Importance of recaptuure rate

[megpetrie]

Hello,

I'm getting a lot of doubt as to the reliability of my reported results (from people outside the MARK community) because I have some recapture probabilities that they consider low (e.g. in this specific case they are around 50% according to my RMark results). It is my understanding that the value of the recapture rate doesn't matter as much since it is incorporated into the model and accounted for whether it is 50% OR 90%. Is this assumption accurate? if the data pass the GOF tests and aren't producing results near the boundaries (0 or 1) it seems that as long as one considers the SE's the results are reliable. I do have somewhat small sample sizes for calculating these recapture rates but there are recaptures at the subsequent occasion (maybe on the order of 15-20).
If anyone has any thoughts on this I'd really appreciate it.

Thanks,
Megan

[bacollier]

Hello,

I'm getting a lot of doubt as to the reliability of my reported results (from people outside the MARK community) because I have some recapture probabilities that they consider low (e.g. in this specific case they are around 50% according to my RMark results). It is my understanding that the value of the recapture rate doesn't matter as much since it is incorporated into the model and accounted for whether it is 50% OR 90%. Is this assumption accurate? if the data pass the GOF tests and aren't producing results near the boundaries (0 or 1) it seems that as long as one considers the SE's the results are reliable. I do have somewhat small sample sizes for calculating these recapture rates but there are recaptures at the subsequent occasion (maybe on the order of 15-20).
If anyone has any thoughts on this I'd really appreciate it.

Thanks,
Megan

Megan,
I have had this discussion with folks before as well, hit same wall. Your right, given what you said above. If recapture rate is correctly (unbiased, etc.) estimated, then your resultant population estimates (or parameter estimates, not sure what your working on) should be accurate. Throw Williams/Nichols/Conroy on their desk (but not to hard, its pretty heavy) and tell them to get educated

bret