How to add omodel and fitstat commands to Stata11?

[laxmsb]

Hi

I have run ordered logistic regression and to measure parallel regression assumption I need "omodel" and "fitstat" commands.
But when I searched in Stata by using findit command, there is no sufficient results to install from the net
does anyone know how to add these command?
Thanks

[Sergiy Radyakin]

Search results hardly depend on the version of Stata. Even in Stata 10 the search results contain links to the omodel and fitstat commands:
omodel from http://fmwww.bc.edu/RePEc/bocode/o
fitstat from http://fmwww.bc.edu/RePEc/bocode/f

Generally Google also brings some relevant results. How many results are sufficient?

Best, Sergiy

[laxmsb]

So, how can I retrieve these commands given in these sites?
There are enough source to get but it failed when I have tried to install them

[Richard Williams]

Don't use the version of fitstat from SSC though; it is very, very old. Instead do

findit spost13_ado

and install it. Despite the name, many/most commands will work with Stata 11, sometimes with reduced functionality.

Also, fitstat isn't the command you want anyway. You want -brant-, which is also part of the spost13 package. Or better yet, just do

findit oparallel

oparallel, by Maarten Buis, does most or all of the common tests of the parallel lines assumption, as well as some tests that aren't otherwise available in Stata.

[Sergiy Radyakin]

On the technical side, the necessary commands to install packages are shown in the viewer:



And all files can also be found in the RePEc pages, e.g.:
http://econpapers.repec.org/software...de/s320901.htm

On the conceptual side, listen to Richard's advice above.

Best, Sergiy

[laxmsb]

I am sorry but I do not understand

I have to use commands of omodel, prvalue, fitstat etc.. but I couldnt install them. I have a limited time to finish my assignment but I still did not sort my problem.
Richard, you mean I should to use oparallel instead of omodel ? but I couldnt install them either.

Please help me

[Richard Williams]

Use the findit command, e.g.

findit oparallel

Click on the description of the command and then click on install.

The prvalue command is outdated but if you need to use it, instead of installing spost9 do

findit spost9_ado

[Richard Williams]

If you are taking a class I am mildly surprised if you didn't receive any instructions on how to download and install programs. You may wish to work with another student or double-check your notes or handouts. Or, check with the instructor.

[laxmsb]

It did work, thank you so much.

Additionally I want to illustrade my predicted probabilities of my dependent variable by using plot.
Which commands should I use, respectively?

[Richard Williams]

If that is what your instructor wants you to do then I suspect you already have the necessary instructions or the resources to find them. Are you using Long and Freese's book? It has a lot of information about such things. See

http://www.stata.com/bookstore/regre...les/index.html

Otherwise googling phrases like "plot predicted probabilities stata" will yield many hits.

In general you should first try to find the answers yourself and then turn to the list if you get stuck. This is especially true for homework assignments.

[Scott Howell]

Has the bug with oparallel in Stata 15 been addressed? I am still having issues with it.

[Kazim Hashim]

Hi,
I am facing a similar problem in that I have run an ordered logistic regression but now cannot find the 'omodel' and 'fitstat' commands. I am currently using Stata/SE16.0.
Does anyone know where I can download these commands from?

Thanks

[Nick Cox]

For example

search omodel

points to a 1997 package and gives you a way to install it. How far it applies if you have Stata 16 I really can't say.

This is really the same advice as given earlier in this thread. findit still works.

[Kazim Hashim]

Hi Nick,

Thank you for your reply, I have downloaded the omodel package but when I try to run this code,

'omodel ologit QualGroup i.FathersQuals i.MothersQuals EngLang ImmigAgeOver16 Male i.Location i.AgeBands i.Ethnic'

It comes up with an error and states that factor-variables are not allowed. However my dependent variable Qualification Group is ordinal in nature, hence the reason for me conducting my Ordered logit Regression. What would be the correct way to code the omodel test in this instance?

thank you in advance

[Nick Cox]

As said omodel is from 1997 and nothing allows the time travel of it supporting syntax introduced in 2009 (help whatsnew10to11). Conversely, the thread title really doesn't apply to you because you have Stata 16.

More crucially, I am ignorant here and can't help further with your question. I don't even know what the omodel test is in this or in any other context. What test is it you want? If it's standard, official Stata is likely to support in conjunction with ologit.

I would start a new thread. Explain your real problem from the outset. Spell out the test(s) you want and ask how to get it or them in Stata 16.