I am running a logistic regression on financial data of several companies. The dependent variable is the company to be a smoother (binary variable: 0, 1), and there are several independent variables: company age, company size, number of board meetings, etc. I tried to run a glm that detect any complete separation in this model through the R-package "detectseparation." The code in R was:

detect<-glm(formula=smth~0+Big4+age+lev+assets_log+committee+PctIndependentDirectors+meetings+boardsize+afe+value_log+revgrowth_log,data=data2,family = binomial("logit"),method = "detect_separation")

the result was: enter image description here

We can see that there is always 0 next to each IV. What does 0 mean? How about -inf and +inf?


1 Answer 1


As it says in the printout, these values indicate existence of the maximum likelihood estimates. 0 means they are finite (i.e. no complete separation), whereas infinite values are for completely separated predictors (0 for -Inf and 1 for Inf respectively).

You should use the standard glm methods like summary and anova to extract parameter estimates and inferences - all this object's print shows you right now is that there don't seem to be any complete separation issues.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.