Data sampling is nothing new. Anyone who has lived through at least one election in Canada has been exposed to sampling. It’s the caveat at the end of a political poll that reads something like “This poll is accurate within 3 percentage points, 19 times out of 20”. This indicates that, instead of asking every eligible voter in the country their opinion, the pollsters asked a small percentage (typically around 1,000 people) and used statistical sampling methods to extrapolate statistically relevant results.
E-discovery deals with records, not votes. However, since a typical file will involve lots of records, and the cost to review all of them is prohibitive, the same type of sampling that is used for polling can be used with e-discovery.
There are a number of references on sampling methodologies specific to e-discovery. One is the EDRM Search Guide, which offers a basic outline of how sampling is applied to e-discovery data sets.
The Sedona Conference’s article titled ‘Achieving Quality in the E-Discovery Process’ delves further into sampling, discussing specific applications, purposes, and court case examples where it has been used. The Sedona Conference paper mentions the current norm used by many e-discovery professionals to cull data, which is termed “judgmental sampling”. This is where “…the practitioner has a general sense of which of the several custodians and date range is most likely to offer the greatest yield.” It compares this approach to “statistical sampling”. A key point was that “One area where statistical sampling has an advantage is that quantifiable measures of error and confidence intervals are possible, while judgmental sampling has no such formal measurement.”
Sampling is already incorporated into many e-discovery tools. Many of the more advanced Early Case Assessment systems employ it, and Predictive Coding is predicated on the use of statistical analysis. However, with a minimal amount of work, statistical sampling can be applied to any e-discovery review, dramatically reducing costs without sacrificing quality.
For more information on how sampling can help you achieve better e-Discovery review results, contact Chuck Rothman at Wortzman Nickle.