c-hat and robust design
Posted:
Thu Feb 16, 2012 4:40 pmby Lyndon
I've fitted a closed RD model (28 primary samples, 132 samples). I'm not aware of how overdispersion may be assessed for this model. Can anybody recommend an approach? One possibility that may be reasonable may be to use a c-hat from a CJS model based on data collapsed to primary samples. Does anyone have a view on that?
Re: c-hat and robust design
Posted:
Thu Feb 16, 2012 7:43 pmby cooch
Lyndon wrote:I've fitted a closed RD model (28 primary samples, 132 samples). I'm not aware of how overdispersion may be assessed for this model. Can anybody recommend an approach? One possibility that may be reasonable may be to use a c-hat from a CJS model based on data collapsed to primary samples. Does anyone have a view on that?
Quick search of forum shows multiple answers to this question. Try the following:
viewtopic.php?f=1&t=189&p=4576&hilit=goodness+of+fit+robust#p4578
Re: c-hat and robust design
Posted:
Mon Feb 20, 2012 9:56 pmby Lyndon
Thanks Evan. I hadn't found these posts in my search.
Re: c-hat and robust design
Posted:
Tue Feb 21, 2012 8:13 amby cooch
Lyndon wrote:Thanks Evan. I hadn't found these posts in my search.
No worries. If I could create a search engine that always yielded exactly what you were looking for, I'd put Google out of business. I'll let you know if that happens.
