Matching and Weighting Multiply Imputed Datasets

What’s New

Version 1.0.0

This is an update to improve documentation and to add several new features: 1. mimira and mimipo objects (the output of with() and pool(), respectively) now inherit from the mice classes mira and mipo. This means mice methods work with these objects. 2. coxph() when used with with() now correctly uses the robust standard errors. 3. A cluster argument has been added to with.mimids() to control whether cluster-robust standard errors should be used to account for pair membership when the model is a svyglm()-type model from the survey package. The default is to include pair membership when present and there are 20 or more unique subclasses (i.e., pairs). This works by supplying the pair membership variable (subclass) to the ids argument of svydesign(). 4. cbind() methods have been exported and documented. 5. mimids and wimids objects are now much smaller, now containing only the supplied mids object and the matchit() or weightit() outputs. and 6. Added trim() to trim estimated weights. This relies on WeightIt::trim() and uses the same syntax (thanks Nicolas!).

Version 0.9.3

This is an update to fix few bugs.

Version 0.9.2

This is an update to change the definition of the complete() function to evade name clashes with the tidyr package.

Version 0.9.1

This is an update to improve documentation and to fix minor bugs.

Version 0.9.0

This is an update to improve documentation and to implement compatibility for robust estimation of standard errors (compatibility with the svyglm() and svycoxph(), from the survey package, to be used as expressions within the with() function) and for new matching and weighting methods (e.g. the full, genetic, and cem matching methods as well as ebal and optweight weighting methods). Moreover, complete() function is included in the package to replace the and

Version 0.8.2

This is a spit and polish update to improve documentation and to fix minor bugs.

Version 0.8.1

This is the first release of the MatchThem package both on the Github and the Comprehensive R Archive Network (CRAN).


Farhad Pishgar

Noah Greifer

Clémence Leyrat

Elizabeth Stuart