« This year's "Basketbrawl"? | Main | Co-Authorship Models »

Wednesday, December 13, 2006

Citing Wikipedia

I have occasionally had students submit draft seminar papers that cite to Wikipedia, the collaborative internet encyclopedia founded in 2001.  I now warn students against the practice (unless they are citing a point that is about Wikipedia and its contents rather than a use of Wikipedia to support the truth of an underlying assertion made therein). 

Some students are surprised, but the reasons should be obvious.  You wouldn't cite to a standard encyclopedia, so you shouldn't cite to an internet encyclopedia for the same set of reasons.  Additionally, (1) Wikipedia authors may have no expertise in the subjects on which they write, (2) Wikipedia is constantly changing and so it's harder to present all seekers with a consistent copy of the source, and (3) Wikipedia sometimes has intentionally false or malicious information.  All of this is true, even though I realize that in head-to-head tests, Wikipedia has compared rather well to Britannica. 

One of the easiest ways to make the point though is this: If I want to make an assertion about, say, the size of the population in the United States, I could just go on to Wikipedia, add the relevant information, and then I have my citation.  Clearly, such a citation adds nothing beyond my bald assertion of the claim.  I'll be interested in the comments if people want to add other student research pitfalls to avoid.

Posted by Adam Kolber on December 13, 2006 at 11:01 AM | Permalink

TrackBack

TrackBack URL for this entry:
https://www.typepad.com/services/trackback/6a00d8341c6a7953ef00d835040e9969e2

Listed below are links to weblogs that reference Citing Wikipedia:

Comments

Somewhat late to the party...When a court-appointed special master last year rejected the claim of an Alabama couple that their daughter had suffered seizures after a vaccination, she explained her decision in part by referring to material from articles in Wikipedia, the collaborative online encyclopedia. The reaction from the court above her, the United States Court of Federal Claims, was direct: the materials “culled from the Internet do not — at least on their face — meet” standards of reliability. The court reversed her decision ... More than 100 judicial rulings have relied on Wikipedia, beginning in 2004, including 13 from circuit courts of appeal, one step below the Supreme Court.

Posted by: Simon | Jan 29, 2007 9:58:08 AM

I like Andrew's point. If the proposition is something one would consider supporting with a citation referencing Wikipedia, it's appropriate to ask whether the proposition requires a citation at all.

Expertise matters because that's what gives us some reason to believe the material is accurate (unless there are additional citations). Perhaps the collaborative process makes up for this potential lack of expertise. But if Wikipedia is cited at a particular point in time, it's not clear if the collaborative process has caught up yet.

Shouldn't whether one should trust Wikipedia's collaborative process turn on how accurate Wikipedia proves to be in general, across many topics? The "at a particular point in time argument" is too strong, in that would rule out crediting Wikipedia, even after a hundred years and statistical demonstration that it's authoratative with 99.99999% reliability. One can quite legitimately say that Wikipedia now is not in general reliable enough to cite, but one should be skeptical of a priori arguments that a collaborative process can never be trusted.

Re(2): Yes, I agree you can cite a particular version of Wikipedia, so that everyone knows precisely what source you're using. This can be a somewhat ugly citation in print media, however. Also, if I decide to proceed to use that same source, I will likely want to spend some time comparing the old version to the current ones. But, basically, I agree with you on this point.

As someone who writes legal scholarship about the Internet, I inflict a great many ugly citations on the reading public. Cleaning up versioning citations for Internet sources in general is going to be one of those monumental tasks facing future Bluebook editors.

Re(3): Because one can edit anonymously on Wikipedia, there is often no reputational interest to prevent one from posting intentionally false information. Maybe you're right that this is statistically unlikely for any given piece of information, but it is also unlikely that this will occur in a peer-reviewed journal. So I guess we'd need some more empirical analysis to resolve the issue.

One other thing to note on this point is that anonymous edits are devalued within Wikipedia and are comparatively less trusted. Credibility comes in part from a proven record of engagement with the community and with productive edits in general. The biggest danger with these sneaky anonymous edits is when they're submarined--snuck into otherwise-obscure articles.

Posted by: James Grimmelmann | Dec 13, 2006 7:57:31 PM

I'm a big two spaces after period believer. I think it looks much better.

