##############Estimating total and mean in survey data#################

###stratified random sampling
library(survey)
library(foreign)
data(hospital)
hospital
dstr <- svydesign(id = ~1, strata = ~oblevel, fpc = ~tothosp, data = hospital)
smean<-svymean(~births, dstr)
smean
confint(smean, level=.95)
stotal<-svytotal(~births, dstr)
stotal
confint(stotal, level=.95)
svyby(~births, ~oblevel, dstr, svymean, keep.var = TRUE)
svyby(~births, ~oblevel, dstr, svytotal, keep.var = TRUE)
svyquantile(~births, dstr, c(0.25,0.5,0.75), ci=TRUE)


#####one stage cluster sampling
ex.data<-read.table(file="W:/teaching/stat579/data/gpa.txt", header=T)
design<-svydesign(id=~suit,weight=~wt,fpc=~rep(100,20),nest=TRUE,data=ex.data)
summary(design)
svytotal(~gpa,design)
smean<-svymean(~gpa,design)
confint(smean,level=.95)
anova(~gpa,design)

#####two-stage cluster sampling
# The Academic Performance Index is computed for all California
# schools based on standardized testing of students.
# The data sets contain information for all schools with at least 100 students
# and for various probability samples of the data.
# apipop contains the entire population.
# apistrat contains a sample stratified by stype.
# apiclus1 contains a cluster sample of school districts.
# apiclus2 contains a two-stage cluster sample of schools within districts.
data(api)
help(api)
clu2design<-svydesign(id=~dnum+snum,fpc=~fpc1+fpc2,data=apiclus2)
dimnames(apiclus2)[2]
summary(clu2design)
svytotal(~enroll,clu2design,na.rm=TRUE)
smean<-svymean(~enroll,clu2design,na.rm=TRUE)