R Avry Wilson Wrote:
-------------------------------------------------------
> Jon Ellison Wrote:
> --------------------------------------------------
> -----
> > Martin Stower Wrote:
> >
> > > I consider reliable dates to be worth more.
> > You
> > > have an objection?
> > >
> > > M.
> >
> > You can consider whatever you like.
> > I consider all tabulated dates equally
> reliable.
> > All of equal worth.
>
> Then you do not comprehend basic statistical
> analysis. Given 100 samples, 99 say date X, 1 says
> date Y, and you have the audacity to think the 1
> date of Y defines the whole set. No, Jon, it is
> more likely the 1 sample is an anomalous result
> that can be removed. The importance here is that
> all dates were reported, not thrown out
> before publication. They included data no matter
> how odd it seemed, and that's honest science.
>
> But then, you don't even care to ask the question
> of why these dates are what they are, you just run
> with them. So how are you going to explain mortar
> made from local mud deposits with centuries old
> flora in them?


"More likely", you say? Is that your standard for throwing out data points?

It seems to be you who doesn't understand statistical analysis. No, you may not throw out data just because it "seems" to be an outlier. You must run a statistical analysis to determine whether it must be thrown out. It's not Jon who's not asking why the dates are what they are, it's the authors of the study who didn't ask what they are (or at least they didn't tell their audience). That's not honest science.