Forget "cite/cite to" - I can't bear "I graduated college in 2003".

Posted by: anon | Dec 13, 2006 6:19:26 PM

You might take a look at Alan Liu's guideline (for students) on appropriate uses of Wikipedia, which I think gets all the important points across:
http://kairosnews.org/developing-a-wikipedia-research-policy

One way to teach students why Wikipedia is not appropriate, while at the same time giving them some extra incentive to do good work, is to have them write their seminar papers FOR Wikipedia. Of course, the policy of "no original research" means you have to either have a separate portion of the assignment for the types of individual analysis usually expected in seminar papers, or modify what you expect from the assignment.

I just completed such a project with my students:
http://en.wikipedia.org/wiki/User:Ragesoss/HIST_236

Posted by: Sage | Dec 13, 2006 5:33:40 PM

I don't know about student papers. At my law review, however, I have noticed more and more faculty pieces citing to Wikipedia. They seem, however, to prove either very broad propositions ("Deontologists focus on duties and rights to provide content to moral obligations") or items of general knowledge ("the Second Battle of Bull Run was fought between the forces of General Lee and General Pope")--things that, in any other discipline, would not require a citation at all. In that respect, they seem to follow Ubertrout's advice.

They don't seem to be particularly offensive on the printed page, either--the Bluebook, as well as most other citation systems, requires some kind of "Last Accessed" entry for web pages, along with a date. This actually makes sense in the case of a wiki or other system where you can see what a page looked like at any point in the past.

Posted by: Andrew | Dec 13, 2006 5:30:17 PM

Ubertrout: i suppose it depends on what you mean by the "proper source." the house or senate journal would be definitive, of course, and therefore necessary in, say, a disertation -- when one purpose is to demonstrate one's scholarly chops. for most purposes, however, i think that wikipedia would suffice.

Prawfette: correction accepted (though i think that "cite" is sometimes used as shorthand for "provide a citation," in which case the preposition would seem okay). most importantly, it seems that you are not a prescriptivist as feared by dan markel. if you were a prescriptivist, you would have said, "most important."

Posted by: steve lubet | Dec 13, 2006 3:37:22 PM

You mentioned a "set of reasons" for not citing standard encyclopedias and some additional reasons for not citing internet encyclopedias. The concerns about internet encyclopedias are well-taken, but what are the unstated reasons for not citing, say, the Encyclopedia Britannica? I know that we are not supposed to, but I have never heard a good reason for why we are not supposed to. For basic facts, they are great resources. What's the problem with them?

Posted by: Crito | Dec 13, 2006 3:22:51 PM

Prawfette, I wonder if you've just triggered a war on prescriptive vs descriptive usage for cite/cite to. Additionally, I wonder if Wikipedia will now agree with the Word Court from the Atlantic's recent column: I just learned that there's supposed to be only one space after periods when one types now. All along, and still today mostly, I have been using two.

Posted by: Dan Markel | Dec 13, 2006 3:22:06 PM

most importantly, you shouldn't "cite to" anything. You should only cite things. Review of transitive and intransitive verbs, folks.

Posted by: prawfette | Dec 13, 2006 3:11:17 PM

I agree with Prof. Lubet that whether one should cite Wikipedia depends in large part for what is being cited. However, I disagree that any detailed information should be cited from Wikipedia. I'd think that for Harrison's dates as Governor of the Indiana Territory, the proper source is the House or Senate Journal (conveniently available online via LOC from pre-1877). However, if you wanted to give a general and quick link to a bio of Harrison, without using the facts therein on your own, that would be ok. Just my opinion, but probably best practices.
I'd note seperately that for concepts or people mentioned in passing, a wikipedia link is probably a good idea if a quick bio in the note is unnecesary, especially since more and more of these articles are being read online.

Posted by: Ubertrout | Dec 13, 2006 2:59:28 PM

i see no harm in citing to wikipedia for readily ascertainable facts; one might even conisder it the source of choice.

for example, earlier today i wanted to find out the dates of william henry harrison's term as governor of the indiana territory (relevant to my current project, not just idle curiousity). wikipedia made that a snap, and there is no reason to think it's less reliable than other sources, so i would not be embarrassed to drop a footnote to wikipedia in an article.

on the other hand, i would not rely on wikipedia (or any similar or no-so-similar encyclopedia or website) for something more nuanced or interpretive, say, harrisons views on slavery in the territories. in that situation, expertise truly matters.

sl

Posted by: steve lubet | Dec 13, 2006 2:44:32 PM

Great comments!

