Sure. One of the extraordinary features of
this method is that it has been re-invented
numerous different times and under numerous
different names. The oldest use I know is
by G.U. Yule about 1910. Deming's use with
F. Stephan circa 1940 appears to have been
unaware of that, but Yule just used it in
passing and didn't sound a fanfare --
didn't even toot a toy trumpet once --
about what he was doing.
One implementation is -mstdize- which has
been on SSC since 1998.
There is also a matrix-based version
called -mstdizem- which is also on
SSC, as part of the -matodd- package.
I also have a Mata-based -mstdizem2-
which isn't in the public domain.
In this case my impression is that
most problems people run are of moderate
size and Mata doesn't offer much of a
speed-up, just fun for the programmer.
Ade [Adrian] Mander approaches the problem from
a different direction in a program of his.
But although either might be a direct way
of approaching your problem, in principle you
shouldn't need a special program. You should just
be able to use -glm- to do it by fitting a loglinear model.
I forget the details, but it hinges on using
offsets, and I once did it by looking by looking
at Alan Agresti's example with SAS code in his
main categorical text (Categorical data analysis,
Wiley) and mimicking it in Stata.
The SAS code in this case can't have been that
obscure; otherwise I wouldn't have been able to
translate a language I never use into one I do use
a lot.
Nick
[email protected]
Dick Campbell
> I wish to do Deming adjustment of a set of tables. This
> methods uses iterative proportional scaling to adjust
> the frequencies in a sample table to a set of known marginals.
> I used to have a fortran program for it years ago, but is has
> disappeared in the course of a move. Is anyone aware of a
> Stata routine for doing this?
*
* 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/
