diazrenata
commented
4 years ago - bootstrapping doesn't seem all that much better than just going off the sample mean
- low abundance creates persistent problems depending on the seed
- we will def need to deal with this, because the neon smammal sads are unsurprisingly highly uneven with relatively many rare species
- rare species can drive overlap way down depending on sampling, and can really mess with the capacity to detect clustering.
- so this may send us back to weighting based on nb individuals represented
- but i also think this signals that there will be communities for which we just can't say if we're getting overlap more consistent with clustering (i.e. some pairs of species overlap a lot) or dispersion or randomness, and that the highly uneven nature of the sad compounds this a lot