統合分析是一門綜合不同研究中的資訊，以針對所感興趣的共同特徵加以分析而得到統計推論的學科。本篇論文所關注的是在統合分析中如何結合不同研究的變異。除了採用典型的常態模型外，我們同時也使用了伽瑪分配及逆高斯分配以探討如何研究並結合變異以得到正確的結論。;Meta-analysis is a discipline that gathers information from different studies and provides a summary statistic (or statistics) about the characteristic of common interests in the studies. This dissertation is concerned with how variations of various studies should be incorporated in Meta-analysis. In addition to the typically adopted normal model, we use gamma and inverse Gaussian distributions to explore how investigators should integrate variations to draw conclusions.
