Design of Standards and Regulations
从统计学视角探讨如何用假设检验等工具提高法律、标准和法规的精确性，减少因模糊性导致的诉讼，对立法者和标准制定者有参考价值。
Contrary to popular perception most legal activities take place outside the courtroom, and it ought to remain that way. Otherwise the system will be clogged up by the resolution of trivial and repetitive problems. Ambiguity in the specification of laws, however, is a frequent but often preventable cause for the involvement of the court system. A central issue therefore is the development of more accurate and concise legal specifications. In this paper we address the issue from a statistical point of view. It is common in introductory statistics books to introduce hypothesis testing and related concepts by referring to their similarity to legal decisions. It is not so often recognized that the analogy can be beneficial in the opposite direction. The concise concepts and language of statistical decision theory and hypothesis testing can be of great help in writing laws, standards and regulations as well as in understanding their underlying structure. Moreover, statistical tools such as the power function and prior and posterior distributions can help in evaluating the performance of proposed and existing laws or standards that involve quantitative evidence. Stretching the analogy between hypothesis testing and legal decisions further, it can be seen that writing a law or standard in situations that require quantitative evidence is conceptually equivalent to specifying (designing) an experiment that will provide admissible evidence (data) so that the subsequent test has 'good' power in discriminating between what we believe ought to, and what ought not to, be legal. It is because of this conceptual equivalence that we like to introduce the term design for the activities associated with drafting legal specifications.