D.D.C.: Low Precision Rate Justifies Extension of Discovery Deadline
Today, Judge Colleen Kollar-Kotelly issued a decision, In re Domestic Airline Travel Antitrust Litig., MDL No. 2656, No. 15-1404 (CKK), 2018 U.S. Dist. LEXIS 155775 (D.D.C. Sept. 13, 2018), granting the Plaintiffs' Motion for an Extension of Fact Discovery Deadlines Pursuant to FRCP 16(b)(4). The Plaintiffs in this suit alleged that the Defendants conspired to restrain trade in the sale of air passenger seats.
The Plaintiffs argued that one of the Defendants, United, used flawed TAR technology which resulted in only 17% of its production of 3.5 million documents being responsive. A scheduling order may only be modified for good cause, and the moving party must show that it was diligent in attempting to perform discovery pursuant to the deadlines of the order.
The Plaintiffs loaded the production into their review platform as soon as it was received and engaged a staff of 70 attorneys to review the documents. Despite the fact that the Defendants questioned whether or not the Plaintiffs had actually engaged so many attorneys for their review, the Court did not require the submission of the names of the attorneys or their billing reports. The Plaintiffs rejected the contention made by the Defendants that a review rate of 3 documents per minute was a reasonable pace. Judge Kollar-Kotelly did not find the fact that United had 180 contract attorneys assist with the production of the documents pertinent to the issue of whether or not the Plaintiffs had to deal with unanticipated circumstances.
The TAR protocol entered into by the parties set a minimum 75% estimated recall rate for United's production and required it to attempt to achieve a higher rate, "if that rate may be obtained with a reasonable level of precision through reasonable additional training effort, taking into account the concept of proportionality and the deadline for substantial completion of document production." Id. *22-23.
United found 85% recall and 58% precision in its control set, but the Plaintiffs' analysis of a validation sample indicated 97.4% recall and only 16.7% precision. United ultimately conceded that the Plaintiffs' estimate was correct. The Plaintiffs said their own TAR tool could not be trained to remove non-responsive documents from the United production.
The Court concluded that the Plaintiffs could not have anticipated the precision level of the United production. It also ruled that the delay in filing the motion until more than three months after the production was diligent given the fact that the parties continued to discuss the reasons for the recall and precision rates.