Interaction with mitools

[angelrodriguez2020]

Dear fellows,
I have a set of imputed files created with MICE that I need to analyze with survey. When I use mitools to transfer them:

> library(mitools)
> imputedsurv<- imputationList(lapply(1:5, complete, x = imp))

This is what I get:

Error in UseMethod("complete_") :
no applicable method for 'complete_' applied to an object of class "c('integer', 'numeric')"
In addition: Warning message:
complete_() is deprecated as of tidyr 1.0.0.
Please use complete() instead.

Any idea why this is happening?

Thank you,

Ángel

[stefvanbuuren]

mice doesn't use the long-ago deprecated tidyr::complete_. My suggestion is to direct your question to the mitools authors. Or live with the warning.

[angelrodriguez2020]

Dear Stef,

This is the answer I got from the author of mitools. Apparently, there is a typo in the syntax you suggest in: mice: Multivariate Imputation by Chained Equations in R. Journal of Statistical Software, December 2011, Volume 45, Issue 3, p.53

"The code you were given

imputedsurv<- imputationList(lapply(1:5, complete, x = imp))

is incorrect, because the data argument to mice::complete is actually called 'data'. If you do

imputedsurv<- imputationList(lapply(1:5, complete, data = imp))

it will work. The reason for the mysterious calls to tidyr is that mice imports the generic function complete() from tidyr. If your call to complete() doesn't have a mids object for the data= argument (because you've called it 'x'), the complete.mids() method will not be called. Instead the tidyr::complete.default method will be called.

-thomas

Thomas Lumley
Professor of Biostatistics
University of Auckland"

[stefvanbuuren]

Thanks for clearing this up.

1. I wasn't aware that the code you cited came from our mice 2011 JSS paper.
2. Add S3 method for complete.mids to improve integration with tidyverse #212 changed complete() to evade interference with tidyr::complete(), and I had to rename the first argument. Back in 1999 when I wrote complete() I couldn't have dreamt that that would be needed some day.
3. I cannot retrospectively change the JSS paper, so I think we need to live with it.