The Search Term Dilemma

Have you ever had the unenviable task of coming up with an all-inclusive search term list in your case that will only bring back relevant documents? You work diligently for hours on your list and have others on your team and opposing counsel “tweak” the terms only to get back from your search a ton of non-responsive junk, including photos of your clients’ children in bunny suits!

This brings up the problem with many searches done by law firms today. Many law firms rely solely on search terms when they should be using a two-tiered search philosophy which is crucial to obtaining relevant documents in a cost-effective manner. Not only do you need to come up with the best search terms possible, but you also need to have a “conceptual” search done on your data. For a good, detailed definition of conceptual searches see David T. Chaplin’s article entitled “Conceptual Search- ESI, Litigation and the Issue of Language.”  Concept searching often involves mathematical and linguistic models and usually someone must teach the computer to recognize the correct terms and concepts. This search logic relies on a thesaurus to capture documents that use different words to state the same meaning. The process is invaluable in order to capture all the terms needed to bring back relevant data from your searches.

As indicated in the ABAJournal article “In Search of the Perfect Search” by Jason Krause, George Paul, a partner at Lewis and Roca in Phoenix, states that the language issues are “not a computer problem…words don’t stand for behavior, but are elastic and change their meaning depending on their context.”

Furthermore, this article describes the TREC Legal Track [Text Retrieval Conference] research that has identified a few practices that improve upon the basic keyword search. First, one must work with opposing counsel to help identify search terms and Boolean search strings. The most critical step is to have sampling done of your search terms to make sure your results bring back relevant documents. I heartily agree with this approach. In one of my cases the approach of using agreed upon search terms and sampling brought back approximately an 80% relevancy rate compared to the more common 20%.

I’m not saying using both search terms and conceptual searches will eliminate the need to review junk e-mails; however, it will definitely lessen the amount of non-responsive data that comes back from your EDI searches and will make the task of document review quicker and less horrendous.