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 Output Interpretation

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

02-10-2015 07:11 AM

I'm learning Proc Genmod at the moment. Novice to modelling as well.

I need someone to guide me in interpreting the proc genmod output although I've some idea to interpret P values. What does the table 'Creteria for accessing goodness of fit' actually explains? From the documentation I understand it explains how our data fits the model, but I don't understand what it actually tells us.

What is intercept in 'Analysis Of Maximum Likelihood Parameter Estimates' output table?

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

02-10-2015 01:32 PM

The intercept is the predicted value when all covariate values are set to zero.

Steve Denham

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

02-11-2015 06:03 AM

Thanks!

How about the other question?

*'I need someone to guide me in interpreting the proc genmod output although I've some idea to interpret P values. What does the table 'Creteria for accessing goodness of fit' actually explains? From the documentation I understand it explains how our data fits the model, but I don't understand what it actually tells us.'? *

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

02-11-2015 07:46 AM

This would be a good time to pick up the seminal text in the field, MCCullagh and Nelder's *Generalized Linear Models, 2nd ed.*

Information criteria allow you to "rank" models, provided the data are unchanged.

I always look at the scaled Pearson's chi squared divided by the degrees of freedom as a guide as to whether the distribution chosen fits the data.

The point of this is that there is not a well-agreed upon measure of goodness of fit for generalized linear models, akin to Rsquared for ordinary least squares fits.

Steve Denham