Differences in estimates between CAPTURE and RCapture

[mspinola]

Dear list members,

I obtained quite different estimates for the Mth Chao from CAPTURE (75, 95% CI: 61 - 108) and from RCapture (61, 95% CI: 53-84).
Any idea why?
Thank you very much in advance.
Best,

Manuel Spínola

[jlaake]

For those that do not know, RCapture is not RMark.

See http://ideas.repec.org/a/jss/jstsof/19i05.html

RCapture is an R package that fits log-linear models to capture-recapture data. I suggest that you contact the author/maintainer of the RCapture package with this question if no one offers any suggestions here.

--jeff

[cooch]

Dear list members,

I obtained quite different estimates for the Mth Chao from CAPTURE (75, 95% CI: 61 - 108) and from RCapture (61, 95% CI: 53-84).
Any idea why?
Thank you very much in advance.
Best,

Manuel Spínola

In addition to Jeff's comment, you are strongly advised not to use CAPTURE - it has been entirely superseded by the closed estimators in MARK. See Paul Lukacs' chapter in 'the book' (chapter 14).

[murray.efford]

On a tangent that may or may not be relevant to mspinola: CAPTURE grappled (largely unsuccessfully) with the problem of density estimation, which, as far as I know, is entirely missing from MARK. For a modern approach to density estimation you could try DENSITY (www.otago.ac.nz/density). Note also (1) there are (at least) two Mth Chao estimators, which may conceivably explain the discrepancy(DENSITY provides both of them) and (2) assuming your data are spatial (e.g. from a trapping grid), a large part of the individual heterogeneity 'h' driving your need for Mth may be due to location of home ranges: this is modeled directly when you use spatially explicit capture-recapture (e.g., DENSITY).

Murray