calculating detection probability

158 views
Skip to first unread message

Neil Carter

unread,
Jun 5, 2012, 3:24:16 PM6/5/12
to unma...@googlegroups.com
Hello,
I've run several pcount models for different study areas.  In some cases the best models were constants only (i.e., lam[.]p[.]) and in some cases covariate models performed best.  I'd like to calculate the detection probability given presence for each of the models, but I don't think I'm doing it correctly.  I'm using the predict function for each model to determine detection probability but the probability is really small <0.008 for the constants only models, and it is only around 0.04 or so for the covariate models.  This seems far too low as c. 75% of sites detected an animal.  Any help in this regard would be really helpful.  Thanks.
--Neil 

Richard Chandler

unread,
Jun 6, 2012, 4:32:23 PM6/6/12
to unma...@googlegroups.com
Hi Neil,

It would be helpful if you could show some of your data (the replicated counts) and perhaps the model output.

Richard
 
_____________________________________
Richard Chandler, post-doc
USGS Patuxent Wildlife Research Center
301-497-5696



From: Neil Carter <nhcar...@gmail.com>
To: unma...@googlegroups.com
Date: 06/05/2012 03:24 PM
Subject: [unmarked] calculating detection probability
Sent by: unma...@googlegroups.com


Reply all
Reply to author
Forward
0 new messages