Berislav Horvatic wrote:since the result is not a set of estimates. In short, "a measure of deviance to
EACH n°/ha", which you asked for, does not exist
Well, sorry, Bero, but as far as I know (which is not much, I admit...), this is not true, or at least renders the calculated population size estimate weak.
To me, it's also a basic scientific principle - all estimates require some indication of confidence.
Berislav Horvatic wrote:(= is not given by the procedures).
Thus, not "=", imho.
Berislav Horvatic wrote:Using several methods gives several different estimates, which can then be compared among themselves,
to get the feeling how much they differ.
If each estimate is associated with a huge variance, it is irrelevant whether they produce similar population size estimations or not. In contrast, similar estimates with dissimilar variances CAN be indicative for preferring a method.
Berislav Horvatic wrote:But it makes no sense to calculate the average value of these
results and the corresponding SE
Of course not. The variance to a population size estimation is not really the same as the statistical variance formula.
http://en.wikipedia.org/wiki/Mark_and_recapture
Berislav Horvatic wrote:(Anyhow, that's NOT what you asked for.)
Euh... Indeed.
I rest my case.
