st: Akaike's Information Criterion

[Fred Wolfe <fwolfe@arthritis-research.org>]

Does anyone have a rule of thumb regarding how much difference in Akaike's 
Information Criterion is considered to be important. I am familiar with the 
rough rule that is available with BIC. For example, in comparing nested 
models 3 and 4, is model 4 "better" or "slightly better," of doesn't it 
make any sense to characterize it this way. I am trying to write a brief 
description in the methods section of a "medical" paper to aid readers in 
understanding the AIC and differences between models.

Model 1 5288.47***
Model 2 5253.47***
Model 3 5254.87***
Model 4 5241.32***

Thanks,

Fred


Fred Wolfe
National Data Bank for Rheumatic Diseases
Wichita, Kansas
Tel +1 316 263 2125
fwolfe@arthritis-research.org


*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/