Origyptian Wrote:
-------------------------------------------------------
> Martin Stower Wrote:
> -------------------------------------------------------
> > Origyptian Wrote:
> > -------------------------------------------------------
> > > Martin Stower Wrote:
> > > -------------------------------------------------------
>
> First of all, lets put back what Femano has chosen
> to omit: the source of the link and his comment
> which I was answering:
>
> > > > > > R Avry Wilson Wrote:
> > > > > > -------------------------------------------------------
> > > > > > [en.m.wikipedia.org]
>
> > > > > Is there anything in that link that doesn't
> > > > > support exactly what I've been saying? Where in
> > > > > your link does it give you permission to omit any
> > > > > data sample merely because it "seems" to you that
> > > > > it should be discarded? The link seems to clearly
> > > > > indicate that outliers are determined by robust
> > > > > statistical tests, not by licking your thumb and
> > > > > raising it to the wind.
> >
> > > > At the link I find this: “There is no rigid
> > > > mathematical definition of what constitutes an
> > > > outlier; determining whether or not an observation
> > > > is an outlier is ultimately a subjective
> > > > exercise.”
> > > >
> > > > I find also this: “Naive interpretation of
> > > > statistics derived from data sets that include
> > > > outliers may be misleading.”
> > > >
> > > > Why aren’t you explaining this to your buddy
> > > > Jon? He’s the one excusing himself from testing
> > > > for outliers.
> > >
> > > Because you misunderstood what you read.
> >
> > Bull. You’re not a mindreader, Femano.
> > You’ve shown time and time again that your
> > comprehension of simple English prose is poor.
> > Don’t presume to lecture me on what I
> > understood.
> >
> > > Where does it say you may throw out a sample
> > > because it just doesn't look right to you? The use
> > > of "subjective" in the context of that Wiki entry
> > > clearly is refering to the freedom to choose among
> > > the many mathematical tests for outliers as stated
> > > in the rest of that same paragraph, such as:
> >
> > [Femano’s attempt to blind us with science SNIPPED]
> >
> > > It doesn't say you don't have to use any test
> > > other than your own wishful thinking.
> >
> > First of all, Femano, where did Avry say any such
> > thing as you’ve imputed to him? Quote it and
> > link to it. Let’s see if you understood what
> > you read.
> >
> > Second, my having reinstated the comment I was
> > answering, everyone can see what you wrote: “Is
> > there anything in that link that doesn't support
> > exactly what I've been saying?”
> >
> > Clearly there is: “There is no rigid
> > mathematical definition of what constitutes an
> > outlier; determining whether or not an observation
> > is an outlier is ultimately a subjective
> > exercise.”
> >
> > It does not say that the only subjectivity in the
> > case is that of choosing “freely” among
> > different tests (which would be subjective
> > enough).
> >
> > What it says is more radical than this. It says
> > that there is nothing definite to search for in
> > the first place. There is no (mathematical)
> > definition of an outlier which is common to all
> > tests. They are not all equivalent. They are not
> > all searching for the same thing. In choosing a
> > test, we choose what counts as an outlier for our
> > purposes, hence “ultimately a subjective
> > exercise”.
> >
> > Two things: (1) you are misrepresenting Avry and
> > (2) you have not understood what you claim I
> > misunderstood.
> >
> > M.
>
>
> 1. How did I misrepresent Avry?

A reminder of what you’ve ignored above: “First of all, Femano, where did Avry say any such thing as you’ve imputed to him? Quote it and link to it. Let’s see if you understood what you read.”

It’s not like it’s easy to miss.

I take it that you can not quote or link to a post in which Avry expressed the view you’ve imputed to him. This is how you misrepresented Avry (my first numbered point).

> 2. What was the point of you quoting the sentence
> with the word "subjective"? Nowhere does your link
> allow you to omit any data points just because you
> "subjectively" claim to "think" they "seem" to
> "likely" be wrong. What was your point in citing
> that quote?

Femano,

It is sufficient for me that you exhibit your incomprehension.

You have yet to show that anyone has advanced the view you persist in imputing.

A long list of tests for outliers is in itself a red flag, an indication that there is no fixed definition of the thing being tested for. You naïvely suppose that plugging figures into a formula makes the procedure scientific, never mind that choosing which formula to plug them into entails precisely the arbitrary and subjective choice (of what counts as an outlier) you deny.

Consider the class of detection methods mentioned first: the graphical ones. What’s going on here, if not visualising the data and then relying on an intuitive and subjective notion (hopefully intersubjective) of what counts as an outlier?

You have not understood what you claim I misunderstood (my second numbered point).

M.