I'm trying to use regTermTest function in R survey package to test the significance of each variables for logistic regression.
However, I got a solver error for one of my variable, "fun". The error is
Error in solve.default(V) : 'a' is 0-diml
My code for the logistic regression is
model2 <-
glm(decision~samerace+race_o+field+goal+attr+sinc+intel+fun+amb+shar+like+prob,data=trg2, family=binomial)
regTermTest(model2, "fun")
I also encountered p = NA result for another variable "amb".
regTermTest(model2, "amb")
Wald test for amb
in glm(formula = decision ~ samerace + race_o + field + goal + attr +
sinc + intel + fun + amb + shar + like + prob, family = binomial,
data = trg2)
#F = NA on 1 and 5977 df: p= NA
My fun & amb variable are both numeric/integer from 1 to 10. They are a rating scale and their types are the same as some other variables such as sinc, attr etc. (the function works well for the other variables except "fun" & "amb")
