Two-species model with equal detection probabilities

[Orien Richmond]

I'm running the two-species model using the phi parameterization. My data set consists of 168 sites surveyed up to three times each for two secretive marsh birds. I set up the detection history data with 168 rows for the first species and 168 rows for the second (336 total rows). My first run estimated all 8 parameters (psiA,psiB,phi,pA,pB,pA',pB',pAB) and went smoothly. The model with the assumption that detection probability is independent (psiA,psiB,phi,pA,pB,pA',pB') also went fine. However, when I ran the model which assumes equal detection probabilities (psiA,psiB,phi,pA,pB,pAB) it should have 6 parameters, but in the results browser and in the model output it only shows 5; the pAB parameter was not computed, even though it appeared correctly in the design matrix. Also, when I tried to run the model which assumes that occupancy is independent (psiA,psiB,pA',pB',pA,pB,pAB), Presence crashed several times and the results file was incomplete. Any help much appreciated!

Cheers,
Orien

[darryl]

Orien,
I'll start with the obvious stuff.

1. How recently have you downloaded PRESENCE? There's were some issues with the two-species model a few months ok that have since been fixed.

2. Did you remember to unselect the 'detections are independent' option before running your second model?

Darryl

[Orien Richmond]

Hi Darryl,

I downloaded Presence yesterday, it's the current version.

Yes, I did unselect the 'detections are independent' option before running the model which assumes equal detection probabilities.

[darryl]

Orien
I just downloaded the latest and tried to fit models with another data set and struck the same problem. Seems like Jim may have accidentally introduced a bug at some point as it use to work. I'll let him know. Thanks for picking this one up.
Darryl