Combine Multiple Imputation Inferences

DESCRIPTION:

Average estimates, and compute combined standard errors or covariance estimates from multiple imputations.

USAGE:

miMeanSEMatrix(object, columns = 1:2, ...) 

REQUIRED ARGUMENTS:

object
an mi object whose imputations are matrices with at least two columns, one of which contains parameter estimates and the other containing standard errors for those parameters.

OPTIONAL ARGUMENTS:

columns
integer vector of length two, indicating which columns of the matrices contain the pameter estimates and standard errors.
...
additional arguments which are passed to miMeanSEDefault, including df and n, for degrees of freedom and sample size, respectively.

VALUE:

a list, whose components are described in the help file for .

DETAILS:

This function extracts data from object, then calls miMeanSEDefault to perform calculations.

If df and n are not supplied (or the corresponding elements of object do not exist) then some components of the output will contain missing values.

SEE ALSO:

, , , , .

EXAMPLES:

fit <- miEval(lm(chol14~., data = cholesterolImpExample)) 
sumfit <- miEval(summary(fit)) 
coefs <- miEval(coef(sumfit)) 
coefs[[1]]  # note estimates and std errors 
# pass matrices, cols 1:2 are estimate & std err 
miMeanSEMatrix(coefs) 
miMeanSEMatrix(coefs, df = nrow(cholesterolImpExample)-3)