I have data that is autocorrelated because it is acoustic data from transects. I'm interested in
running a glm:
glm biomass year season tod
where season is 1 or 2 (Jan or June) and tod is time of day (1=day, 2=night, 3=dusk). At the moment,
everything is significant but that may be becuase of the autocorrelation. To deal with the
autocorrelation, I want to randomly resample the data and then run the glm to test for year,
seaoson and time of day effects. If I have season or time of day effects, I then need to break the
data down further. I'm getting nowhere using the stata maunal or help files for bootstrap and
bsample.
I run this:
bs sum_abc, strata(year season tod) saving(mybs2): glm sum_abc year season tod, f(gaussian)
link(id)
and the result is not what I want. Do I need to bsample after breaking it into smaller subsets by
year, season and time of day and then append the files? If I try to bsample the data with strata,
cluster or both, I can't get a large enough sample size while retaining the 3 X.
Thank you.
Jennifer
><>><>><>><>><>><>><>><>><>><>><>><>><>><>><>><>><>><>
Jennifer Devine
Doctoral Candidate
Memorial University
4 Clark Place
St. John's NL
Canada
A1C 5S7
Tel: 709-737-8833
Fax: 709-737-3121
[email protected]
><>><>><>><>><>><>><>><>><>><>><>><>><>><>><>><>><>><>
*
* 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/
