The research began in 1980 and sought to determine how often the police deviate from rules of evidence in deciding whether or not to prosecute, how often weak cases are prosecuted, and whether an independent body of prosecutors could reduce the numbers of weak and cautionable cases prosecuted. The research used interviews and an examination of 400 cases. However, the research became more acceptable as a management survey than as an academic survey. Since most of the cases were strong, the research gave unreasoned support to the idea that the prosecution process works fairly in general. Despite the problems in the research, it is important to use research for policy purposes. However, this will be effective only if policy does not define the problem in the first place. Notes.