Categorize income variable due to quartiles

[Larissa Klee]

Hello,

I assume the solution is very simple, but I'm just standing at a loss....

I have to categroize the income variable due to the quartils-distance. So the first category should include all people whose income is lessequal to p25. Category 2 should include all between p25-p75 and category 3 all n whose income is higher than p75.

I suppose there is a stata command which I just don't find / get.

Thank you so much in advance!

Best regards,
Larissa

[Nick Cox]

Use xtile to throw away information as desired and then recode to throw away yet more.

. sysuse auto, clear
(1978 Automobile Data)

. xtile binned=mpg, nq(4)

. tabstat mpg, by(binned) s(n min max)

Summary for variables: mpg
     by categories of: binned (4 quantiles of mpg)

  binned |         N       min       max
---------+------------------------------
       1 |        27        12        18
       2 |        11        19        20
       3 |        22        21        25
       4 |        14        26        41
---------+------------------------------
   Total |        74        12        41
----------------------------------------

. recode binned 2/3 = 2 4 = 3, gen(binned2)
(36 differences between binned and binned2)

. tab binned*

         4 | RECODE of binned (4 quantiles of
 quantiles |               mpg)
    of mpg |         1          2          3 |     Total
-----------+---------------------------------+----------
         1 |        27          0          0 |        27
         2 |         0         11          0 |        11
         3 |         0         22          0 |        22
         4 |         0          0         14 |        14
-----------+---------------------------------+----------
     Total |        27         33         14 |        74

This example is typical: ties can result in bins with unequal frequencies in practice.