ianfiske
commented
14 years ago I'm not really sure off-hand of the correct way to do this since you don't observe a site-level occurrence. A residual is observed minus fitted, so what would you use for observed? Maybe the empirical Bayes predictor? You might also be fine just aggregating your detection-level residuals within each site -- maybe the mean?
Would anyone have any recommendations for extracting residuals for an occ object? Using residuals (x) produces residuals for every detection event, but I am interested in extracting residuals for only the occupancy process. The entire model fits several covariates for detection and occupancy, but I would like to plot the residuals for the model having a single residual for each site. Is this possible? Am I thinking about this incorrectly? Thanks!