[Air-l] Wikipedia warning -- Wikipedia is not a reliable information source

Judd Antin jantin at SIMS.Berkeley.EDU
Sun Dec 4 11:46:39 PST 2005

I know, as Elijah mentioned, that issues surrounding quality of 
information on Wikipedia have been widely discussed, but I wanted to 
offer a few comments:

- Wikipedia handles information quality with brute force. It assumes 
that many eyeballs make good information. Whether or not this is true in 
the end, the question is whether it is true at the moment that one might 
access the article. Has the quality of the information been recently 
compromised? And if so, has ample time passed for inaccuracies to be 

- Compounding the above point is the fact that Wikipedia's quality 
assurance mechanism only works on a macro-level. We have little 
information about whether individual articles have been properly vetted 
by many users at a specific point. We end up deciding in general whether 
we believe the mechanism works or not.

- One common tool that people use to make decisions about information 
quality is authorship. But in the name of promoting the 
commons-authorship model, Wikipedia actually obscures the notion of the 
author at an individual level. If you wanted to find out who wrote an 
individual piece an article, there would be no reasonable way to do 
that. Though MediaWiki keeps detailed page histories, wading through 
them, especially for articles that have 1000s of edits, is overwhelming 
if not impossible.

Although there are certainly problems with Wikipedia, I do not agree 
with Ken's assessment that it should be banned as a reference. Wikipedia 
puts its community of users in charge, and so we should not assume that 
any action on our part will result in their instituting some mechanism 
of control. Rather we should take it upon ourselves to promote both the 
production of high quality information in Wikipedia and its responsible 
use. I do agree, however, that Wikipedia should rarely - or never - be 
the only reference. For my own part, I often find that Wikipedia is a 
valuable place to start exploration, especially in areas that I know 
little about. Wikipedia articles often provide me with the background 
and context I need to begin exploration. This, IMHO, is the type of use 
we should promote.

Finally, I want to advertise that we are working on a software-tool that 
will allow users to navigate Wikipedia based on the notion of trusted 
authors through a variety of interactive tools and visualizations. Its 
core is an algorithm that sifts through a page's revision history to 
assign an author to each line of text in the current revision of an 
article. History Flow (http://researchweb.watson.ibm.com/history/) is 
another wonderful tool that works on a similar principal.

We would welcome any suggestions for functionality or applications of 
this tool. Once we've reached a reasonable point in the implementation, 
we will also open source the tool.



--Judd Antin
School of Information Management & Systems (SIMS)
University of California Berkeley
jantin at sims.berkeley.edu
blog: http://technotaste.com/blog

Ken Friedman wrote:
> Dear Elijah,
> I am trying to stir up some action that may lead to a solution.
> You are mistaken in the idea that these issues are evident to most
> Wikipedia users. These problems are evident to people on this list --
> but far too few students recognize the need to use multiple sources.
> No matter how much I encourage people to use multiple sources
> for ALL facts, something like 40% of my students use only one
> source for most facts, and that source is increasingly Wikipedia.
> My larger point is that if students are satisfied with only one source,
> I want that source to be Encyclopedia Britannica or a similar
> excellent reference source.
> I'd prefer that more students checked issues and facts against
> many sources. Until this becomes a common habit of mind, I am
> no longer willing to accept Wikipedia. My second large point is
> that if enough of us take the same stand, Wikipedia will necessarily
> find a way to do better fact-checking and to prevent egregious
> cases of libel, defamation, hoax information, and fraud.
> Yours,
> Ken
>>What's your larger point?  Or are you just trying to stir things up?
>>These are issues that have been widely discussed.  And largely 
>>things that are very evident to users of wikipedia... and often to 

More information about the Air-L mailing list