methods-mitml.list: Concatenate lists of imputed data sets

In mitml: Tools for Multiple Imputation in Multilevel Modeling

Concatenate lists of imputed data sets

Description

These functions allow concatenating lists of imputed data sets by data set, row, or column.

Usage

## S3 method for class 'mitml.list'
c(...)
## S3 method for class 'mitml.list'
rbind(...)
## S3 method for class 'mitml.list'
cbind(...)

Arguments

...	One or several lists of imputed data sets with class mitml.list as produced by mitmlComplete (or similar).

Details

The c, cbind, and rbind functions allow combining multiple lists of imputed data sets in different ways. The c method concatenates by data set (i.e., by appending additional data sets to the list), rbind concatenates by row (i.e., appending additional rows to each data set), and cbind concatenates by column (i.e., by appending additional columns to each data set).

These functions are intended for experienced users and should be used with caution. Appending rows or columns from multiple imputation procedures is usually unsafe unless in special applications (see 'Examples').

Value

A list of imputed data sets with an additional class attribute mitml.list.

Author(s)

Simon Grund

Examples

# Example 1: manual imputation by grouping variable

data(studentratings)
fml <- ReadDis + SES ~ ReadAchiev + (1|ID)

imp1 <- panImpute(subset(studentratings, FedState == "SH"), formula = fml,
                 n.burn = 1000, n.iter = 100, m = 5)

imp2 <- panImpute(subset(studentratings, FedState == "B"), formula = fml,
                  n.burn = 1000, n.iter = 100, m = 5)

implist1 <- mitmlComplete(imp1)
implist2 <- mitmlComplete(imp2)

rbind(implist1, implist2)

# Example 2: predicted values from linear model

imp <- panImpute(studentratings, formula = fml, n.burn = 1000, n.iter = 100, m = 5)
implist <- mitmlComplete(imp)

pred <- with(implist1, predict(lm(ReadDis ~ ReadAchiev)))
cbind(implist, pred.ReadDis = pred)

mitml documentation built on March 31, 2023, 7:01 p.m.