• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文期刊>International Journal of Statistics and Applications >Does Data Weighting Improve Propensity Scores?
【24h】

Does Data Weighting Improve Propensity Scores?

机译：数据加权可以改善倾向得分吗？

获取原文
           
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

Propensity Score Matching (PSM) is the current popular method used to evaluate an impact of the programmes compared to other non-experimental methods. The method is the most widely used type of matching in which the comparison group is matched with the treatment group on the basis of a set of observed characteristics or by using the propensity score (PS). Although it has emerged as the most widely used method in measuring non-experimental design, it suffers from the debates. One of the major debates of PSM is its functional form. The paper intended to analyse the effects of weighted data in generating PS. The data were collected through the mixed approach as the data collected were both numerical and textural from five Tanzania regions namely: Kagera, Mwanza, Mara, Simiyu and Kigoma. PS was generated using the Logistic Regression (LR) model. The findings show that the weighted data slightly improves propensity scores than unweighted data. Basing on the findings, it can be concluded that the representative sample for the population produce better PS compared to unrepresentative samples. As the weighted data provides better PS compared to unweighted data, the efficiency of PSM is improved. For better results of an impact of the programme, the logistic regression model with the sample data which is representative of the population should be used.
机译：倾向得分匹配（PSM）是与其他非实验方法相比，用于评估程序影响的当前流行方法。该方法是使用最广泛的匹配类型，其中根据一组观察到的特征或通过使用倾向得分（PS）将比较组与治疗组进行匹配。尽管它已成为衡量非实验设计最广泛使用的方法，但它受到了争论。 PSM的主要争议之一是其功能形式。本文旨在分析加权数据在生成PS中的作用。数据是通过混合方法收集的，因为从五个坦桑尼亚地区（即Kagera，Mwanza，Mara，Simiyu和Kigoma）收集的数据既是数值数据又是纹理数据。使用Logistic回归（LR）模型生成PS。研究结果表明，加权数据比未加权数据略微改善了倾向得分。根据这些发现，可以得出结论，与无代表性的样本相比，具有代表性的人群样本具有更好的PS。由于加权数据与未加权数据相比提供更好的PS，因此提高了PSM的效率。为了使计划产生更好的效果，应该使用具有人口代表性的样本数据的逻辑回归模型。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号