miceadds (version 2.2-0)

scale_datlist: Adding a Standardized Variable to a List of Multiply Imputed Datasets or a Single Datasets

Description

Adds a standardized variable to a list of multiply imputed datasets or a single dataset. This function extends base::scale for a data frame to a list of multiply imputed datasets.

Usage

scale_datlist(datlist, orig_var, trafo_var, weights = NULL, M = 0, SD = 1,  digits = NULL)

Arguments

datlist

A data frame, a list of multiply imputed datasets of one of the classes datlist or imputationList or a list of nested multiply imputed datasets of one of the classes nested_datlist or NestedImputationList.

orig_var

Vector with names of the variables to be transformed

trafo_var

Vector with names of the standardized variables

weights

Optional vector of sample weights. Alternatively, the weights can also be a string indicating the variable used from datlist.

M

Mean of the transformed variable

SD

Standard deviation of the transformed variable

digits

Number of digits used for rounding the standardized variable

Value

A vector or a matrix

See Also

base::scale, ma.scale2

Examples

## Not run: 	
# #############################################################################
# # EXAMPLE 1: Standardized variables in list of multiply imputed datasets
# #############################################################################
# 	
# data(data.ma02)
# datlist <- data.ma02
# 
# #--- object of class 'datlist'
# datlist <- datlist_create( datlist )
# 
# # mean and SD of variable hisei
# ma.wtd.meanNA(data=datlist, weights = datlist[[1]]$studwgt , vars = "hisei" )
# mean( unlist( lapply( datlist , FUN = function(data){
#         stats::weighted.mean( data$hisei , data$studwgt )  } ) ) )
# ma.wtd.sdNA(data=datlist, weights = datlist[[1]]$studwgt , vars = "hisei" )
# mean( unlist( lapply( datlist , FUN = function(data){
#         sqrt( Hmisc::wtd.var( data$hisei , data$studwgt ) ) } ) ) )
# 
# # standardize variable hisei to M=100 and SD=15        
# datlist1a <- scale_datlist( datlist=datlist, orig_var="hisei", trafo_var="hisei100", 
#                weights = datlist[[1]]$studwgt , M=100, SD=15 )
# 
# # check mean and SD
# ma.wtd.meanNA(data=datlist1a, weights = datlist[[1]]$studwgt , vars = "hisei100" )
# ma.wtd.sdNA(data=datlist1a, weights = datlist[[1]]$studwgt , vars = "hisei100" )
# 
# #--- do standardization for unweighted sample with books <= 3
# # -> define a weighting variable at first
# datlist0 <- mitools::imputationList( datlist )
# datlist2a <- within.imputationList( datlist0 , {
# 	         # define weighting variable
#                  wgt_books <- 1 * ( books <= 3 )
#                     } )
# 
# # standardize variable hisei to M=100 and SD=15 with respect to weighting variable
# datlist2b <- scale_datlist( datlist=datlist2a, orig_var="hisei", trafo_var="hisei100",
#          weights = "wgt_books" , M=100, SD=15 )
# 
# # check mean and SD (groupwise)
# ma.wtd.meanNA(data=datlist1a, weights = datlist[[1]]$studwgt , vars = "hisei100" )
# ma.wtd.sdNA(data=datlist1a, weights = datlist[[1]]$studwgt , vars = "hisei100" )
# 
# #--- transformation for a single dataset
# dat0 <- datlist[[1]]
# dat0a <- scale_datlist( datlist=dat0 , orig_var="hisei" , trafo_var="hisei100" , 
#                     weights = dat0$studwgt , M=100, SD=15 )
# stats::weighted.mean( dat0a[,"hisei"] ,  w = dat0a$studwgt )                    
# stats::weighted.mean( dat0a[,"hisei100"] ,  w = dat0a$studwgt )
# sqrt( Hmisc::wtd.var( dat0a[,"hisei100"] ,  weights = dat0a$studwgt ) )
# 
# #--- Standardizations for objects of class imputationList
# datlist2 <- mitools::imputationList(datlist)   # object class conversion
# datlist2a <- scale_datlist( datlist=datlist2 , orig_var="hisei" , trafo_var="hisei100" , 
#                weights = datlist[[1]]$studwgt , M=100, SD=15 )
#                
# #############################################################################
# # EXAMPLE 2: Standardized variables in list of nested multiply imputed datasets
# #############################################################################
#                
# # nested multiply imputed dataset in BIFIEsurvey package
# data(data.timss4, package="BIFIEsurvey")
# datlist <- data.timss4
# wgt <- datlist[[1]][[1]]$TOTWGT
# 
# # class nested.datlist
# imp1 <- nested.datlist_create( datlist )
# # class NestedImputationList
# imp2 <- NestedImputationList( datlist )
# 
# # standardize variable scsci
# imp1a <- scale_datlist( datlist=imp1, orig_var="scsci", trafo_var="zscsci", weights =wgt)
# # check descriptives
# ma.wtd.meanNA( imp1a , weights=wgt , vars = c("scsci" , "zscsci" ) )
# ma.wtd.sdNA( imp1a , weights=wgt , vars = c("scsci" , "zscsci" ) )
# 
# #############################################################################
# # EXAMPLE 3: Standardization of variables for imputed data in mice package
# #############################################################################
# 
# data(nhanes, package="mice")
# set.seed(76)
# 
# #--- impute nhanes data
# imp <- mice::mice(nhanes)
# #--- convert into datlist
# datlist <- mids2datlist(imp)
# #--- scale datlist (all variables)
# vars <- colnames(nhanes)
# sdatlist <- scale_datlist(datlist, orig_var = vars, trafo_var = paste0("z",vars) )
# #--- reconvert to mids object
# imp2 <- datlist2mids(sdatlist)
# #*** compare descriptive statistics of objects
# round( mean0( mice::complete(imp, action=1) ) , 2 )
# round( mean0( mice::complete(imp2, action=1) ) , 2 )
# ## End(Not run)