James,
Re(1): Expertise matters because that's what gives us some reason to believe the material is accurate (unless there are additional citations). Perhaps the collaborative process makes up for this potential lack of expertise. But if Wikipedia is cited at a particular point in time, it's not clear if the collaborative process has caught up yet.

Re(2): Yes, I agree you can cite a particular version of Wikipedia, so that everyone knows precisely what source you're using. This can be a somewhat ugly citation in print media, however. Also, if I decide to proceed to use that same source, I will likely want to spend some time comparing the old version to the current ones. But, basically, I agree with you on this point.

Re(3): Because one can edit anonymously on Wikipedia, there is often no reputational interest to prevent one from posting intentionally false information. Maybe you're right that this is statistically unlikely for any given piece of information, but it is also unlikely that this will occur in a peer-reviewed journal. So I guess we'd need some more empirical analysis to resolve the issue.

Aaron,
My post is written in 2006. In five years time, we may have enough evidence of Wikipedia's reliability to change the scholarly convention. If so, my views may well change.

Leo,
In that example, I wasn't supposing that the author was doing something intentionally inappropriate. He simply believes he knows a certain fact and has entered it into Wikipedia. Then, why is it unethical to cite to one's own point in Wikipedia but acceptable to cite to someone else's? Fact X's being in Wikipedia just means that one person decided to add X to the system. Perhaps fact X's remaining in Wikipedia over a period of time is the critical factor. If so, then it doesn't matter who entered it into Wikipedia.

Posted by: Adam Kolber | Dec 13, 2006 2:29:46 PM

It's unethical to cite Wikipedia if you're not citing it, but instead citing yourself through it. It's also unethical to quote selectively from a book in a misrepresentative way. A draconian ban on Wikipedia is not the solution to possible individual abuses.

Posted by: Leo | Dec 13, 2006 12:36:01 PM

Adam, posts like these make you sound like a luddite. In five years, you will look upon this posting, shake your head, and realize how clueless you were back in old-2006. The power of open-content information is just beginning and will shift our paradigm with regards to "expert" produced information.

The power and pervasiveness of open-content information has just begun. Wikipedia is just a tip of large iceberg that will become a cornerstone of the internet.

Posted by: Aaron | Dec 13, 2006 12:28:36 PM

I would also note the response to the I could just go on to Wikipedia, add the relevant information, and then I have my citation point is twofold. First, you'd be surprised at how quickly and how often other Wikipedia editors find and remove astroturfed content. Second, anyone reading the paper could go back and check the history of the page on that point, to see who added the information and when. In that sense, Wikipedia is more auditable than many other sources.

Posted by: James Grimmelmann | Dec 13, 2006 12:16:30 PM

On (1), why does the particular expertise of the author matter if what the article says is accurate and useful? We could look to that expertise as one factor in deciding whether the information is of high-quality, but Wikipedia also builds quality both through its collaborative process. That is, a thousand people each adding a little bit to an article on a topic they know a little about can aggregate to a quite useful and knowledgable article, given the collaborative editing process.

(2) can be adressed by citing to a particular revision of the article. (e.g. here is the entry for Germany, as of 4:39, 13 December 2006).

And as for (3), all information sources sometimes contain intentionally false or malicious information. The question is how often, and Wikipedia's overall statistical properties are quite good. It is very unlikely that any particular cited point is intentionally false or malicious.

I would still dicourage citation to Wikipedia, but principally because citing to Wikipedia is typically a sign that the author has not bothered to delve deeper into more specific and comprehensive sources on the topic.

Posted by: James Grimmelmann | Dec 13, 2006 12:12:36 PM

I agree. I like Wikipedia very much, but I'm nervous about its use in scholarship. My own blog is a Wiki-free zone.

Posted by: Daniel Goldberg | Dec 13, 2006 12:10:10 PM

The comments to this entry are closed.