At a panel, a financial-services company’s general counsel described a study he carried out regarding customer arbitrations. The general counsel had staff pull one out of eight arbitrations at random going back 12 years and compile approximately 50 data points about each of them. Among the data points were whether the arbitration was handled by outside counsel or within the law department, the total cost of resolution, and how long the arbitration process took.
That mass of data amply justified a multiple regression analysis (See my post of Aug. 14, 2005 on regression analysis; and April 22, 2007 with several comments on statistics.). The general counsel used an employee who knew multiple regression and could find out from it what factors influenced the costs of arbitrations and to what degree. The analysis disclosed many surprising findings, such as that the ages of the more costly plaintiffs were toward the younger end of the client scale. Other law departments, with other masses of data, could benefit from regression studies.