handleNA {matchMulti} | R Documentation |
Handle Missing Values
Description
Preprocesses a dataframe of matching covariates so the Mahalanobis distance can be calculated.
Usage
handleNA(X, verbose = FALSE)
Arguments
X |
a matrix or dataframe of covariates to be used for matching |
verbose |
logical value indicating whether detailed output should be provided. |
Details
Preprocessing involves three main steps: (1) converting factors to matrices of dummy variables (2) for any variable with NAs, adding an additional binary variable indicating whether it is missing (3) imputing all NAs with the column mean. This follows the recommendations of Rosenbaum in section 9.4 of the referenced text.
Value
a matrix containing the preprocessed data.
Author(s)
Luke Keele, Penn State University, ljk20@psu.edu
Sam Pimentel, University of California, Berkeley, spi@berkeley.edu
References
Rosenbaum, Paul R. (2010). Design of Observational Studies. Springer-Verlag.