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)