On Sunday, May 18, 2003, at 02:33 AM, Taeho wrote:
I want to get "inequal" values such as coefficient of variation and
Gini
coefficient for multiple units such as towns, cities, etc in a dataset
which has house selling prices over 1000 cities and towns.
Stata ado files don't allow "by" options. Please let me know the
easiest
way to get the values for multiple units. I am a STATA bigger having
limited knowledge in programming.
Thanks
It is NOT true that user-written files do not allow by _prefixes_,
which is the usual way to 'loop' a command over a by-group. I have
taken the liberty of hacking Rich Goldstein's _rspread_ (a blast from
the version 3.0 past, published in STB-23) to make it 'byable', in the
Stata parlance (see [p] byable). I believe it calculates all the stats
that are mentioned. It only prints out the results for each by-group
(and having a byable routine generate new variables is slightly harder)
but the easiest way to return a volume of stats such as this is in a
Stata matrix (which can always be unwound into variables). I have not
made that change yet. I have also disabled the graph option in my
version.
I believe the routine is working properly (or at least replicating
Rich's results, for the cases for which I have tried it). You can
validate his against mine by running rspread with an 'if' clause, which
should match the equivalent by-group. The new routine, rspread2, is
written for Stata 7. Caveat executor; since Rich lives around the
corner, perhaps we'll get together for a beer and clean this routine
up, if there is interest in bringing it fully up to date.
To get the new ado-file,
copy http://fmwww.bc.edu/ec-p/baum/cert/rspread2.ado rspread2.ado
Also get the rspread package from ssc, since it contains some necessary
auxiliary routines and the help file.
E.g.
. use http://fmwww.bc.edu/ec-p/data/hayashi/griliches76.dta
. bys s year: rspread2 lw
________________________________________________________________________
_______
-> s = 9, year = 66
Measures of Absolute and Relative Dispersion (or Inequality):
| Mean Dev. about
Max.
variable| Mean Median MeanDif CV CD Gini SEMean
% Dev.
--------
|---------------------------------------------------------------------
lw | .32 .32 .46 0.0752 0.0626 0.0443 .13
13.17%
________________________________________________________________________
_______
-> s = 9, year = 67
Measures of Absolute and Relative Dispersion (or Inequality):
| Mean Dev. about
Max.
variable| Mean Median MeanDif CV CD Gini SEMean
% Dev.
--------
|---------------------------------------------------------------------
lw | .32 .32 .63 0.0869 0.0614 0.0614 .32
6.14%
...
Thanks
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/
