miSummary(miList containing lm objects) miSummary.lm(object1, object, correlation = T)
lm object, which should be the same as the
first imputation from
object.
lm objects.
call and
terms components of the
lm objects.
lm objects.
df.orig because of additional uncertainty due
to missing values.
sigma^2
produces an estimated covariance matrix for the
coefficients.
Some of the components above are
miList objects, containing
results for each imputation. Others are consolidated results,
combining information from all imputations. In particular,
sigma is the square root of the average squared value of
sigma
from the multiple imputations,
r.squared is computed as the ratio of averages
of numerator and denominator sums of squares across imputations,
fstatistic,
df.coef, and
cov.unscaled
are obtained using consolidation methods described in the references
below.
Hesterberg, T. (1998),
Combining multiple imputation t, chi-square, and F inferences ,
Insightful Technical Report number 75.
Rubin, D. B. (1987),
Multiple imputation for nonresponse in surveys ,
John Wiley, New York.
Schafer, J. L. (1997),
Analysis of Incomplete Multivariate Data ,
Chapman & Hall, London.
fit <- miEval(lm(chol14~., data = cholesterolImpExample)) miSummary(fit)