Nhanes数据库介绍及使用（三）--方差估计怎么办

1. 复杂抽样的方差估计

复杂抽样与简单随机抽样不同，方法学上的差异也导致了数据分析时的差异。以方差估计为例，简单随机抽样的标准统计软件包计算的方差估计值通常太低且有偏差，没有考虑差异加权和集群内样本人员之间的相关性。

2.子人群的方差估计

在实际数据分析过程中，很多时候我们需要分析的都是部分人群的数据，英文中可能称为subpopulations,domains
官方建议不要随便删除不需要的人群，仍然将这部分人群保留在数据集中，通过分析软件特定语句呈现数据集中感兴趣人群的结果

For example, to estimate mean body mass index (BMI) and its standard error for men aged 20 and over, the entire dataset of examined individuals who have an exam weight, including females and those younger than 20 years, must be read into the statistical software procedure. -官方给出的例子

3.SAS方差估计代码

PROC SURVEYMEANS data=one varmethod=taylor nomcar;
  STRATA sdmvstra;  /*SAS数据集中有相应变量，基本上每次都是用这个*/
  CLUSTER sdmvpsu;  /*SAS数据集中有相应变量，基本上每次都是用这个*/
  WEIGHT WTMEC4YR;
  DOMAIN Select;
   * more statements...;
run;

逐句解释一下，第一句中的surveymeans 是SAS中用于复杂抽样设计的分析代码，相对于普通设计的就是means。除了surveymeans之外，用于复杂抽样的代码还有surveyselect, surveyfreq, surveyreg, surveylogistic, surveyphreg （详细介绍见SAS帮助）；Taylor是官方推荐的方法，一般都是选择这个；nomcar是非完全随机缺失

However, if there is evidence indicating that the nonrespondents are different from the respondents for your study, you canuse the NOMCAR option to compute descriptive statistics among respondents while still counting the number of nonrespondents.
Considering the possibility that those students who did not respond spend differently than those students who did respond,you can use the NOMCAR option to request the analysis to treat the respondents as a domain rather than exclude the nonrespondents.
Although the point estimates are the same as the analysis without the NOMCAR option, for this particular example, the variance estimations are slightlyhigher when you assume that the missingness is not completely at random -SAS教程中关于nomcar的具体描述

strata是指定第一阶段的分层变量
cluster是指定定群变量
weight就是之前提到的权重
domain是用于呈现数据集中感兴趣人群的结果。相对于在不删除不感兴趣人群的情况下分析部分感兴趣人群的结果。与by语句类似

4.参考内容

https://wwwn.cdc.gov/nchs/nhanes/tutorials/module4.aspx 以及 sas help

最后编辑于：2021.02.10 15:42:30