Combines estimates and standard errors from m complete-data analyses performed on m imputed datasets to produce a single inference. Uses the technique described by Rubin (1987) for multiple imputation inference for a scalar estimand.
mi.inference(est, std.err, confidence=0.95)a list with the following components, each of which is a vector of the
same length as the components of est and std.err:
the average of the complete-data estimates.
standard errors incorporating both the between and the within-imputation uncertainty (the square root of the "total variance").
degrees of freedom associated with the t reference distribution used for interval estimates.
P-values for the two-tailed hypothesis tests that the estimated quantities are equal to zero.
lower limits of the (100*confidence)% interval estimates.
upper limits of the (100*confidence)% interval estimates.
estimated relative increases in variance due to nonresponse.
estimated fractions of missing information.
a list of $m$ (at least 2) vectors representing estimates (e.g., vectors of estimated regression coefficients) from complete-data analyses performed on $m$ imputed datasets.
a list of $m$ vectors containing standard errors from the
complete-data analyses corresponding to the estimates in est.
desired coverage of interval estimates.
Uses the method described on pp. 76-77 of Rubin (1987) for combining the complete-data estimates from $m$ imputed datasets for a scalar estimand. Significance levels and interval estimates are approximately valid for each one-dimensional estimand, not for all of them jointly.
See Rubin (1987) or Schafer (1996), Chapter 4.