i.varname - Factor specification in Regression analysis

[Anton Ivanov]

Hello forum members!

I have fit the following model:

regres pr size hsa smf fb_star alexa i.census i.notforprofit, vce(robust) beta, where:

i.census - factor with 4 groups (values 1-4), and i.notforprofit - factor with 2 groups (binary coded), all other variables are numerical.

In the results table below - the factor "census" - was it automatically dichotomized by STATA? Shall I interpret the results for factors just as I would do it for dummies?

Thank you!

[Rich Goldstein]

as you can see, Stata has made 3 indicator (dummy) variables for census (which, with 4 categories, is exactly what you would expect); by default, Stata has chosen "1" as the reference group, but you can change this if you want; see -h fvvarlist-

[Anton Ivanov]

Thank you, Rich. That is exactly what I wanted to double-check.