I decided to add a feature that might be useful for some people. Sometimes it's easier for people to create a list of matchit or weightit objects by running the functions on imputed datasets rather than using matchthem() or weightthem(), but that leaves the user with a list of objects rather than something that cobalt functions can interpret. I added the functions as.mimids() and as.wimids() which allow a use to supply a list of such objects and the imputation object (e.g., the mids object), and the function constructs a mimids or wimids object, which work as though matchthem() or weightthem() had been called on the original imputation object with approach = "within".
I also updated the documentation to use markdown, which makes it easier to read and update. I made a few small changes here and there to the code, but nothing else major. I also bumped the version number up to 1.2.0. It should be ready to submit.
Hi Farhad,
I decided to add a feature that might be useful for some people. Sometimes it's easier for people to create a list of
matchit
orweightit
objects by running the functions on imputed datasets rather than usingmatchthem()
orweightthem()
, but that leaves the user with a list of objects rather than something thatcobalt
functions can interpret. I added the functionsas.mimids()
andas.wimids()
which allow a use to supply a list of such objects and the imputation object (e.g., themids
object), and the function constructs amimids
orwimids
object, which work as thoughmatchthem()
orweightthem()
had been called on the original imputation object withapproach = "within"
.I also updated the documentation to use markdown, which makes it easier to read and update. I made a few small changes here and there to the code, but nothing else major. I also bumped the version number up to 1.2.0. It should be ready to submit.
Noah