王慧,李长平,胡良平.复杂抽样调查设计多值有序资料一水平多重Logistic回归分析[J].四川精神卫生杂志,2019,32(5):400-405.
复杂抽样调查设计多值有序资料一水平多重Logistic回归分析
One-level multiple Logistic regression analysis of the multi-value ordered data collected from the complex sampling survey design
投稿时间:2019-09-27  
DOI:10.11886/j.issn.1007-3256.2019.05.004
中文关键词:  复杂抽样设计  多值有序资料  Logistic回归分析  抽样权重
英文关键词:Complex sampling survey  Multi-value ordered data  Logistic regression analysis  Sampling weights
基金项目:国家高技术研究发展计划课题资助(2015AA020102)
作者单位邮编
王慧 天津医科大学公共卫生学院卫生统计学教研室 300070
李长平 天津医科大学公共卫生学院卫生统计学教研室世界中医药学会联合会临床科研统计学专业委员会 100029
胡良平 世界中医药学会联合会临床科研统计学专业委员会军事科学院研究生院 100850
摘要点击次数:
全文下载次数:
中文摘要:
      本文目的是比较不同分析策略对复杂抽样调查设计多值有序资料一水平多重logistic回归分析结果的异同。通过实例分析,利用四种不同的分析策略(将复杂抽样视为单纯随机抽样,考虑抽样设计不考虑抽样权重,考虑抽样权重不考虑抽样设计,同时考虑抽样设计和抽样权重)对复杂抽样设计多值有序资料进行建模。在四种不同分析策略的累积logistic回归模型拟合的结果中,自变量的偏回归系数、标准误差及P值均有所不同。在对复杂抽样调查设计的多值有序资料回归建模时,将抽样设计和抽样权重纳入统计分析,会得到更准确、更稳健的分析结果。
英文摘要:
      To compare the results of one-level multiple logistic regression analysis of multiple-value ordered data collected from the complex sampling survey design by using different analysis strategies. Four different analysis strategies (treating complex sampling as simple random sampling, considering sampling design without considering sampling weights, considering sampling weights without considering sampling design, and considering both sampling design and sampling weights) were used to model the multi-value ordered data of complex sampling design. In the cumulative logistic regression model fitting results of four different analysis strategies, the partial regression coefficients, standard error and P value of independent variables were all different. In the regression modeling of multi-value ordered data of complex sampling survey design, more accurate and reliable analysis results could be obtained by incorporating sampling design and sampling weights into building regression models.
      附件
查看全文  HTML  查看/发表评论  下载PDF阅读器
关闭
手机扫一扫