CODEWITHSUNDEEP
surveyweighted
Y Huang
Y Huang

Reputation: 1

Estimates in subpopulations with weighted data using survey() package

ftp://cran.r-project.org/pub/R/web/packages/survey/vignettes/domain.pdf

The complete dataset is tch2012. However, I am only interested in the subpopulation of tch2012 in which two criteria are met: age <= 5 and gender == "female". And within that subpopulation, I want to compare those with the disease (disease == "1") and without the disease (disease == "0").

This is the code I wrote: 

library(survey)

tch2012.tsl.dsgn <- svydesign(id= ~HOSP_KID, strata= ~KID_STRATUM, weights = ~DISCWT, data = tch2012, nest = TRUE)

create a pointer to subpopulation of female pediatric age 5 years and under

tch2012_f_age5.tsl.dsgn <- subset(tch2012.tsl.dsgn, AGE <= 5 & gender == "female")

weighted data of total number with and without the disease in female pediatric patients age 5 years and under

svyby(~count, ~disease, design=tch2012_f_age5.tsl.dsgn, svytotal)

However, I got the below error message when I ran the svyby()

Error in sum(sapply(covmats, ncol)) : invalid 'type' (list) of argument

Since I am not very familiar with dealing with weighted data, I am no clue how to trouble shoot.

Thanks in advance for the help!

Upvotes: 0

Views: 441

Answers (1)

Anthony Damico
Anthony Damico

Reputation: 6094

this code works

library(survey)
data(api)
dclus1<-svydesign(id=~dnum, weights=~pw, data=apiclus1, fpc=~fpc)
x <- subset( dclus1 , sch.wide == 'Yes' )
svyby(~api00, ~stype, design=x, svytotal)

please edit your question by adding a minimal reproducible example How to make a great R reproducible example?

Upvotes: 1

Related Questions