Input data must have class mids

Question

I'm working on a school project where I need to impute missing data and after the imputation with mice I'm trying to produce completed data sets with the complete-function.

When I run them one by one everything works fine, but I'd like to use a for-loop in case I want to have more than just m = 5 imputations. Now, when trying to run the for-loop, I always get the error

Error in complete(imputation[1]) : Input data must have class 'mids'.

However when I look up the class it is mids, what's going wrong here?

This is my code:

imputation <- mice(data = data, m = 5, method = "norm", maxit = 1, seed = 500) 
m <- 5
for(i in 1:m){
  completeData[m] <- complete(imputation[m])
  print(summary(completeData[m]))
}

Could someone maybe help me out here?

Answer 1

We are getting error because the class is not mids:

imputation[1]
# $call
# mice(data = walking, m = 5, maxit = 0, seed = 500)

class(imputation[1])
# [1] "list"

From the manual for ?complete:

Usage

complete(x, action = 1, include = FALSE)

library(mice)

# dummy data imputation
data(walking)
imputation <- mice(walking, max = 0, m = 5, seed = 500)

# using for loop
m <- 5
for(i in 1:m){
  completeData <- complete(imputation, m)
  print(summary(completeData))
}

# I prefer to use lapply
lapply(seq(imputation$m), function(i) summary(complete(imputation, i)))