How to Jump-Start Early Case Assessment: A Case Study Based on TREC 2008

Contact: Warwick Sharp, Vice President, Marketing and Business Development, Equivio Inc.
(800) 851-1965 | info@equivio.com | www.equivio.com

Early case assessment is becoming more and more important, but at the same time, it's also becoming more and more difficult. It's more important because increasing document volumes are driving up the costs of litigation. Increasing costs, in turn, increase motivation to go the early case assessment route, explore options and settle.

But, here's the catch — increasing document volumes are also making early case assessment ever more difficult and challenging. To do early case assessment effectively, you need an effective way to isolate the most relevant documents as early as possible in the process. (By the way, the same problem plagues internal investigations where corporations seek out the most relevant documents so they can make a quick decision on whether to pursue the case.)

Unfortunately, the standard way of finding relevant documents in a collection, via keyword search, does not do the job. For over a decade, the National Institute of Standards and Technology ("NIST") has co-sponsored the Text Retrieval Conference ("TREC") to promote research in information retrieval from large text collections. Of greatest interest to the legal community, though, is the TREC Legal Track, which focuses on measuring different methods, including traditional keyword search, for retrieving relevant discovery documents in litigation matters.

The Equivio>Relevance™ product, released last month, addresses the early case assessment challenge by scoring the relevance of each document in a collection. To test the efficacy of Equivio>Relevance, Equivio ran the software on a data set of 7 million documents from the TREC 2008 Legal Track.

The results of the test are detailed in the case study How to Jump-Start Early Case Assessment: A Case Study Based on TREC 2008. The key finding shows that use of this technology almost tripled the recall and precision rates of keyword search.

The case study enables a comparison between the use of machine-learning technology in early case assessment and keyword search. Via machine-learning technology, the user can:

• Prioritize early case assessment by focusing on the most relevant documents.


• Read fewer documents in total.


• Read more relevant documents.