Categorical explanatory variable and dummies

[Edward Smith]

Hi all. I am new here so excuse me if I do something wrong.
Well, I am a newbie in econometrics and stata too. I have a case of probit model where one of my explanatory variables is a categorical variable (number of siblings, 0-5). I just put it in the code for running the probit reg as i.nsib and in the output I get nsib and under this, 1,2,3,4,5 with their coeff, z values etc. So does it mean that I have 5 dummies, one for every number of siblings? For instance, does it mean that when someone has 3 siblings the 3.nsib will be 1 while the other will be 0?

One more thing I want to ask is how can I interpret the above case if nsib 1,2,3,4 are not significant while nsib 5 is significant?

Thanks in advance!

[Marcos Almeida]

Welcome to the Stata Forum / Statalist.

When you used factor notation, zero is the reference. Therefore, compared to 'no sibling' , having 5 siblings provided a significant coefficient.

With regards to your first question, yes, for 6 levels, that would be like you have 5 dummies.

You surely risk getting more insightful advice, shall you share command and output, as recommended in the FAQ.

[Carlo Lazzaro]

Edward:
as an aside to Marcos' helpful advice, you're implicitly wondering about the so called dummy trap (please, see https://en.wikipedia.org/wiki/Dummy_...le_(statistics)), which is covered in any decent econometrics textbook.
I would also add that both -margins- and -marginsplot- can favour a better understanding of your regression output.

[Edward Smith]

Dear Marcos and Carlo thank you very much for your response. I understand the dummy variable trap (that can cause collinearity issues), all I wanted to know is whether every response for the number of siblings (1,2,3,4,5) is a dummy and whether the case of 0 siblings works as a base category so it is clear now.