-----------------------------------------
Maarten L. Buis
Institut fuer Soziologie
Universitaet Tuebingen
Wilhelmstrasse 36
72074 Tuebingen
Germany
http://home.fsw.vu.nl/m.buis/
-----------------------------------------
--- On Sun, 12/7/09, Song Nhac wrote:
> I employed Weibull model and Log logic model to test my
> data, but, I received two different sign in alpha and beta
> In Weibull: beta= -.3503182 and alpha=2.669465 (negative in
> beta) but in Log logic: beta= .8955857 and alpha= -.4768999
> (negative in alpha)- I expect beta is negative in this model
> but it did not.
> I use the same data set, in which the dependent variable
> is (survival) as in attached data set and independent
> variable is (lenght).
To avoid the spread of viruses people are not allowed to sent
attachments over the Statalist and your message would normally
have been automatically blocked by the server, as is discussed
in the Statalist FAQ:
http://www.stata.com/support/faqs/res/statalist.html#toask.
These two models model something different: the Weibull models
the log hazard of experiencing the event, while the log-logistic
model the log time till the event. If the hazard is high then the
time till the event should be small and vice versa, so you would
expect the coefficients of the two models to have opposite signs.
Hope this helps,
Maarten
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
