学术预告

学术预告:Modern Biostatistical Methods: ~ Challenges

发布日期:2015-07-07 发表者:辛西 浏览次数:

报告题目:Modern Biostatistical Methods: Capabilities, Cautions and Challenges

报告人:Xin M. Tu(Department of Biostatisticsand Computational Biology, Department of Psychiatry, University of Rochester)

报告时间:2015年7月13日(周一)16:00-17:00

报告地点:逸夫楼C座314会议室

摘要:

Modern statistical methods provide a powerful fool toaddress complex statistical issues arising in clinical research andpractice. Harnessing the power of suchmethods can be quite a challenging task, even for professional biostatisticians. Too often, a range of competing methods isavailable and it can be quite difficult to determine the approach mostappropriate to address the issues at hand, especially when one needs to defendthe approach used, as is often the case in grant applications and manuscriptpublications. Many statistical modelsmay seem quite similar, but can actually produce quite different results withcompletely different interpretations. Also, as statistical packages become ever “user-friendly”,it is tempting to use them without having a good understanding of theunderlying statistical methods as well as the quality of the implementation ofthe methods by the packages. In thistalk, we discuss a range of popular statistical methods for clinical researchand appropriate contexts for their applications using data from both real andsimulated studies.

报告人简介:

Xin M. Tu ,1982年毕业于复旦大学,1989年于杜克大学获得统计博士学位,1990-1992年在哈佛公共卫生学院生物统计系进行博士后研究。现任罗切斯特大学(University of Rochester)生物统计系教授,同时兼任统计咨询服务(StatisticalConsulting Services)和精神病统计学部(Division of PsychiatricStatistics)主席。Xin M. Tu在生物医药、精神病研究的统计应用领域具有深入和丰富的研究经验,他长期从事纵向数据分析、结构方程模型、因果模型、自由分布模型等领域的统计研究,发表SCI论文近200篇,出版学术专著5部。