+ Reply to Thread
Results 1 to 2 of 2

Thread: Modeling ordinal responses with nested data

  1. #1
    Points: 384, Level: 7
    Level completed: 68%, Points required for next Level: 16

    Thanked 0 Times in 0 Posts

    Modeling ordinal responses with nested data


    I am preparing a study that would involve modeling responses an ordinal scale. These responses identify the behaviors of high school students who go on to attend particular universities; I want to predict their responses as a function of both individual-level observations and university-level factors (e.g., institutional control [public/non-profit private/for-profit private], size, type [four-year vs. two-year], etc.). All of the effects to be tested are fixed, not random. What is the appropriate statistical technique to use here? I'm thinking an HGLM equivalent of ordinal regression but am unsure (I'm given to understand HGLM is a mixed-effects modeling technique). There are several variables to be studied at both levels of the hierarchy, and no interaction terms are included at either level.

    Additionally, if anyone is familiar with SAS, what would be the appropriate family of procedures with which to run the analysis (i.e., PROC MIXED, GLM, etc.).


  2. #2
    Omega Contributor
    Points: 38,253, Level: 100
    Level completed: 0%, Points required for next Level: 0
    hlsmith's Avatar
    Not Ames, IA
    Thanked 1,185 Times in 1,146 Posts

    Re: Modeling ordinal responses with nested data

    I have not ran these models for ordinal data, but I beleive you would be looking at logistic regression for ordinal data. Thus, the proc glimmix and nlmixed would be the procedures. These approaches may also be called cumulative logit models.

    I know, Multilevel Models: Applications Using SAS, by Wang, Wie, Fisher, briefly cover these models on pages139-146.
    Stop cowardice, ban guns!

+ Reply to Thread


Posting Permissions

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

Advertise on Talk Stats