On the approximations of critical values and power comparisons of several combination test methods under the dependence assumption

• Main Authors: Huang Fang Yi, 黃方意
• Other Authors: John Jen Tai, Chia-Ding Hou
• Format: Others
• Language: zh-TW
• Published: 2014
• Online Access: http://ndltd.ncl.edu.tw/handle/28583660408837546265

碩士 === 輔仁大學 === 統計資訊學系應用統計碩士班 === 102 === Combining P-value from separate tests is a very useful approach for gathering information of significance, and these methods have been widely used in various aspects of application. Accordingly, comparing the power performances of different combination methods for helping researchers in selection of a suitable method in a practical study turns out to be an important issue in this area. Through simulation, Loughin(2004) had systematically compared six well-known combination methods under the assumption that the separate tests are independent. However, there is no research that has been done on comparing these combination methods under the scenario that separate tests are dependent. In this thesis, the respective approximate formulas are proposed for the critical values of five well-known combination methods under the dependence scenario. A series of simulations are conducted to compare power performances of these methods.