In Monique Da Silva Moore v. Publicis Groupe & MSL Group, Case No. 11-cv-01279 (S.D.N.Y. Feb. 24, 2012), U.S. Magistrate Judge Andrew J. Peck for the U.S. District Court for the Southern District of New York endorsed the use of predictive coding to locate electronically stored information in a document-intensive, employment discrimination case involving 3 million emails. This is the first case to date that endorses a protocol for the use of predictive coding to locate documents relevant to litigation in ESI.
Predictive coding has many benefits -- it is cost-effective and can cut review time down to a few weeks. In contrast, manual review and keyword searches can cost up to $8.50 per document.[FOOTNOTE 1] In document-intensive cases, the costs can add up to millions of dollars and take many months to complete. Even more troubling, keyword searches can leave up to 80 percent of relevant ESI undiscovered.[FOOTNOTE 2] Predictive coding reduces the number of documents that need to be manually reviewed, which results in a significant reduction of e-discovery costs.
Unlike other search methods that are used throughout the legal industry, like keyword searches, predictive coding uses mathematical formulas that are derived from document coding choices made by experts, who are usually senior attorneys working on the case. The expert codes random sets of documents taken from the corpus of e-discovery and specifies the relevancy of the documents. The computer learns how the lawyer codes the documents and develops a formula for "relevancy." The formula is then applied to the entire document collection to locate all relevant documents in the case.
To Continue Reading: Click Here
By: Rebecca N. Shwayri