Confidence intervals

[Poochican]

Hi,
I am very new to github and not sure where/how to put these types of comments. This is not really an issue but would make the package more usable. In order to actually test difference in resource use vs availability you need to include confidence intervals (95%). These are readily available in the adehabitatHS package (see Manly 2002 for more details), however I prefer the various selection options you provide in this package!

Thanks!

[DesiQuintans]

Hi, I can investigate this further because a meaningful 95% CI would be great to have, but there’s some limitations with the electivity score as a number, if that makes sense. From my thesis:

Like all electivity indices, the numeric score itself cannot be directly compared
between species or sites, but the Relativised Electivity Index does allow for meaningful
comparison of rank order between sites that have different food availability (Lechowicz,
1982).

I ended up using Kendall’s coefficient of concordance (W) to see if the rank order of an insect’s preferred foods was consistent between different sites, even if the availability of those foods differed.