« On to Berlin | Main | The Art of Question-Asking »

Tuesday, July 24, 2007

SSRN: A "Hero Sheet" Sacrificing Functionality for Rankings?

Let me begin by saying that I think SSRN, by facilitating easy access to scholarship, is a fantastic contribution to legal academia -- especially for those who don't otherwise have an easy means to upload their works to the internets (e.g., those who don't have a school with a good IT staff).

But I do have one criticism of SSRN:  On each author's SSRN page, articles are listed by number of downloads, not chronologically or in any other logical order.  This order helps you figure out who’s “winning” the downloads race, but it’s otherwise useless and inconvenient.  If I chose to visit an author’s SSRN page, it’s because I want to see his or her body of work (1) in chron order, to see how the line of scholarship developed (e.g., more or fewer con law publications over time? how prolific?), or (2) in any other sensible order the author chose, like article type (e.g., academic articles vs bar articles, WIP vs published, etc.).

To be clear:  I'm not one of those folks complaining "we shouldn't give a 'score' to scholarship with these download counts."  I don't have any conceptual problem with download counts; my complaint is just that by making it mandatory that authors' works are listed in order of download counts, SSRN is (1) making itself useless as a way to look at an author's budy of work and (2) communicating that download counts are what really matters about that body of work.

SSRN’s biggest boosters have pitched it as the future of legal scholarship.  Maybe it is, but right now, the extent of its focus on "ranking" of scholars, papers, and schools makes it look more like an old-fashioned Hero Sheet focusing on superficial metrics at the expense of more meaningful ways to evaluate folks' scholarship.

Posted by Scott on July 24, 2007 at 11:15 PM in Life of Law Schools | Permalink


TrackBack URL for this entry:

Listed below are links to weblogs that reference SSRN: A "Hero Sheet" Sacrificing Functionality for Rankings?:


Thanks for the comments on the way results appear on SSRN. We have made it possible and simple for users to rank the search results in ways other than download counts (which seems to be popular). If you look on the left hand side of any search results page you will find 4 links. They are: Number of downloads; Date posted, descending; Abstract title, ascending; Abstract title, descending. We have in development many more options, including options for author control of the initial screen on author pages. Sorry we can't get it all done instantly, but when we give most everything away for free and make it up on volume it does limit our budget for development.

Posted by: Michael C. Jensen | Aug 4, 2007 11:53:00 AM

Orin: I completely agree that SSRN's focus on "the downloads race" reflects its desire for eyeballs, which increase both its revenue and its prominence (and I'll plead agnosticism as to whether money, prominence, or idealism motivates the SSRN folks more). I just think that, like US News's sensationalistic (but economically rational) law school rankings, SSRN's sensationalistic rankings focus is unfortunate, and in tension with SSRN's professed academic focus. I suppose I'm just railing against a market equilibrium I don't like, but I'll take comfort in that virtually all the law school deans (a pretty smart bunch) spends a lot more ink and time doing that with respect to the popularity of the US News rankings.

Posted by: Scott Moss | Jul 27, 2007 12:16:36 AM


SSRN is a for-profit business, and I gather their business strategy is to focus on the downloads race because that's their market advantage. The more important SSRN downloads are, the more important SSRN is, and eventually, the more SSRN can charge its customers. I tend to think most-recent-first would be better from a scholarly perspective, but I am not surprised that SSRN ranks by downloads instead.

Posted by: Orin Kerr | Jul 26, 2007 3:29:31 AM

Anon: I rarely respond to anonymous comments because (like yours, "Yawn...") they're often snotty, but you bring up a factual point that either (1) I don't understand or (2) is not true. You say, "you'll just have to add a couple of extra words into the search line, and your pick will come up first" -- but as far as I know, there just is not any way to get, say, the Scott Moss articles to come up in any order other than by download count. Are you saying there is?

Posted by: Scott Moss | Jul 25, 2007 5:01:08 PM

BEPress's Selected Works lets authors categorize their articles and group them along two different axes, one of which is almost entirely unconstrained. It's not perfect, but it does what Scott would like. If you create and host your own page (as I do), you can arrange them however you like. Schools that have publications listed on faculty pages (to my knowledge all of them) typically go chronologically, possibly broken down by format.

Why do these other systems default to non-popularity based orders? Because ordered by downloads is a great way to arrange search results and a less helpful way to arrange items for browsing. SSRN has a search engine of its own; download order is nice for the results of that search when one is looking for a particular paper by an unknown author or papers on a subject. It's perhaps less appropriate when you're browsing an author's papers and are looking to see what they've written.

I tend to prefer some kind of other explicit order -- by subject, by author's favorites, by year -- over popularity when one creates a site. Leave the popularity rankings for the search engines, which specialize in it and can do a great job at it. In SSRN's case, because they have their own search engine, that means they don't need to discard the download counts the way that Google would. All the information is available where it's most useful: authors' preferences on author pages, download counts on SSRN search, and hyperlink popularity on third-party search.

Posted by: James Grimmelmann | Jul 25, 2007 9:25:51 AM

It is a standard technique of search engines to order objects of search according to the interest of prior searchers. This is known to minimize search time. So, when SSRN lists papers by the number of downloads, it simply puts upfront the paper that an average user most likely wants to see. If you have idiosyncratic preferences, you'll just have to add a couple of extra words into the search line, and your pick will come up first. Yawn. A slow news day?

Posted by: anon | Jul 25, 2007 1:04:01 AM

Post a comment