Create four quartiles

[Soumya Upadhyay]

I have a longitudinal data set that contains data from 2013, 2014 and 2015. There is one variable that contains the number of employed physicians. I would like to create a separate categorical variable that contains 4 quartiles on the number of employed physicians. Please help and let me know how can I do that. Thank you.

[Clyde Schechter]

-help pctile-

[Soumya Upadhyay]

-help pctile-

I wrote the following code:
pctile hsp_quartile=ftehsp,nquantiles(4)

But, it doesn't work. I am looking to create a categorical variable that contains 0,1,2,3 as four categories that represent four quartiles on ftehsp variable. Please help.

[Stephen Jenkins]

Part of your confusion is terminological. What you really want is a variable identifying quantile groups, not quantiles. Remember: there are 3 quantiles -- the 3 values that define 4 quantile groups. (Similarly, there are 9 deciles defining 10 decile groups, and so on.) Whatever, Clyde Schechter directed you to exactly the right place, i.e. help pctile. Look at the discussion for xtile.

[Chris Engel]

I wrote the following code:
pctile hsp_quartile=ftehsp,nquantiles(4)

But, it doesn't work. I am looking to create a categorical variable that contains 0,1,2,3 as four categories that represent four quartiles on ftehsp variable. Please help.

Just do:

xtile hsp_quartile=ftehsp,n(4)

and if you're married to using 0 as start number just do some replace'ing of hsp_quartile

[Attaullah Shah]

If you have a large data set and want some speed advantage, you can use astile from SSC. astile is much faster compared to Stata's official xtile

ssc install astile
astile hsp_quartile=ftehsp,nq(4)

And if you want to decrement quantile categories by one, then after generating hsp_quartile, do the following

replace hsp_quartile = hsp_quartile -1