I would like to include the following variables in my quantile regression model:

y=real hourly wages x1= sex (male=1, female=2) x2= yeaedu (years of education) x3= race (has many categorie, around 8)

I run regression:

regression I wonder how I should include variable race in this model? Should I create first a dummy variable for each category ? Or is there a formula/shortcut in state?

I assume that the other two variables sex and years of edu are correctly included the model?

How do I interpret coefficient of var sex?

Thanks a lot!


You should dummy it out:

  • For sex, make a dummy = 1 if women, of 1 if male. Only include one of these. Stata might do this on it own, you have data setup correctly. But I can't tell from the picture
  • For race you do the same, include seven dummy variables all equal = 1 if said person if of that race. And you leave one group out as reference.

Note that it does not matter which race or sex you include in the regression, the result will always be the same.

| cite | improve this answer | |
  • $\begingroup$ If my dependent variable in quantile regression is skewed, should use a log transformation of my y? here y is hourly wage. $\endgroup$ – Student3456 May 7 '16 at 19:41
  • $\begingroup$ Depends on your purpose, one of the reasons for median regression is that it is much less sensitive to outliers so you might not need to do anything. $\endgroup$ – Repmat May 7 '16 at 19:45

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.