Multilevel Selection in Litigation Data

The selection effects in litigation data are one of the most daunting problems facing legal researchers. We develop a bounds approach to dealing with multiple levels of selection. We build on work by Helland and Yoon on the English rule's effect on litigation outcomes. The English rule prescribes that the loser of a lawsuit pays the winner's litigation costs. When we take selection due to settlement and to drops into account, the bounds analysis suggests that some conclusions in the works of Hughes and Snyder (1990 and 1995) may not be robust to the most extreme forms of selection.