in their presentation they were trying to speculate why the different sites had different distributions of question types, and so I talked to them a little about the culture of AskMe and the rules against chatfilter.

I would also point out their finding that AskMe has both the longest average question length and the most answers on average. We win(?)
posted by DiscourseMarker to MetaFilter-Related at 7:52 AM (48 comments total) 8 users marked this as a favorite

Do we win?
posted by smackfu at 7:53 AM on July 9, 2010


(I may have posted that comment before I got to the end of the post.)
posted by smackfu at 7:54 AM on July 9, 2010


Yahoo also produced the only two questions that the coders rated as “not a question”: The full text of these questions is: “myspace layouts?” and “heavy metal music? I am doing a project about it and need info.”

Heh.
posted by Horace Rumpole at 8:05 AM on July 9, 2010


We always win.
posted by rtha at 8:33 AM on July 9, 2010


I believe chatfilter is both the second and third type in their taxonomy. Is that right?

If you look at Figure 2, you can see the low occurrence of chatfilter with regards to the other sites.
posted by vacapinta at 8:40 AM on July 9, 2010


The categories assigned to the questions were neat. Of course "advice" dominates in terms of length.
posted by Phire at 8:42 AM on July 9, 2010


yay!
posted by mathowie (staff) at 8:43 AM on July 9, 2010


We can now make it official:

AskMeFi: Scientifically proven to be better than other leading Q&A sites.
posted by Salvor Hardin at 8:44 AM on July 9, 2010 [1 favorite]


I really need to collect all the Mefi-related research papers that have been published into a nice binder to put on my bookshelf.
posted by cortex (staff) at 8:52 AM on July 9, 2010


Someone should phrase an AskMefi question in such a way that it breaks their methodology. Like a very specific Godel sentence.
posted by griphus at 8:55 AM on July 9, 2010 [2 favorites]


cortex, can you at least collate them into some links for now?
posted by carsonb at 9:07 AM on July 9, 2010


I really need to collect all the Mefi-related research papers that have been published into a nice binder to put on my bookshelf.

What a dork.

Oh well, back to working on a web site of all the Civil War Monuments in my state. I sure hope my package of plastic army men (and a really cool tank!) shows up today in the mail!
posted by marxchivist at 9:11 AM on July 9, 2010 [7 favorites]


"Advice" and "Identification" are both more Relationshipfilter than chatfilter. The middle two types - "appproval" and "quality" seem closer to chatfilter questions to me.

I'm so excited that I now have a term to use when talking about how the responses to a question help people other than the OP by being retrievable in future searches: "Archival value."

Thanks for posting this - I only skimmed, but quite neato.
posted by Miko at 9:11 AM on July 9, 2010


cortex, can you at least collate them into some links for now?

Good wiki project...
posted by Miko at 9:12 AM on July 9, 2010 [1 favorite]

What’s the next band you want to see get a Rock Band “special edition”? I wish they would do the Ramones. I would want to be Dee–Dee. Who would you be?
I haven't posted on the green for quite a while, but I'm so tempted.
posted by Wolfdog at 9:30 AM on July 9, 2010 [1 favorite]


I feel like a better comparison would be AskMe v. Reddit's AMA v. Something Awful's Ask/Tell.

Make it happen, internet.
posted by reductiondesign at 9:57 AM on July 9, 2010


"Advice" and "Identification" are both more Relationshipfilter than chatfilter. The middle two types - "appproval" and "quality" seem closer to chatfilter questions to me.

Advice is definitely Relationship filter, but I think that Identification, certainly as they've phrased the sample question, would get flagged as chatfilter more often than not. I mean seriously, would you not flag
"What’s the next band you want to see get a Rock Band “special edition”? I wish they would do the Ramones. I would want to be Dee–Dee. Who would you be?"
as chatfilter? I certainly would. What's the problem to be solved here?

I think the (Dis)Approval category is pretty clearly the type of question not allowed on AskMe (and I told them as much in the presentation).

I think the Quality category is borderline--the question they have as an example seems clearly chatfilter-y, but there are frequently questions along the lines of "I need a new laptop/smartphone/web host/etc. Given that I want to do X, Y and Z, what is the best choice?" I would imagine that question being coded in their taxonomy as Quality, but it seems perfectly legitimate for AskMe, since there is enough detail to solve a specific problem.
posted by DiscourseMarker at 10:10 AM on July 9, 2010 [1 favorite]


Metafilter: (I may have posted that comment before I got to the end of the post.)
posted by dersins at 10:12 AM on July 9, 2010 [5 favorites]


Yeah, I guess I read identification wrong. I was looking more at the purpose:

Directed at establishing a focused discussion (and potentially building relationships) among people with a shared commitment to a topic.

than the example, which I would definitely flag as chat. I guess an "identification" question is more about finding the people to identify with than about the focused discussion.
posted by Miko at 10:14 AM on July 9, 2010


I haven't posted on the green for quite a while, but I'm so tempted.

