Premium
Statistical Evidence in Discrimination Cases
Author(s) -
Gastwirth Joseph L.
Publication year - 1997
Publication title -
journal of the royal statistical society: series a (statistics in society)
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.103
H-Index - 84
eISSN - 1467-985X
pISSN - 0964-1998
DOI - 10.1111/1467-985x.00063
Subject(s) - circumstantial evidence , context (archaeology) , statistical evidence , statistical hypothesis testing , statistical inference , statistical analysis , test (biology) , inference , statistical model , econometrics , process (computing) , psychology , promotion (chess) , affect (linguistics) , actuarial science , computer science , statistics , law , political science , mathematics , history , economics , null hypothesis , artificial intelligence , paleontology , archaeology , biology , operating system , communication , politics
Statistical data and analyses are used as circumstantial evidence indicating that race, gender, age or another legally prohibited factor was a significant cause of an unfavourable employment or governmental decision. This paper discusses the role of statistical evidence in the context of the legal process. Changepoint techniques to determine when an employment practice changed and Cornfield’s approach to examining alleged flaws in a statistical analysis to ensure that they are sufficiently severe to affect the final inference are illustrated. The importance of understanding the underlying process before deciding on the appropriate statistical methodology is described in the context of a promotion case where the routine use of Fisher’s exact test, without checking that the original hirings were fair, can lead to anomalous results.