Decision in W.D. Pa. concerning statistical sampling

On March 21, 2018, Judge Mark Hornak issued a decision in EEOC v. Fedex Ground Package Sys., 2018 U.S. Dist. LEXIS 45884 (W.D. Pa.) denying the defendant's motion to preclude the discovery of ESI. The plaintiff's request concerns, "roughly 193 gigabytes of e-mail data (consisting of 363,765 e-mails and attachments) in the possession of twenty-six corporate custodians." (id. at *3) Fedex's motion was made on the basis of the fact that the search would not turn up sufficient responsive material in order to justify such a broad search. The court held that the request was relevant and proportional, and that it could not be said that the emails were not reasonably accessible under Fed. R. Civ. P. 26(b)(2)(b).

Fedex relied on a statistical sample that showed of about 1000 emails containing selected search terms, only roughly 8% would be responsive. It arrived at this low figured after its review indicated some hits would be withheld on the grounds of privilege, and others would be produced as per other discovery requests. It also pointed out that none of the responsive emails constituted a "smoking gun". Judge Hornak found that, "The Court agrees with the EEOC that FedEx incorrectly attempts to frame relevancy as a binary test where information is either irrelevant or a "smoking gun." {Id. at 8}. The cost of the search was estimated to only be $28,000, plus approximately $3,000 in monthly storage fees. The court gave the plaintiff the opportunity to reverse engineer the statistical sample using the same data set.

A review that would yield 30,000 documents should be considered sufficient reason to search the data set in question. Judge Hornak noted that, "the fact that these repositories create complex mechanisms to store huge amounts of information cannot be used in and of itself as a shield to avoid discovery requests otherwise permitted under the Federal Rules of Civil Procedure." [Id. at *10]. The opinion doesn't go into detail but apparently part of Fedex's argument was that it stored its employees emails on systems that did not facilitate the recovery of the data.