>You seem a trifle sensitive about models and modeling - statistics 
>are just tools.

Fully agree......but these tools that have to be applied correctly.

>  It makes no sense to say that something is wrong with the data.

I see data sets all the time that are wrong; mistakes, typos, wrong 
coding of categorical variables, blanks, improper variable names, 
etc, etc. These can make a GLM routine crash.

>Either the program works for its intened purpose or it doesn't.

Here is another reason for non-convergence: models with interactions 
applied on highly unbalanced data; not enough observations in a 
particular two- or three way interaction combination.

>If one of the statisticians who helped debug the program for SAS and 
>another professional statistician/programmer cannot get the program 
>to work with a data set I'd say that the functionality of of the 
>algorithm depends on the data set - it is a tool that sometimes 
>can't handle the data.

Improper application of a particular method, over-complicated models 
with too many parameters,  or "funny" things with the data comes to 
my mind first. I have even seen people applying a Poisson GLM on 
negative data, and wondering why it crashed.

Perhaps you are willing to upload your data to a website, so that we 
can all have a go at it? That would solve this discussion for once 
and for all. Else we will never know why, and how you managed to 
crash the GLM routine in SAS.

Alain








Dr. Alain F. Zuur
First author of:

Analysing Ecological Data (2007).  Zuur, AF, Ieno, EN and Smith, GM. 
Springer. 680 p.
URL: www.springer.com/0-387-45967-7

Analysing Ecological data using GLMM and GAMM in R. (2008). Zuur, AF, 
Ieno, EN, Walker, N and Smith, GM
Springer.

Other books: http://www.brodgar.com/books.htm

Statistical consultancy, courses, data analysis and software
Highland Statistics Ltd.
6 Laverock road
UK - AB41 6FN Newburgh
Tel: 0044 1358 788177
Email: [EMAIL PROTECTED]
URL: www.highstat.com
URL: www.brodgar.com

Reply via email to