SPSS Statistics

Genlin warning: "There are no valid cases for the log link function"

  • 1.  Genlin warning: "There are no valid cases for the log link function"

    Posted Tue August 25, 2020 05:23 PM
    Hi, I'm conducting nested negative binomial regression models and am getting a strange error at the first step with 4 independent variables ("There are no valid cases for the log link function. Only the iteration history is displayed. Execution of this command stops.") When I specified that the dispersion variable be held constant at 1, there is no problem. However, I want to have the dispersion variable freely estimated and this is where I keep getting the error.

    When I run the second step model with 6 variables, there is no problem with the model converging. The same is true for the 3rd (10 variables) and 4th models (12 variables) that have more IVs.

    I have selected only participants who have complete data at the final step and I'm still getting this error.

    Does anyone have a solution or recommended fix? Thanks for your guidance in advance.



    ------------------------------
    Evan
    ------------------------------