R/all_funs.R
convey_prep.Rdstores the full survey design (needed for convey functions that use a global poverty threshold) within the design. this function must be run immediately after the full design object creation with svydesign or svrepdesign
convey_prep(design)a survey design object of the library survey.
the same survey object with a full_design attribute as the storage space for the unsubsetted survey design
functions in the convey package that use a global poverty threshold require the complete (pre-subsetted) design in order to calculate variances correctly. this function stores the full design object as a separate attribute so that functions from the survey package such as subset and svyby do not disrupt the calculation of error terms.
library(survey)
library(laeken)
data(eusilc) ; names( eusilc ) <- tolower( names( eusilc ) )
# linearized design: convey_prep must be run as soon as the linearized design has been created
des_eusilc <- svydesign( ids = ~rb030 , strata = ~db040 , weights = ~rb050 , data = eusilc )
des_eusilc <- convey_prep( des_eusilc )
# now this linearized design object is ready for analysis!
# # # CORRECT usage example # # #
des_eusilc <- svydesign( ids = ~rb030 , strata = ~db040 , weights = ~rb050 , data = eusilc )
des_eusilc <- convey_prep( des_eusilc )
sub_eusilc <- subset( des_eusilc , age > 20 )
# since convey_prep() was run immediately after creating the design
# this will calculate the variance accurately
SE( svyarpt( ~ eqincome , sub_eusilc ) )
#> eqincome
#> eqincome 58.57576
# # # end of CORRECT usage example # # #
# # # INCORRECT usage example # # #
des_eusilc <- svydesign( ids = ~rb030 , strata = ~db040 , weights = ~rb050 , data = eusilc )
sub_eusilc <- subset( des_eusilc , age > 20 )
sub_eusilc <- convey_prep( sub_eusilc )
#> Warning: this function must be run on the full survey design object immediately after the svydesign() or svrepdesign() call.
# since convey_prep() was not run immediately after creating the design
# this will make the variance wrong
SE( svyarpt( ~ eqincome , sub_eusilc ) )
#> eqincome
#> eqincome 58.57576
# # # end of INCORRECT usage example # # #