Presenting logistic regression tables for non-binary categorical variables with multiple coefficients

[Dawudi Goaliath]

Hello everyone,

I'm new to the forum so forgive me my rudimentary description of the problem.

For my logistic regression model I have three non-binary categorical predictors (income group, education, and volunteering/month) with more than three categories. The resultant output for the logit command when specifying income, education, and volunteering as factor variables (e.g. i.educ), obviously displays coefficients for each category of the predictors (e.g. low income/middle-income/high-income). My question now is how to present findings for these variables elegantly; is there any other way than to include all coefficients for each variable when using the estout command for example? By doing this, the table would obviously be extremely large. What is the standard procedure here? Skimming through regression tables in scientific articles, I find that non-binary categorical predictors sometimes nonetheless exhibit only on coefficient.

Thanks for any help in advance.

Dawudi

[Carlo Lazzaro]

Dawudi.
welcome to the list.
As far as the coefficient of each three-level predictor is concerned, in general Stata reports 2 out of 3 of them, leaving the reference category 0 (-omitted-).
As far as I know, articles published in decent journals (peer-reviewed) reports in the regression outcome table all but the reference category coefficient for categorical variables.

[Dawudi Goaliath]

Thank you very much Carlo!

[Maarten Buis]

In many cases where non-binary categorical variables have one parameter, these variables were added as if they are a continous variable. For example, education as years of education.

[Rita Ryan]

I used estout and esttab and still just get the same coefficients lissted in the main logit output. I am looking for individual non-binary categorical variable coefficient- I can see from the professor's example this automatically comes in the output from SPSS but I can't see how to get it from Stata.

[Clyde Schechter]

Duplicate post. See https://www.statalist.org/forums/for...n-coefficients if interested.