During LegalTech, there were probably half a dozen panels on E-Discovery with a focus on predictive coding. One of the more interesting panels offered a predictive coding case study, featuring Dr. David Lewis. With David’s permission, I post these tweets of the slides–though I would stress that these findings are preliminary and subject to change.
The first slide compares the precision rate for old-fashioned linear review, and for predictive coding. The former is at 76%, and the latter is at 87%. In other words, using predictive coding is more likely to identify the correct documents, and less likely to identify the wrong documents.
Precision for predictive coding is 87% and 76% for linear review #LTNY pic.twitter.com/ANMIX8KM5X
— Josh Blackman (@JoshMBlackman) February 4, 2014
Second, predictive coding yielded superior results to even manual review, which is much more time-intensive.
Comparing Manual Review with Predictive Coding #ltny pic.twitter.com/cQZYWWQTT9
— Josh Blackman (@JoshMBlackman) February 4, 2014
The precision for identifying privileged documents was also much higher.
Predictive Coding for Privilege #LTNY pic.twitter.com/hZrKjOENpS
— Josh Blackman (@JoshMBlackman) February 4, 2014
And predictive coding was better at identify the “hot doc” (that is the one document you do not want to produce).
Comparing location of “hot document” with predictive coding and manual review. pic.twitter.com/t0NQaehehh
— Josh Blackman (@JoshMBlackman) February 4, 2014
In an earlier presentation, Judge Peck offered a helpful summary of predicting coding caselaw.
Judge Peck’s summary of predictive coding case law #LTNY pic.twitter.com/fvH6FbfUMH
— Josh Blackman (@JoshMBlackman) February 5, 2014
But there are still many pending questions.
Unsettled questions about predictive coding #LTNY pic.twitter.com/tymeRSD910
— Josh Blackman (@JoshMBlackman) February 5, 2014