plot - How to plot interaction of logistic regression model (glm) with multiple imputed data (MICE)?

1
votes

I created an interaction term with iv*sex and imputed the data with mice. Then used the imputed data to run a logistic regression model (glm):

model <- with(data=imp, glm(dv~control+iv+sex+iv*sex, family="binomial"))

The following are the abbreviations of the variable names: dependent variable=dv, independent variable=iv, moderator=sex, interaction term= iv*sex

There is significant interaction for iv*sex and I would like to plot a graph for the interaction but I couldn't find how to. It will be greatly appreciated if any solutions is offered. Thanks!

1
rplotglminteractionr-mice

1 Answers

0
votes

I've just run into the same issue, and with effects package I solved it.

e <- effects::effect("iv*sex", model)
e <- as.data.frame(e)
ggplot2::ggplot(e, ggplot2::aes(iv, fit, color=sex, group = sex)) + 
  ggplot2::geom_point() + 
  ggplot2::geom_line() +

"fit" is your dependent variable, in your case "dv".