Open Access. Powered by Scholars. Published by Universities.®
- Publication
Articles 1 - 2 of 2
Full-Text Articles in Law
What's Good In Theory May Be Flawed In Practice: Potential Legal Consequences Of Poor Implementation Of A Theoretical Sample, Melanie S. Williams, A. Lynn Phillips, G. Michael Phillips
What's Good In Theory May Be Flawed In Practice: Potential Legal Consequences Of Poor Implementation Of A Theoretical Sample, Melanie S. Williams, A. Lynn Phillips, G. Michael Phillips
Melanie S. Williams
The article discusses the problem of the use in litigation of statistical sampling. Sample-based research is increasingly used in cases as diverse as products liability, antitrust, intellectual property, and criminal law, among others. Sample-based research provides objective evidence upon which decisions, damages and liability may rest. Despite its importance, however, statistical evidence is often misused and misunderstood by attorneys who may be unfamiliar with the underlying form of analysis. The paper explores common errors when using litigative samples, comments upon best practices for the use in law of sample-based research, and demonstrates the importance of sound statistical sampling and data …
Lessons From Single-Company Event Studies: The Importance Of Controlling For Company-Specific Events, Scott D. Hakala
Lessons From Single-Company Event Studies: The Importance Of Controlling For Company-Specific Events, Scott D. Hakala
Scott D Hakala
Single-company event studies are commonly employed in applied practice, such as in analyzing market efficiency, reliance, and damages in securities litigation. However, the presence of significant company-specific events among the observations used to estimate the market model results in significantly biased, overstated standard errors (a well-known omitted variables problem) and less reliable coefficient estimates in such studies. This is a frequently over-looked or neglected issue that renders the statistical inferences in single-company event studies employing using more traditional event study techniques biased and often unreliable. This paper demonstrates through simulation and actual examples that, even allowing for errors in implementation, …