Question

我正在尝试使用调查包计算一部分穷人。所以我创建了一个子集，其中包含总数的变量（一个）并称为n（我不确定它是否是正确的方法）。所以，我有：

str(base2015$q)
 num [1:117939] 0 0 0 0 0 0 0 0 0 0 ...
str(base2015$one)
 int [1:117939] 1 1 1 1 1 1 1 1 1 1 ...

和分类变量区域。（是我的svydesign）我尝试使用像这样的svyratio函数，但他给我所有数据集的结果，我希望按子组（区域）。我希望每个地区都有这个（见图） proportion of poor

svyratio(numerator=~q, 
         ~one, 
         design = y, separate=FALSE, na.rm=TRUE, 
         formula, covmat=FALSE, deff=FALSE)

我还尝试使用群体创建子集，然后使用svyby：

n <- subset( y , one == 1 )
View(n)
svyby( ~ q , ~ region , n , vartype=c("se","ci","cv"), svytotal )
svyby( ~ q , ~ region , n , vartype=c("se","ci","cv"), svymean )
  region  q  se ci_l ci_u  cv
1      1 NA NaN  NaN  NaN NaN
2      2 NA NaN  NaN  NaN NaN
3      3 NA NaN  NaN  NaN NaN
4      4 NA NaN  NaN  NaN NaN
5      5 NA NaN  NaN  NaN NaN

他给了我这个。我不知道我做错了什么。有人可以帮我这个吗？

Answer 1

data(scd)

scddes<-svydesign(data=scd, prob=~1, id=~ambulance, strata=~ESA,nest=TRUE, fpc=rep(5,6))

svyby( ~ alive , denominator = ~ arrests , by = ~ ambulance , scddes , svyratio )

# which missings do you want to remove?  read `?svyby`
svyby( ~ alive , denominator = ~ arrests , by = ~ ambulance , scddes , svyratio , na.rm = TRUE , na.rm.by = TRUE , na.rm.all = TRUE )

穷人的比例

1 个答案: