turn on suggestions

Auto-suggest helps you quickly narrow down your search results by suggesting possible matches as you type.

Showing results for

Find a Community

- Home
- /
- Analytics
- /
- Stat Procs
- /
- Proc Genmod & step by step approach for modeling

Topic Options

- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Highlight
- Email to a Friend
- Report Inappropriate Content

08-11-2015 01:13 PM

Hello, I am working on my dissertation. I have issues with, normality, linearity, heteroscedasticity, sparse data, and clustered data. I noted that a GEE model in proc genmod can account for many of these issues. The problem is that I have not had matrix algebra or calculus so many of books and journals are beyond my current competency. Does anyone know of a "For Dummies" resource for proc genmod and GEE?

Sincerely,

David

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Highlight
- Email to a Friend
- Report Inappropriate Content

08-12-2015 11:39 AM

Since you are working on your dissertation, I suggest you ask your advisor. He or she would know your skills/background and also standard resources in your field that would be accessible to you. What is your field of study?

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Highlight
- Email to a Friend
- Report Inappropriate Content

08-12-2015 12:10 PM

Thanks Rick for the reply.

I have spoken with my chair/advisor as well as other members of my committee without much success. My doctorate is in counseling education. I believe I am going in the correct direction but need a source (Book, Journal) which I have not discovered tha provides more scaffolding so I can gain a fundamental understadning of GEE and proc genmod for my defense.

dj

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Highlight
- Email to a Friend
- Report Inappropriate Content

08-12-2015 03:25 PM

1) Find an elementary and intuitive introduction to linear regression in your school's library or online. This should be easy to find and master. For example, a good undergraduate textbook should suffice.

2) Find an elementary and intuitive introduction to logistic regression. This also shouldn't be too hard to locate.

Logistic regression is the simplest type of generalized linear model, in which the logit of a binary response variable is a linear function of the regression coefficients. Again, you ought to be able to find a good undergraduate text or a survey article that describes logistic regression in an accessible manner.

3) Next look at Poisson regression, which models a response variable that represents a count (0, 1, 2, ....). This is another special kind of generalized linear regression.

4) The theory of generalized linear models is mostly abstractions that enable you to handle logistic models, Poisson models, and related models in a single framework. If you can start with the simple subcases, I think you'll have an easier time building up to your final objective.

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Highlight
- Email to a Friend
- Report Inappropriate Content

08-12-2015 03:29 PM

Thanks again! I will look into it. . .