Dear Statalisters,
I am using the -gam- procedure (Royston and Ambler) to look at trend data.
My response variable is zero inflated count data. The data are
longitudinal with multiple observations on a site over time. The data
are:
tsset site year
In its simplest formation the model statement is:
gam count year site, family (poisson) link(log) df(10)
Question 1.
I would like to bootrap confidence intervals from the gam procedure. Does
anyone know how to do this, given the longitudinal nature of the data?
Question 2.
Also, -gamplot- simply plots the smooth function produced by the model. I
would like to plot the 'smoothed' counts on the original scale. Does
anyone have suggestions as to how I might do that?
Thank you very much for your time.
Brad
*
* 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/
