Differences in "p" between Presence and RPresence output

questions concerning analysis/theory using program PRESENCE

Differences in "p" between Presence and RPresence output

Postby DMNH » Thu Jul 11, 2019 10:04 am

Hi all,

I am attempting to perform multi-season occupancy modeling in RPresence. I haven't been able to find any tutorials or walkthroughs for performing occupancy modeling with RPresence and have been figuring it out mainly through trial and error. It is very plausible that I have made a mistake somewhere in my code and haven't detected it. In order to "confirm" that I was performing the analysis correctly I used the same .pao and performed the same analysis in RPresence and in the Presence software. I first noticed the difference in my own data and assumed that it was an error within my dataset. I then took a step back and followed the guided instructions for the northern saw-whet owl example. I am making the assumption that the default option for Presence, psi(.),gamma(.),eps(.)p(.), would provide the same results as this code in RPresence:
Code: Select all
NSOPao<-readPao('nso.pao')
mod1<- occMod(model=list (psi~1, gamma~1,epsilon~1, p~1),data= NSOPao, type= "do.1")

When I compared the outputs of these models the AIC values were the same as well as the psi,gamma, and epsilon values.However, the p values were not the same. The Rpresence output help p at .49474 and the Presence output started at .49474 and had some missing values and then some random numbers throughout.

Can anyone please explain to me where I could have went wrong?

Thank you,
Devin
DMNH
 
Posts: 4
Joined: Fri Apr 19, 2019 2:13 pm

Re: Differences in "p" between Presence and RPresence output

Postby darryl » Thu Jul 11, 2019 4:31 pm

Hi Devin,
Your should get the same estimates as both RPresence and PRESENCE use the same underlying code (aside from version differences, or bugs...), it's probably just an output feature (or bug, depending on your perspective).

So you're saying its the PRESENCE output that is giving you missing and random values? You get missing p estimates when there was no corresponding survey (ie a '-' in the pao file), although you shouldn't get random values being output. What version are you using? If not the latest, try updating.

Cheers
Darryl
darryl
 
Posts: 466
Joined: Thu Jun 12, 2003 3:04 pm
Location: Dunedin, New Zealand

Re: Differences in "p" between Presence and RPresence output

Postby jhines » Thu Jul 11, 2019 4:36 pm

Hi Devin,

Darryl's right that Presence and RPresence use the same code to run models, but since I had to make the C++ code into a DLL for RPresence, some of the initial code is different. I found the cause of the "feature/bug" and will post an updated version of RPresence tomorrow (12July). The random estimates you're getting correspond to missing data and you can ignore them. The rest of the output is correct.

Jim
jhines
 
Posts: 469
Joined: Fri May 16, 2003 9:24 am
Location: Laurel, MD, USA


Return to analysis help

Who is online

Users browsing this forum: No registered users and 1 guest