When I do a binary logistic regression to calculate adjusted odds ratios, I get the warning "Due to redundancies, degrees of freedom have been reduced for one or more variables." I'm aware that's due to collinearity and I've looked at the iteration history to see which variables have been affected, but I want to know which specific categories have been omitted from the affected variables. All my predictor variables are categorical, and I used the first-coded category as the reference category (if relevant). Is there any way to which categories have been dropped from the analysis or is that not even relevant since it's collinear with another variable? Thank you!

The Variables in the Equation table shows you which variables were still included, and categorical variables that were entirely excluded are not shown in the Categorical Variables Codings table.

Of course, less than perfect collinearly may still be present in the included regressors: the omitted ones would have been the perfectly collinear ones.

Hi, thank you for the reply! None of the variables were entirely excluded; rather, some values/categories of specific categorical variables (ie. White, Black, Asian for a person's race) were excluded. All potential values/categories are shown in the Variable Codings table, so I'm not sure which were excluded. Do you know if there's a way to find this out?

The included categories for a variable are shown in the table of coefficients in the Variables in the Equation table in parentheses after the variable name. The numbers refer to the Categorical Variables Codings table.