Trial vs io configuration

[Lyndon Brooks]

Hi,
I've been trying to fit a point idependence model to some mrds data. I'm having a hard time getting a detection function to fit. The dsmodel fits well but the mrmodel doesn't.
I supect that there may have been some sort of failure of the independence protocol leading to something like behavioual response in the mr data.
This may well be why a trail configuration is more robust than a full independence configuration.
My question is this: is trial less powerful than full independence configuration?
It seems we'd lose some mr data but perhaps the dsmodel would make up for that?
Thank you.

[Lyndon Brooks]

Hello,

To answer my own question: On our data, the trial configuration was a better fit (the mrmodel component of GOF chisq was p>0.05 at least) but the SE of p-hat was slightly larger. Perhaps this is a compromise we must make.

Lyndon.

--
You received this message because you are subscribed to the Google Groups "distance-sampling" group.
To unsubscribe from this group and stop receiving emails from it, send an email to distance-sampl...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/distance-sampling/9ec13e6a-fb62-4197-a4ec-83015156c693n%40googlegroups.com.