Problem with Firth Logistic Regression

stdald15

New Member
Hello all,

I am having some issues trouble shooting a Firth Logistic regression model I have run on municipal level data. So, I have a dataset on all municipalities under analysis and I am looking at an event with low frequency (81), while the number of non-events is more than 2000.

The issue I keep receiving an infinite estimate for the model intercept, while the IVs are ok (at least they look ok to me).

Any help troubleshooting this would be much appreciated. I could definitely use a fresh pair of eyes.

Here is the output I am stuck from R (logistf). I also attached the printout as the table below comes out a little convoluted. The covariances didn't strike me as particularly troublesome (though I'm all ears if someone disagrees). Also, I used the standardized option in the model for the IVs. All the variables are continuous.

Model fitted by Penalized ML
Confidence intervals and p-values by Profile Likelihood Profile Likelihood Profile Likelihood Profile Likelihood

coef se(coef) lower 0.95 upper 0.95 Chisq p
(Intercept) -3.947938e+00 1.643859e-01 -4.283353e+00 -3.639557e+00 Inf 0.000000e+00
Total_Fixed_Ass 3.923257e-09 4.396373e-09 -3.774015e-09 1.342160e-08 0.9401887 3.322293e-01
Infrastructure 3.079252e-01 5.709936e-02 1.993128e-01 4.240471e-01 44.1480759 3.044509e-11
Salary_Rate 7.017842e-03 1.662845e-03 3.883563e-03 1.033340e-02 16.7389526 4.289128e-05

Likelihood ratio test=109.3604 on 3 df, p=0, n=2454
Wald test = 76.8872 on 3 df, p = 1.110223e-16

Covariance-Matrix:
[,1] [,2] [,3] [,4]
[1,] 2.702274e-02 3.614560e-12 -8.852098e-04 -1.726696e-04
[2,] 3.614560e-12 1.932810e-17 -1.292892e-10 -1.467278e-12
[3,] -8.852098e-04 -1.292892e-10 3.260337e-03 -5.979337e-07
[4,] -1.726696e-04 -1.467278e-12 -5.979337e-07 2.765053e-06