[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]
Re: st: use of fracpoly with Proportional Odds model?
At 04:56 PM 7/5/2006, Rodrigo A. Alfaro wrote:
I am not sure that distribution matters. frac_chk classifies
routines by distribution, but it looks like fracpoly only
distinguishes between stcox and everything else. But I may not have
followed all the ins and outs of the coding, as various other
programs are getting called too.
Richard suggests to make a copy of frac_chk.ado. Use which
to find it. For example in this pc:
Make a copy (for example in your ado folder) typing
copy C:\Stata8\ado\base\f\frac_chk.ado C:\ado\frac_chk2.ado
Modify frac_chk2.ado (using notepad or any text-editor).
The problem is how!! Specifically what distributions will be
asociated with ologit? Experts should know the answer.
Here is what I tried. I wish, though, that there were problems with
known solutions so i could tell if this was working right.
* Make copies of the most current versions of fracpoly.ado,
frac_chk.ado, and frac_154.ado and place them in your personal
folder. Rename them, adding a 2 or a B to the names or
whatever. Edit each of your copied programs to use these new names.
* On your copy of frac_chk.ado, on the tokenize command, add ologit
to the list of blessed programs.
That is all I did. When I do this, I get answers. But if
distribution does matter, then more coding, possibly much more, could
Anyway, I think this is a start, but there may well be more to it! I
continue to wonder why ologit isn't there in the first place, e.g. is
this something that you shouldn't even be trying to do or is the
omission of ologit simply an oversight?
Richard Williams, Notre Dame Dept of Sociology
OFFICE: (574)631-6668, (574)631-6463
WWW (personal): http://www.nd.edu/~rwilliam
WWW (department): http://www.nd.edu/~soc
* For searches and help try: