[R] logistic regression with package 'mice'

[Antonello Preti]

Dear all, I request your help to solve a problem I've encountered in using
'mice' for multiple imputation.
I want to apply a logistic regression model.
I need to extract information on the fit of the model.
Is there any way to calculate a likelihood ratio or the McFadden-pseudoR2
from the results of the logistic model?
I mean, as it is possible to extract pooled averaging and odds ratio...

Thank you in advance,
Antonello

Here an example of logistic regression on imputed data:


library(mice)

imp <- mice(nhanes)

# logistic regression on the imputed data

fit <- glm.mids((hyp==2)~bmi+chl, data=imp, family = binomial)

summary(fit)

summary(pool(fit)) ### pool averaging across all imputed dataset

summary(pool(fit, method = "rubin1987"))    ### pool across all imputed
dataset

### odds ratio

su <- summary(pool(fit, method = "rubin1987"))[,c(1,6,7)]

stime <- data.frame(exp(su))

names(stime) <- c("OR", "95% low", "95% high")

options(scipen=999)
stime
options(scipen=1)

	[[alternative HTML version deleted]]