Is there an easter egg in rock band that does whatever that question was?
posted by fixedgear at 10:16 AM on July 9, 2010


Good wiki project...

Definitely, yeah. The best thing in the world would be to be able to actually collect those papers for access by interested mefites, but there may not be any practical way to do that in a publicly-hosted manner.

I'll try to collect the references I know of at some point here. There are four or five papers I'm aware of (at least one of which isn't actually out yet), they've all had metatalk posts at some point I think.

I gave JSTOR a quick search for the hell of it and didn't turn up any of the papers I had in mind, but I did find a bunch of references to (mostly language-related) journal articles that mentioned mefi—but it was all behind additional paywalls, so, feh. What I could manage in terms of excerpts suggested it was mostly footnote references and such, but it'd still be neat to be able to put that all together in one place. Anybody who has really great journal access, lemme know, I'd love to go hunting.
posted by cortex (staff) at 10:17 AM on July 9, 2010


There's also this paper, which was also presented at AoIR in the fall. And them seem to be offering it for free right now! (full text)
posted by DiscourseMarker at 10:54 AM on July 9, 2010 [1 favorite]


I have fairly decent journal access, I think. At work now, ping me.
posted by carsonb at 10:55 AM on July 9, 2010


Pinged!
posted by cortex (staff) at 11:10 AM on July 9, 2010


> There's also this paper, which was also presented at AoIR in the fall.

From that paper's meetup list:

most attendees                          most meetups
city   number of attendees    city   number of meetups

New York, NY      28                          New York, NY      49
San Francisco, CA 27                        San Francisco, CA 22

NYC rules!
posted by languagehat at 11:23 AM on July 9, 2010


Oooh! Can't wait to read this article. I would also love to help out or contribute to any MeFi research/paper collection or presentation projects, however they develop. I have a lot of time and interest to devote to this. :)
posted by iamkimiam at 11:27 AM on July 9, 2010


As my lifelong hero would say . . . "Fascinating."
posted by bearwife at 11:31 AM on July 9, 2010 [1 favorite]


The end result of all that parsing of questions seems to be what they consider to be a 'better taxonomy' to allow sites like AskMe to more efficiently and uniformly categorize questions. Isn't this solving a problem that sites don't necessarily want solved (or at least not necessarily solved in this fashion)?
posted by Hardcore Poser at 11:43 AM on July 9, 2010


There's also this paper, which was also presented at AoIR in the fall.

Here are the other references from that paper:

Silva, L., Goel, L. & Mousavidin, E. (2009) ‘Exploring the dynamics of blog communities: the case of MetaFilter’, Information Systems Journal, vol. 19, no. 1, pp. 55–81.

Ali-Hasan, N. (2005) ‘MetaFilter: an analysis of a community weblog’, [Online] Available at: http://www.noor.bz/pdf/alihasan_metafilter.pdf (1 April 2009).

Lawton, P. (2005) Capital and Stratification within Online Community: A Case Study of Metafilter.com. Master’s thesis, University of Lethbridge, [Online] Available at: http://www.metafilter.com/trash/Capital_and_Stratification.pdf (31 August
2009).
posted by smackfu at 11:48 AM on July 9, 2010 [2 favorites]


From that article:

"One attendee based in New York City attended 33 meetups."
posted by smackfu at 11:51 AM on July 9, 2010


Social question and answer (Q&A) Web sites field a remarkable variety of questions: while one user seeks highly technical information, another looks to start a social exchange. Submissions from the latter are deleted.

Wow, they definitely have us down.
posted by Durn Bronzefist at 11:57 AM on July 9, 2010


The end result of all that parsing of questions seems to be what they consider to be a 'better taxonomy' to allow sites like AskMe to more efficiently and uniformly categorize questions. Isn't this solving a problem that sites don't necessarily want solved (or at least not necessarily solved in this fashion)?

Their stated goal in the paper, and what I remember from their presentation, was to try to help information systems designers design "better" Q&A sites. I really have no idea if there are, in fact, designers out there trying to build a better Q&A mousetrap, but there are computer scientists on the research team, so presumably someone thought this research might be useful.

