I am trying to update my gologit2 program to work with Stata 9's new
-svy- features and svy: prefix. It is turning out to be more
difficult than I expected. Test statistics are coming out wrong and
some options (e.g.autofit) are not working right.
I'm curious, then, exactly how does svy do its magic? Does it go
line through line of your program, adding an svy: prefix to every
command that supports it, e.g. ologit? Or does it wait until the
entire program has run, and do something with the final stored results?
There is a short chapter on adding SVY support in the 3rd edition of
the Stata Maximum Likelihood book, but it appears that gologit2 will
need more than what is described in that chapter. If I understood
better how svy works I might be able to pinpoint the problems better.
Incidentally, gologit2 has always supported a svy option, e.g.
gologit2 y x1 x2 x3, svy
I seem to be able to get that to work fine with Stata 9 svyset
files. But, the svy: prefix continues to cause me problems. In
short, one way or another I think I will get my program to work with
the new svy features, but getting it to work in the officially
blessed way with the svy: prefix may be another matter.
