Here is the code to solve the problem:

library(survey)

mydesign <- svydesign(id=~C17SCPSU,strata=~C17SCSTR,weights=~C1_7SC0,nest=TRUE, data=ECLSK)
options(survey.lonely.psu="adjust", survey.ultimate.cluster = TRUE)

svymean(~C3BMI, mydesign, na.rm = TRUE)

svymean(~SEX_MALE, mydesign, na.rm = TRUE)

This code defines the survey design using svydesign(), adjusts for PSU lonely cases, and then uses svymean() to calculate the mean of C3BMI and SEX_MALE. The na.rm = TRUE argument is used to remove missing values from the calculations.

Last modified on 2025-04-20