Also, their acknowledgments indicate that this research was attached to 2 NSF grants, so maybe it's the government who cares.
posted by DiscourseMarker at 12:34 PM on July 9, 2010


I'm in ur jurnls taking ur tax dlrs heh amirite stfu
posted by norm at 12:38 PM on July 9, 2010 [1 favorite]


Lawton, P. (2005) Capital and Stratification within Online Community: A Case Study of Metafilter.com.

Yeah, this is the one for me, even though it's old. I love this sort of thing (probably why I'm a sociology major). Can't wait to read it.
posted by Danila at 12:53 PM on July 9, 2010


> "One attendee based in New York City attended 33 meetups."

Gotta be jonmc.
posted by languagehat at 1:47 PM on July 9, 2010 [1 favorite]


Yeah, this is the one for me, even though it's old. I love this sort of thing (probably why I'm a sociology major). Can't wait to read it.

That one is hosted locally.
posted by carsonb at 2:40 PM on July 9, 2010


I have great journal access. Memail me if you need more help with finding stuff.
posted by emilyd22222 at 2:41 PM on July 9, 2010


The meetup paper is bullshit. We've had over 20 meetups in Chicago. Hell, we've had over 10 this year. And we've definitely had over 30 attendees on more than one occasion: the block party alone cracked 40 or so.
posted by eamondaly at 2:55 PM on July 9, 2010


And we've definitely had over 30 attendees on more than one occasion: the block party alone cracked 40 or so.

Yeah, and the Portland 10th had 60-some, although admittedly many were from out of town.
posted by dersins at 3:58 PM on July 9, 2010


And we've definitely had over 30 attendees on more than one occasion: the block party alone cracked 40 or so.

Yeah, and the Portland 10th had 60-some, although admittedly many were from out of town.


The meetups paper does not include the 10th anniversary. From the article:
I draw on approximately eight and a half years of longitudinal user activity data from the online community MetaFilter (MetaFilter.com), from the community's first post on 14 July 1999 through 31 December 2007
And anyway, we had around 70 people show up at one point or another at the 10th party in Boston, so nyah.
posted by DiscourseMarker at 4:27 PM on July 9, 2010


I'm still not sure I like the idea of bees answering questions, but I like to watch them dance while they're answering them.
posted by not_on_display at 7:42 PM on July 9, 2010 [1 favorite]


"AnswerBag" as a site name always makes me LOL when I read it; that combination of words has such a uneducated, backwater flavor in my head that it makes automatically doubt the quality of answers one would get there. It's like naming a restaurant "FeedBag."

For the record, I have a similar reaction to "Dress Barn."
posted by Nattie at 1:50 AM on July 10, 2010 [2 favorites]


Many years ago, when I was doing some of my earliest field research on blue collar communities in the US south, I was sitting in a bar talking to a couple of retired redneck guys (who later became friends of mine) about what I was doing.

One, who'd seen me before, asked "You still on that RE-search shit? What, you ain't found enough assholes around here YET?"

His friend chimed in "We're just BUGS IN A JAR to you!"

Seemed appropriate to this situation.
posted by fourcheesemac at 5:06 AM on July 10, 2010 [3 favorites]

A study of how question askers select “best answers” found that users evaluate answers differently depending on the type of the question (Kim, et al., 2007). For instance, users who ask information–seeking questions appear to value the answer’s clarity and accuracy, while users who ask for opinions value factors such as agreement with the answer or the presence of emotional support.
Yessss!!! Watch my green check marks stack up!!

For the record, I have a similar reaction to "Dress Barn."

Gaa, I hate the name "Dress Barn"-- it makes us sound like we are all cows... not that I shop there; it is Polyester Central. I also have an aversion to "Shoe Carnival" because that is just ridiculous. Shoes and Clowns together. Tccch.
posted by Secret Life of Gravy at 5:56 AM on July 10, 2010


The issue isn't whether or not a meetup had 30 attendees, though, is it? It's that someone attended 30 meetups. That's pretty damn impressive.
posted by Phire at 10:29 AM on July 10, 2010


I have been to maybe ten or twelve so far, and I'm a relative noob! Now, Eideteker--he has been to every single meetup.
posted by not_on_display at 2:27 PM on July 10, 2010


Yeah, I've got flickr photosets for thirty different meetups in my photostream and I'm not super diligent about remembering to take photos so that's lowballing it. A big dose of that was from the whole Meeting Metafilter thing last fall, but even minus that I've probably done 20+.
posted by cortex (staff) at 3:11 PM on July 10, 2010


"I would also point out their finding that AskMe has both the longest average question length and the most answers on average. We win(?)"

In conclusion, go MetaFilter.
posted by NortonDC at 9:01 AM on July 11, 2010


« Older Sports- WTF   |   lame delete Newer »

You are not logged in, either login or create an account to post comments