Hi all,
Thank you very much for all your comprehensive reply concerning error term distribution.
But, you may ask me why I am so curious about this distribution?
I wanna know how can I model some latent variables (unobserved covariates) in a logistic model. If I know the distribution of the error terms then I can see their effects (latent variable effects) in the likelihood of the model very easily. Then my parameter estimates will be more accurate.
That was the story!
Very eager for your comments!
Bunch of thanks,
Amir
-------------------------------------------
Amir Kasaeian
PhD Student in Biostatistics
Tehran University of Medical Sciences (TUMS)
amir_kasaeian@yahoo.com akasaeian@razi.tums.ac.ir -------------------------------------------
Original Message:
Sent: 07-18-2011 13:03
From: Margot Tollefson
Subject: error termdistribution of logistic regression model
Amir,
The book, Applied Logistic Regression 2nd Edition, by Hosmer and Lemeshow, published by Wiley in 2000 has a section on logistic regression diagnostics, Section 5.3. You may find the section useful.
Margot
-------------------------------------------
Margot Tollefson
Owner
Vanward Statistical Consulting
-------------------------------------------