Using bal.tab() with Multiply Imputed Data
When using bal.tab()
with multiply imputed data, the output will be different from the case with a single data set. Multiply imputed data can be used with all bal.tab()
methods, and the mimids
and wimids
methods for MatchThem objects automatically incorporate multiply imputed data. This page outlines the outputs and options available with multiply imputed data.
There are two main components of the output of bal.tab()
with multiply imputed data: the within-imputation balance summaries and the across-imputation balance summary. The within-imputation balance summaries display balance for units within each imputed data set separately. In general, this will not be very useful because interest rarely lies in the qualities of any individual imputed data set.
The across-imputation balance summary pools information across the within-imputation balance summaries to simplify balance assessment. It provides the average, smallest, and largest balance statistic for each covariate across all imputations. This allows you to see how bad the worst imbalance is and what balance looks like on average across the imputations. The summary behaves differently depending on whether abs
is specified as TRUE
or FALSE
. When abs = TRUE
, the across-imputation balance summary will display the mean absolute balance statistics and the maximum absolute balance statistics. When abs = FALSE
, the across-imputation balance summary will display the minimum, mean, and maximum of the balance statistic in its original form.
There are four arguments for each bal.tab()
method that can handle multiply imputed data: imp
, which.imp
, imp.summary
, and imp.fun
.
imp |
A vector of imputation membership. This can be factor, character, or numeric vector. This argument is required to let |
which.imp |
This is a display option that does not affect computation. If |
imp.summary |
This is a display option that does not affect computation. If |
imp.fun |
This is a display option that does not affect computation. Can be "min", "mean", or "max" and corresponds to which function is used in the across-imputation summary to combine results across imputations. For example, if |
The output is a bal.tab.imp
object, which inherits from bal.tab
. It has the following elements:
Imputation.Balance |
For each imputation, a regular |
Balance.Across.Imputations |
The balance summary across imputations. This will include the combination of each balance statistic for each covariate across all imputations according to the value of |
Observations |
A table of sample sizes or effective sample sizes averaged across imputations before and after adjustment. |
As with other methods, multiple weights can be specified, and values for all weights will appear in all tables.
Noah Greifer
Please choose more modern alternatives, such as Google Chrome or Mozilla Firefox.