st: Re: dummies from continuous variable

[baum <baum@bc.edu>]

--On Thursday, August 15, 2002 2:33 -0400 Deborah wrote:

> In STATA, however, the only way I know to create the same variable is to =
> use gen:
>
> gen age2534 =3D (age>=3D25 & age < 35)
>
> Question:  Is there a better way to generate indicator variables from =
> continuous ones?
A thread earlier this week spoke about doing this from the percentiles of 
the continuous variable (pctile, xtile, _pctile). But if you have arbitrary 
values where you want to bracket the c.v., these two commands will do it:

. gen brackets=recode(mpg,14,18,23,29,35)

. tab brackets, gen(mpgdumm)

  brackets |      Freq.     Percent        Cum.
------------+-----------------------------------
        14 |          8       10.81       10.81
        18 |         19       25.68       36.49
        23 |         24       32.43       68.92
        29 |         16       21.62       90.54
        35 |          7        9.46      100.00
------------+-----------------------------------
     Total |         74      100.00


. summ mpgdumm*

   Variable |     Obs        Mean   Std. Dev.       Min        Max
-------------+-----------------------------------------------------
   mpgdumm1 |      74    .1081081   .3126365          0          1
   mpgdumm2 |      74    .2567568    .439826          0          1
   mpgdumm3 |      74    .3243243   .4713172          0          1
   mpgdumm4 |      74    .2162162   .4144734          0          1
   mpgdumm5 |      74    .0945946   .2946518          0          1

And if you want to apply the same cut-points to a number of variables:


. local brack "14,18,23,29,35"

. g brack2=recode(mpg,`brack')

Kit


*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/