|
Activity Number:
|
90
|
|
Type:
|
Invited
|
|
Date/Time:
|
Monday, July 30, 2007 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Committee on Law and Justice Statistics
|
| Abstract - #307965 |
|
Title:
|
The Effect of Serial Correlation on Statistical Evidence in Legal Cases
|
|
Author(s):
|
Yulia R. Gel*+ and Weiwen Miao and Joseph L. Gastwirth
|
|
Companies:
|
University of Waterloo and Macalester College and George Washington University
|
|
Address:
|
200 University Ave. W., Waterloo, ON, N2L 3G1, Canada
|
|
Keywords:
|
non-parametrics ; legal statistics ; time series ; goodness-of fit tests
|
|
Abstract:
|
When data occur sequentially in time, they typically have a positive serial correlation. Such a dependence among observations is known to affect the distribution of both parametric and nonparametric test statistics. The effect on the p-value of a particular test is related to the dependence structure underlying the data. We will present a study of the dependence effect on various statistical tests utilized in legal cases and possible remedial approaches. The implications of our results for a statistical analysis submitted in a securities law case and a jury discrimination case will be discussed.
|
