[R-sig-Geo] Proper ME{spdep} usage

[Amy Henry]

Hi,

My data consists of two species counts derived from fixed position
traps over 6 months. It is very skewed. I also collected data on
environmental variables such as building characteristic for where the
traps were located.

I would like to make sure I am using the function ME{spdep}
properly.First I ran a negative binomial regression to identify theta
and then a glm with family=nb(theta). I created a correlogram of
residuals using Moran’s I . If spatial autocorrelation was present I
removed it with ME and then compared AIC to determine if it improved
model. I ran another correlog to see if the spatial autocorrelation
was removed. Most of the time spatial autocorrelation is still present
but seemed to be at a different distance class and the model did not
improve.


I have attached a few examples of my script for different
environmental variables and subsets of my data. I am sorry I am not
able to post on website. Thank you very much for your advise.

Best,
Amy
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Examples.R
Type: application/octet-stream
Size: 19652 bytes
Desc: not available
URL: <https://stat.ethz.ch/pipermail/r-sig-geo/attachments/20130801/e4f7a509/attachment.obj>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Data_All.csv
Type: text/csv
Size: 5325 bytes
Desc: not available
URL: <https://stat.ethz.ch/pipermail/r-sig-geo/attachments/20130801/e4f7a509/attachment.bin>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Data_InHDB.csv
Type: text/csv
Size: 2535 bytes
Desc: not available
URL: <https://stat.ethz.ch/pipermail/r-sig-geo/attachments/20130801/e4f7a509/attachment-0001.bin>