generating a composite variable with a continous and a categorical variable

[Ikenna Egwu]

Dear Statalist,

To measure my dependent variable, I would like to generate a composite variable comprising of two variables - a continous and an ordinal categorical variable. The ordinal categorical variable ranges from 0 to 1. Please what is the right way of doing this?

Thank you.

Ikenna

[Joseph Coveney]

what is the right way of doing this?

Unless you know that the distance (however defined) between ordered categories is constant, then I suspect that there isn't one.

You might be better off handling them in a model that keeps them as-is, for example using gsem.

version 16.1

clear *

set seed `=strreverse("1539915")'

sysuse auto
summarize rep78, meanonly
quietly replace rep78 = runiformint(r(min), r(max)) if missing(rep78)

// Something along these lines
gsem (rep78 <- c.length, oprobit) (weight <- c.length), nocnsreport nodvheader nolog

exit

[Ikenna Egwu]

Alright, thanks