Expressing dummy variables in multiple regression

[Lynn Zhang]

Hi!

I have put in the following comand into STATA:

reg ln_real_income year_2018_19 year_2019_20 Female Mature Asian Black Mixed other_ethnicity unknown_ethnicity first_class lower_second_class third_class unclassified

All of the variables, except ln_real_income, are dummy variables and take the the value of either 1 or 0. My question is how would I go about writing this regression in a formula form? Is this something STATA can do?

Thanks!

[Andrew Musau]

A dummy variable (also known as an indicator) is still a variable. So you express it as any other variable. So the regression

regress ln_real_income Female Mature Asian Black

represents the population model:


$$\text{ln_real_income}_i= \beta_0+ \beta_1 \text{Female}_i+ \beta_2 \text{Mature}_i+ \beta_3 \text{Asian}_i+ \beta_4 \text{Black}_i + u_i\;\;\;(i=1, \cdots, N)$$

where the subscript \(i\) indexes individuals and \(u\) is the error term. You will want to explicitly identify the base categories (not included in the equation) explicitly in the text.

Is this something STATA can do?

Official estimation commands in Stata do not write equations.

[Andrew Musau]

On the spelling of Stata, see https://www.statalist.org/forums/help#spelling

[Lynn Zhang]

Got it, thank you!