arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 0803.4057stat.AP
收藏

回复给编辑的信

Response to the Letter to the Editor

Clifford Spiegelman, S. J. Sheather, W. A. Tobin, W. D. James, S. Wexler, D. M. Roundhill

摘要:

这篇论文引起了全世界媒体(电视和报纸)的兴趣。此外,我们还收到了信件、电子邮件和电话。我们最喜欢的一条是一条语音邮件,要求我们回一个电话到澳大利亚,在这一点上,我们将了解谁真正杀死了肯尼迪。我们欢迎有机会对菲奥伦蒂诺先生给编辑的信作出答复。Fiorentino先生声称,我们关于第二个刺客的可能性的声明是基于三个或更多单独的子弹的前提,这显然是错误的。作为回应,我们想简单地引用Gerald Posner的书Case Closed的第327页,这是支持单一刺客理论的最着名的作品之一:“如果康纳利被另一颗子弹击中,它必须是由第二个射手发射的,因为沃伦委员会自己的重建表明奥斯瓦尔德不可能在1.4秒内操作螺栓并重新开火。菲奥伦蒂诺还声称,“第二个致命缺陷是使用了一个基于贝叶斯定理的相当简单的公式。”让$E$表示证据,$T$表示只有两颗子弹(因此只有一个枪手)的理论。我们使用贝叶斯定理从$P(E|T)$和先验概率$P(T)$假设计算$P(T|E)$。为了使$P(T|E)$的可能性比$P(\bar{T}|E)$高十倍,先验概率的比率[即,$P(T) / P(\bar{T})$ ]必须大于15。因此,我们再次得出结论,这对Guinn博士关于证据只支持两颗子弹的结论提出了严重质疑。可悲的是,这远远不是第一次在涉及公共利益的案件中误解和/或滥用概率。英国的一个著名例子是克拉克案。详情参见Nobles and Schiff(2005)。最后,我们欢迎,事实上也鼓励科学界的成员对数据进行其他分析。

英文摘要:

This paper has attracted interest around the world from the media (both TV and newspapers). In addition, we have received letters, emails and telephone calls. One of our favorites was a voicemail message asking us to return a call to Australia at which point we would learn who really killed JFK. We welcome the opportunity to respond to the letter to the editor from Mr. Fiorentino. Mr. Fiorentino claims that our ``statement relating to the likelihood of a second assassin based on the premise of three or more separate bullets is demonstrably false.'' In response we would like to simply quote from page 327 of Gerald Posner's book Case Closed, one of the most well known works supporting the single assassin theory: ``If Connally was hit by another bullet, it had to be fired from a second shooter, since the Warren Commission's own reconstructions showed that Oswald could not have operated the bolt and refired in 1.4 seconds.'' Mr. Fiorentino also claims that the ``second fatal flaw is the use of a rather uncomplicated formula based on Bayes Theorem.'' Let $E$ denote the evidence and $T$ denote the theory that there were just two bullets (and hence a single shooter). We used Bayes Theorem to hypothetically calculate $P(T|E)$ from $P(E|T)$ and the prior probability $P(T)$. In order to make $P(T|E)$ ten times more likely than $P(\bar{T}|E)$, the ratio of the prior probabilities [i.e., $P(T) / P(\bar{T})$] would have to be greater than 15. Thus, we again conclude that this casts serious doubt on Dr. Guinn's conclusion that the evidence supported just two bullets. Sadly, this is far from the first time that probability has been misunderstood and/or misapplied in a case of public interest. A notable British example is the Clark case. See Nobles and Schiff (2005) for details. Finally, we welcome and, in fact, encourage members of the scientific community to provide alternative analyses of the data.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