Results 1 to 6 of 6

Thread: Does the IMSL6.0 support the Ordered Probit Model?

  1. #1
    Junior Member
    Join Date
    May 2009
    Posts
    3

    Does the IMSL6.0 support the Ordered Probit Model?

    I read the User Guide but I can't find the ordered probit.

    If IMSL doesn't support ordered probit, could you tell me the other way to estimate ordered probit using c++?

  2. #2
    Well, two things here...

    First, the categorical_glm routine in IMSL C (which has the same functionality as CTGLM in Fortran) does not support multinomial responses, so only if your ordered response had two possible outcomes would the routine be of help. Assuming you get past this issue, categorical_glm does not fit the models any differently if the response is ordinal as well as categorical. That being said, my understanding is that you can model ordinal responses as just categorical responses and the coefficient estimates will not be biased, you'll just lose efficiency. So, if you have a large enough dataset you might be fine just fitting a nonordered probit model. For example, see pages 6 and 7 in

    Logit and Probit: Ordered and Multinomial Models
    Vani K. Borooah
    Sage University Paper Series: Quantitative Applications in the Social Sciences

  3. #3
    Junior Member
    Join Date
    May 2009
    Posts
    3

    Plz, Give me some help here.

    I tried to make a ordered probit program.
    But it didn't work.

    I've tried and modified a lot but I can't find what's wrong.

    I really appreciate if you can tell me what's wrong.

    Here's the benchmark answer and I attatched code and data.

    Thank you.

    Benchmark answer
    x[0] = -0.004881
    x[1] = 0.074675
    x[2] = -0.020546
    x[3] = -3.269916
    x[4] = -1.399122

  4. #4
    Thanks for posting the code and data. After briefly looking it over its not clear what kind of guidance you are looking for. For example, are you confident that the likelihood function you have coded is correctly formulated? Have you used this approach elsewhere and are just having trouble implementing it with IMSL C?

  5. #5
    Junior Member
    Join Date
    May 2009
    Posts
    3

    Model Description

    I quite confident with the likelihood function,
    cuz it from the Eviews help file.

    But I'm not sure that I implemented it correctly.

  6. #6
    Thanks, as that gives me and other readers a better place to start. I am hoping that someone out there is more experienced with ordered probit/logit models and will offer advice, because it will take me a while to implementing this myself and testings things...

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •