Bryan Caplan  

Optimal Citation in the Google Age

PRINT
Tom Strong, Compelling Coasean... There's No Evidence Like No Ev...

One of the best examples of path-dependence is the format of academic citations. Who cares what city a book was published in? Who ever did? But in the google age, tradition is even sillier. Volume numbers? Page numbers? These days, if someone wants to get a citation, he'll probably just google the article name.

Challenge: If you had the power to set the standard for citations in the google age, what would it be?

My nomination: last name of author(s), title, year. Thus, my book would be: Caplan, *Myth of the Rational Voter*, 2007. For journals, you'd show author(s), article title, journal name, and year.

My rationale: Journal name and year actually provide useful information for readers who don't want to google. Year gives historical context, journal name gives a quality signal. The rest is noise, no?


Comments and Sharing





COMMENTS (14 to date)
Barkley Rosser writes:

Well, there is a lot of stuff, including even sme journal articles, that are not available by googling. Are you able to get at chapters in edited volumes easily online? Not alwasys, especially from older books.

Blackadder writes:

Perhaps it's because I'm an attorney and thus used to it on Westlaw, but searching by volume and page number is quite useful as it requires less typing per search than if you had to type in the title.

LemmusLemmus writes:

I have long wondered why the city a book was published would be of any interest - even before the "google age". The American Psychological Association's citation code even says you need to include the state/country except if we're dealing with a "known publishing city", such as Stockhom (er?).

However, in the case of books I'd include edition given that later editions often differ from earlier ones and sometimes even include a section along the lines of "refutation of criticizms".

In the case of journal articles, I would include the number of the starting page, simply for faciliating finding it; I would also include the number of the finishing page to give people an idea of how much money they have to spend if they want to photocopy the article.

Stuart Buck writes:

For books, the name of the publishing house could signal quality ("Harvard University Press" beats "Lulu"). But city of publisher is pointless.

Michael Makowsky writes:

There is also a redundancy here that secures the citation against noise over second and third hand citations. There are plenty of articles I have found despite misspellings because of volume and year information. Plus in the grand scheme of a google library of knowledge, the more information to sort, organize, and interpret the annals of scholarly history, the better.

Another way of putting it: information is cheaper than ever to record and store by orders of magnitude. Why would we want to start economizing on it?

Nathan Whitehead writes:

In computer science I think we're ahead of the curve. The standard in CS is for authors to put full PDF preprints on their webpage. To get articles written in the Google age, you just google the name and pick out the link from the persons webpage.

The next step that is becoming more popular: embedded PDF links within papers. You see a citation in an online paper, you click it, and up pops your browser with the "official" citation page (usually from the ACM, IEEE, or Springer). Who needs citations when you have hyperlinks?

Eric Crampton writes:

For journal articles, year, volume, issue and pages are useful. At least at Canterbury, many of our journal subscriptions are via ProQuest or one of the other similar services; a Google search then brings you to the journal homepage rather than to your library's subscription. If the citation info has volume and issue, it's much easier to find in ProQuest. If it has page numbers, it's much easier to get via InterLibrary Loan if it's an obscure journal.

Gavin Andresen writes:

It would make my life easier if, instead of Caplan, "The Myth of the Rational Voter", 2007 it was bcaplan@gmu.edu, "The Myth..." etc.

Email addresses are nice globally-unique identifiers for people; I'm working in a research group that's working to solve the 'automatic author coreference' problem, and, absent an email address, it's really tough for a computer to tell whether or not the Caplan that wrote some book over here is the same Caplan that wrote an article on a similar topic over there...

They also, of course, are another quality signal (@gmu.edu being higher status than @aol.com).

Hal Varian writes:

There's a whole standards setting effort on this:
http://dublincore.org/
http://en.wikipedia.org/wiki/Dublin_Core

Hoover writes:

I'm not sure any of this addresses the fact that we're in the "Google Age". It just translates the traditional academic citation format so that it can be used on the internet.

The task seems greater than that. Citations in the Google Age need to be able to refer to blog posts, news articles, forum posts, dynamic pages (i.e. the results of online database searches), and so on.

What's more, in the Google Age they should be hyperlinks. A cite should = a hyperlink. They should include an indication as to whether you will have to pay ten dollars to access the original, or register, or whatever.

Of course the problem is that websites disappear, making citations redundant. Citations for search results depend on search syntax and database structure remaining the same. Authors have names like trixybelle222.

What can a citer do, faced with this complexity and unpredictability?

Answer: this is the Google Age and the law ought to permit permanent storage of information for these purposes by people who didn't create it. So a citation (and citations are hyperlinks, remember) would never send you to a page that says "Oops we couldn't find the page you're looking for". It would send you to a digital copy.

Two principles seem to arise from this: 1) information is flatter, and a forum post has equal value to an academic journal or book. 2) People need to be able to copy and permanently archive other peoples' work.

That, at any rate, is the theory. At least in terms of the Google Age. Whether it's fair, or useful, or will advance learning, is another matter.

dearieme writes:

Call me cynical, but I'd settle for the bastards spelling my name correctly.

Boonton writes:

I would only add page number. Even in full Google/Kindle age reading people will still sometimes read full books and it is easier to have them in printed form (anyway, page number translates easy in electronic format).

A URL is perfectly fine in informal settings like comment threads and blog posts. In more formal work, I think you need either a repository of pdfs or a more formal citing method.

Does anyone know how publishing city got into the citation business? Is the idea that if a writer cited some book no one had you would use the yellow pages to find the publisher and for that you need the city?

K writes:

The publisher of a book also sends an important quality signal, so I'd include that too.

Barkley Rosser writes:

If you have title and year and page numbers that will probably do it. However, there are some oddities out there. Some journals paginate by volume, some by issue. Some journals do not even have volume numbers (Brookings Papers on Economic Activity). And then there is the old chestnut of identifying people by initials or fuller names, there being several pairs of economists with identical initials but different names who publish in the same fields of economics.

Comments for this entry have been closed
Return to top