Combine Multiple Imputation Inferences

DESCRIPTION:

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

USAGE:

miMeanSEList(object, estimate = "estimate", se = "se", 
             df = "df", n = "n", sse = F, cov) 

REQUIRED ARGUMENTS:

object
an mi object whose imputations are lists.

OPTIONAL ARGUMENTS:

estimate, se,
df, n, cov
these arguments may be either character strings containing names of components of imputations of object which contain estimates, standard errors, degrees of freedom, sample size, and covariance matrices, respectively, or may be objects containing those values.
sse
if TRUE, then the numerical values for se are assumed to contain squared standard errors.

VALUE:

a list, whose components are described in the help file for . The first two components are:
est
average of the estimates across imputations.
std.err
standard errors for the estimates.

DETAILS:

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

estimate, and at least one of se and cov , are required. 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 
miMeanSEList(miEval(list(estimate = coefs[,1], 
                         se = coefs[,2])))