To construct a composite variable from multiple categorical variables

[Enoch Park]

My question may be simple. I tried to find a possible answer to my question, but failed to do so.

I am trying to construct a composite variable from multiple categorical variables. I want to get a single value (e.g., mean) from multiple categorical variables.
For instance, I want to estimate the effect of self-directed learning (SDL) on academic achievement. Both are categorical variables. There are five questions that are intended to assess the extent of SDL ability. So I would like to combine the five questions into one variable.

How can I do this in STATA?
egen, concat or egen, group are not what I am looking for because they do not give me a single value(number).
If they are continuous variables, egen, rowmean will work.

Or simply combining multiple categorical variables into a single variable is problematic, isn't it?
Please help me with this.

[Joseph Coveney]

Maybe use a simple sumscore of the five questions' response? Or perhaps something a little more involved, such as a MIMIC-like structural equation model using gsem. See the help file and user's manual for examples.

[Maarten Buis]

An intermediate option to the ones mentioned by Joseph Coveney is a sheaf coefficient, see ssc desc sheafcoef

[Joro Kolev]

And if what Joseph and Maarten are proposing is not what you are looking for, please provide a data sample using -dataex-, and explain how exactly this combined variable that you desire has to look like.

[Sonnen Blume]

An intermediate option to the ones mentioned by Joseph Coveney is a sheaf coefficient, see ssc desc sheafcoef

This is an exceptionally useful concept. Do you have any paper/report using this command showing a real-life use, just to get a better sense of its application. Thank you.