| Meta分析中的异质性检验方法 |
| |
| 引用本文： | 何寒青,陈坤. Meta分析中的异质性检验方法[J]. 中国卫生统计, 2006, 23(6): 486-487,490 |
| |
| 作者姓名： | 何寒青 陈坤 |
| |
| 作者单位： | 浙江大学医学院流行病学与卫生统计学教研室,310031 |
| |
| 摘 要： | 目的比较和探讨meta分析中异质性定量化的检验方法。方法通过比较Q检验以及H和I^2统计量等方法，结合一个实例进行分析，来研究meta分析中异质性的定量化检验方法。结果Q检验容易受到样本量变化的影响，而H和I^2统计量经过对自由度的校正，不会受到文献数目变化的影响，异质性结果检验较为稳健。结论H和I^2统计量计算简单，检验结果也稳定可靠，是meta分析中异质性检验和异质性来源研究中值得推广应用的方法。
|
| 关 键 词： | Meta分析 异质性 统计学检验 |
Methods for Measuring Heterogeneity in a Meta - analysis |
| |
| He Hanqing,Chen Kun. Methods for Measuring Heterogeneity in a Meta - analysis[J]. Chinese Journal of Health Statistics, 2006, 23(6): 486-487,490 |
| |
| Authors: | He Hanqing Chen Kun |
| |
| Affiliation: | Department of Epidemic and Health Statistics , School of Medicine, Zhejiang University (310031 |
| |
| Abstract: | Objective To demonstrate the optimal methods of heterogeneity test in meta- analysis. Methods Q - test, H and 12 statistics were compared to test the existence of heterogeneity in meta - analysis in this study. Results In these methods, Q - test was poor in detecting the true heterogeneity among studies when meta - analysis includes small numbers of studies, while H and 12 statistics were independent of the num-her of studies in meta - analysis. Conclusion We propose H and 12 staffsties as the methods for testing heterogeneity in the meta - analysis, and both of them can be readily calculated. |
| |
| Keywords: | Meta - analysis Heterogeneity Statistical test |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
