Wikipedia’s long tail
I recently read someone (but now I’ve lost the reference! [Found it, thanks to a commenter: It’s Peter Morville.) talking about Wikipedia‘s long tail, by which the person meant (I think) the articles that are rarely read or edited. The question was whether the quality of those articles is as good as that of the oft-edited ones.
It’s a great question. The obvious guess is that, if you believe the wisdom of crowds saves Wikipedia from inaccuracy and mediocrity, then long tail articles ought not to be as good as head of the tail articles. But I find little about Wikipedia conforms to my expectations. In conclusion: I have no guess.
Jimmy Wales has apparently said that research he did in December 2004 showed that about half the edits done by logged-in users are done by 2.5% of logged-in users. (Ironically, I’m having difficulty finding an authenticated source for this stat.) Do we know what the curve of traffic per article or edits per article looks like? (This should probably exclude stubs.) Is it a power law, as would be expected? If so, is there any way to check the quality (accuracy, relevance, completeness, neutrality?) of the long tail articles?
Just how good and/or useful is Wikipedia’s long tail?
(By the way, the Wikipedia article “Why Wikipedia is not so great” is a pretty great example of a Wikipedia article. :) [Tags: wikipedia longtail]
Categories: Uncategorized dw
Here are the numbers I heard from the horse’s mouth (Jimmy Wales speaking at a library meeting) last Friday:
* 50% of all Wikipedia edits are done by just 0.7% of users (615 people)
* 1.8% of users (1500 people) have written more than 72%
so that “500-1500 people wrote the English Wikipedia”. Hmm, that second stat looks a bit odd, but that’s what I wrote down.
Though I’m not sure what time period these stats are taken from.
I don’t know where I’m counted in those percentage numbers, but I’m a Wikipedia contributor. For the most part, I live in the long tail. I’ve picked a few pages where I have knowledge, I’ve written or rewritten the main sections, I’ve added appendices, and I monitor them on my watchlist. If other niche editors work like I do, then there’s no reason that often-read articles will be more accurate then our pet projects.
http://tagsonomy.com/index.php/peter-morville-the-tagsonomy-interview/
Peter Morville mentions his distrust in the long tail of Wikipedia.
Aha! Thanks, Will. And I blogged that Peter Morville interview yesterday, so I really should have remembered.
I’m a pretty active wikipedian, there are so many types of articles that would fit into what I would call the long tail. The kind Jim talks about above that are edited by a few authors that know about the subject are honestly some of the best in wikipedia I think, as long as they are not about anything controversial. Usually they aren’t though, or they would have more editors. They don’t get a lot of traffic from pageviews, but they have a few editors, and are high quality.
The other kind of long tail are the single author pages, written usually by an anonymous contributor who is never heard from again. These are usually about very obscure generally un-useful topics, thankfully, but are usually quite bad, in terms of formatting, naming conventions etc.
In my experience the first type is the most common, which is good for wikipedia. And even in the second type the information is usually factual, just presented poorly.
Those generalizations don’t hold up for controversial topics. There has been some discussion for sorting/checking all the articles that have only been edited by one person, but I don’t think it’s implemented yet.
I prefer Oondi
With oondi nobody can edit your hard work and at least you get something in return (100% of advertisement profits).
See http://www.ccc.de/congress/2004/fahrplan/event/59
for wikipedia data, including the 2.5% stat.
Cheers.