Related Experiment Videos
Biostatistical concepts and methods in the legal setting
J L Gastwirth1, S W Greenhouse
1George Washington University, Washington, D.C. 20052, USA.
Statistics in Medicine
|August 15, 1995
Summary
Biostatistical methods are crucial for U.S. legal cases, offering insights into association measures and regression analysis. This study highlights limitations of Fisher's exact test and demonstrates the utility of adapted Cornfield procedures and Peters-Belson regression for legal evidence.
Area of Science:
- Biostatistics
- Legal Statistics
- Statistical Inference
Background:
- Biostatistical methods are increasingly applied to legal cases.
- Key statistical concepts include measures of association and regression analysis.
- Challenges exist in applying standard statistical tests in legal contexts.
Purpose of the Study:
- To explore the application of biostatistical concepts in U.S. legal cases.
- To address the inapplicability of Fisher's exact test under specific conditions.
- To demonstrate the utility of adapted Cornfield procedures and Peters-Belson regression in legal settings.
Main Methods:
- Analysis of Fisher's exact test applicability in 2x2 tables.
- Adaptation of Cornfield's procedure for assessing omitted variables.
- Application of Peters-Belson regression methodology.
Main Results:
- Fisher's exact test is inappropriate when sample size determination is linked to the hypothesis.
- Adapted Cornfield's procedure can evaluate the impact of unobserved factors on non-significant findings.
- Peters-Belson regression is relevant for equal pay and promotion litigation.
Conclusions:
- Biostatistical methods provide essential tools for legal analysis.
- Careful consideration of test assumptions is vital in legal applications.
- Specific statistical methodologies offer valuable insights for discrimination and pay equity cases.