Hello,
For context, I'm running a multi-scale occupancy model. My models perform as expected, and model outputs are consistent with those from other programs (e.g., Presence). Using the parameter estimates from my top model and the delta method, I have my estimates of standard error around cumulative probability of detection. I now need to estimate confidence intervals, which must be first calculated on the logit scale.
Here's my problem: At first glance, it seemed that RMark had a few options for giving me estimates and the VCV matrix on the logit scale (i.e., functions described in the package like 'compute.link' and 'compute.links.from.reals'). Unfortunately, these lead to the same error suggesting that the functions are not available (e.g., ' Error in compute.link(model = top.mod, design = dm, vcv = TRUE) : could not find function "compute.link" '). The 'real'-related equivalents of these functions (e.g., 'compute.real', 'get.real') work as expected.
Does anyone have insight on the availability of these functions? I do have the most recent versions of both MARK and RMARK (both of which I have uninstalled and reinstalled, just in case). Should I side-step the delta method altogether and instead use bootstrapping to generate my confidence intervals?
Thank you for your time.