RE: Re: st: GLM - offset needed?

[Jennifer Devine <jdevine@mun.ca>]

I've tried both taking the logarithm of the area swept before including it 
as -offset()-, or including the untransformed area swept as -exposure()- 
instead of -offset()-.
Neither have corrected the problem. I also forget to add, I had a weighting 
term in my model - would that affect things?

glm catch yr strat [iw=total area], f(nb) link(log); total area is the 
total amount of area sampled in a year and is on the order of 200,000 sq. 
nautical miles.

Including either an offset (including taking the log of area swept) or 
exposure causes my slope to be positive, instead of negative - that can't 
be correct because the data show a clear decline over time. Also, the 
parameters of the models are identical, regardless if I use offset or exposure.

Is there another way to take into consideration the area swept by the gear 
given that I already use a weighting term?

Thank you,
Jennifer


*
*   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/