Monday, September 11, 2006

The Long Tail of Legal Scholarship
Tom Smith

This is interesting (as of course I would say).  Paul Caron appears to show (if I get his drift) that SSRN downloads are like semi- or quasi-cites, in that they fill out the long tail you would expect in legal scholarship, which my research, looking at citations only, found ended aburptly when 40 percent of articles did not get cited at all.  As you can see in Paul's Table 2, even the bottom 10 percent of SSRN papers by download, do get at least a few downloads.  There is a big difference between just a few, and zero.  So downloads reveal that the long tail is there.  If there were some way to measure "read or skimmed, but not cited," undoubtedly law review published articles would have a much longer tail as well.

I have another theory related to these discussions I hope to write about at some point, which might be summed up as -- Glenn Reynolds is completely wrong!  That is, I think the operation of Lexis and Westlaw to push otherwise obscure articles more to the fore, is purely a function of their currently very archiac search technology.  When legal search technology becomes as sophisticated as Web search, which I pray (and I mean that literally) is soon, then we will be back to a situation where the rich get richer, in terms of garnering eyeballs.  Of course, the rich are already getting richer in this respect.  But not as much as they would be if the technology were any good.  So I agree with Glenn that legal search technology mitigates this "preferential attachment" effect somewhat, but, I think this is only for now, and only because the "randomizing" effect of Lexis and Westlaw is just a symptom of crude technology.  But that will change.

| Permalink

TrackBack URL for this entry:

Listed below are links to weblogs that reference The Long Tail of Legal Scholarship
Tom Smith


From 1866 to 1900, no-one paid any notice to Mendel's papers.

Posted by: dearieme | Sep 11, 2006 1:41:41 PM

When legal search technology becomes as sophisticated as Web search

So you think boolean searching (terms and connectors searching) is less effective than Web searching (a natural-language type of searching)? I'd say that boolean searching is a much better way to search and that Web searching really misses a lot of documents.

Posted by: pat | Sep 12, 2006 6:54:17 AM

Did you ever notice how horrible the search function is at Amazon? I always assumed it was to force you to look at books you're not trying to find just to record the ISBN number and assign them for class.

Posted by: Agricola | Sep 12, 2006 12:22:59 PM

Great post. I agree archaic search algorithms promote items at the long end of the tail. Google's actually designs its complex and cutting edge algorithms precisely to promote the popular and "cut off the tail" to a nub. While the tail is still there, whatever is not on the first page of results is destined for obscurity.

The next generation (I am guessing here) of search algorithms, will customize searches toward your personal preferences, rather than everyone else’s. This will bring back the tail for everyone, rewarding obscure readers and searchers alike.

Posted by: Brett Trout | Sep 15, 2006 11:04:08 AM