A Modified Approach to Random Selection of Bellwether Cases
Bloomberg BNA's Product Safety & Liability Reporter, July 6, 2015
Mass tort litigation can involve hundreds or thousands of plaintiffs who bring suit against one or more defendants in multiple jurisdictions. To process these cases efficiently, courts sometimes first try a subset of plaintiffs who are thought to be representative of the other cases in the docket. The outcomes of these so-called bellwether trials can provide information about the other cases in the docket and the importance of different factors in determining trial outcomes.
Courts can use different methods to select bellwether cases. In some instances, one or both parties make selections, subject to the court's supervision ("party selection"). In other instances, the court selects cases at random ("random selection") or based on input from the parties ("court selection"). Each of these methods has relative merits and drawbacks.
In a recent article, "A Modified Approach to Random Selection of Bellwether Cases," (Bloomberg BNA's Product Safety & Liability Reporter, July 6, 2015) Vice President Dov Rothman, Manager David Toniatti, and DLA Piper partner Matthew Holian introduce a new approach for determining bellwether case pools - termed "random selection with sample veto."
The authors demonstrate how this approach can address criticisms that random selection can produce unrepresentative results -- particularly when applied to a small sample size. In this approach, the court initially draws multiple random samples of bellwether cases, and each party is then permitted to eliminate a certain number of samples of cases, but must retain all cases within each sample, to produce a bellwether pool. The authors show that the random selection with sample veto method retains many of the benefits of random selection, while reducing the risk of an unrepresentative sample. The sample veto process provides a way for the parties to eliminate unrepresentative samples, and decreases the risk that the remaining bellwether sample will be unrepresentative in a way that advantages one party over another. This process also increases the probability that the bellwether sample will reflect the average strength of the docket while also ensuring that cases across the spectrum, including very weak and very strong cases, can be included as bellwether cases.