www.phidot.org

I just added a section to the chapter on closed abundance estimation (Chapter 14), which covers deriving 95% CI for model averaged abundance estimates. A number of folks (myself included) have simply used the reported 95% CI generated using the MARK model average routines.

However, this is not robust - and can generate 95% CI that make no sense (i.e., where the lower-limit is < M(k+1), which is the number known to have been captured). See the -sidebar- beginning on p. 29 of the newly revised chapter (which I uploaded a few moments ago...).

http://www.phidot.org/software/mark/doc ... chap14.pdf

Section 14.9.1 of the closed captures chapter in the MARK book states the model averaged CI for abundance is incorrect and recommends using the 'by hand' approach shown. Is it possible to update MARK to give the correct CI by programing the 'by hand' approach into MARK? Or has this already been done? It seems inefficient to have each MARK user complete the 'by hand' calculations.

Under what conditions will the MARK CI for closed abundance be nearly the same as the 'by hand' approach? Any guidance is appreciated.

brp wrote:Section 14.9.1 of the closed captures chapter in the MARK book states the model averaged CI for abundance is incorrect and recommends using the 'by hand' approach shown. Is it possible to update MARK to give the correct CI by programing the 'by hand' approach into MARK?

.

No, not easily.

Or has this already been done? It seems inefficient to have each MARK user complete the 'by hand' calculations.

No - see above. Doing it by 'hand' is something of a misnomer. You can dump all you need to Excel (or equivalent), and do the needed calculations pretty easily there.

Under what conditions will the MARK CI for closed abundance be nearly the same as the 'by hand' approach? Any guidance is appreciated.

Assume the model-averaged CI is always wrong (its not always the case, but there is little reliable a priori guidance). Do it by hand. No shortcuts recommended.