Journal of the American Statistical Association volume 81, issue 394, P277-286 1986 DOI: 10.1080/01621459.1986.10478271 View full text
|
|
Share
Franklin M. Fisher

Abstract: Statisticians acting as expert witnesses encounter special problems. These include (a) exposition of underlying methods and concepts, (b) the avoidance of data mining when attorneys wish to leave no stone unturned, and (c) the explanation of hypothesis testing versus prediction. Data management practices are particularly important. Appropriate and inappropriate forms of criticism are discussed, as is the problem of maintaining objectivity. Examples drawn from actual experience are given.