Re: Google (not Borland) knows all ;)

From: Bob Dawson (bdawson_at_idtdna.com)
Date: 12/17/04 

Date: Fri, 17 Dec 2004 09:20:58 -0600

"Dennis Landi" wrote
>
> It doesn't matter who is here at any given time don't you see that?
> Everyone is reading/writing to the the delphi NGs because of *delphi*.

At that level of generality, then. we're a self-select and the significance
of our opinion on matters Delphi become suspect.

Many are interested in Delphi--but only a few small number of us end up
here. So maybe there's something odd about our views, or maybe there's
something odd about us--either way it doesn't look good to a statistician.
Or maybe it's all just chance and we really do mirror that wider
world--that's possible. There's just no reasonable basis for asserting it to
be the case.

> I think you are too fixated on individuals Bob. And this is particularly
> what makes your stance so weak, given your self-proclaimed expertise on
> statistics.

The group is an abstraction--a group like this is strongly influenced by the
actions of individuals.

And for the record, I don't believe I claimed any expertise in statistics
passed enough layman's knowledge of the subject to know what's plausible and
what isn't.

> there are other samples (of which these forums are a great example) which
> can be just as a representative as other techniques.

There are multiple selection methods that can lay claim to levels of
confidence. The assertion that this forum conforms to any of them isn't
plausible. As Jake rightly distinquishes, that doesn't mean the group is
unrepresentative, merely that it is not known to be representative and
cannot beclaimed to be so.
bobD

Relevant Pages