Assume your variable is "income"

xtile quartile=income, nq(4)
tabstat income, by(quartile) stat(n min max)      //to check your quartile

Sorry, for "quintile" :

xtile quintile=income, nq(5)
tabstat income, by(quintile) stat(n min max)      //to check your quintile

太谢谢了，我这就去试试。还想请教您为什么做回归的时候总是要把收入做quintile来分析呢？有的还在收入和财富前加log分析呢？

I assume that you are doing multiple linear regression. Most likely income is not in a normal distribution. Transforming it into quintile or a log transformation may make it fit well in the model (avoid the influence of its extremes and outliers), and make the results easy to explain. This is my good guess. For exact reasons, you have to check the literature.

Remember, multiple linear regression does not need each individual independent variable to be in a normal distribution. You can try "original income" in your model and check the model diagnostics yourself.

you are so professional. This explanation is so wonderful. Thank you so much.

I do logistic regression, if i have other question, could I ask for you? Thanks.