----- Original Message -----
From: "Dennis Breslin" <dbreslin@ctol.net>
To: "Jonathan Graham" <grahamj@virtue.cx>; "Bruce Markowitz"
<scosgt@worldnet.att.net>; <agraham@agraham.ca>; <thinkpad@stderr.org>
Sent: Saturday, March 13, 2004 12:52 PM
Subject: Re: [Thinkpad] Dell Vs IBM laptops
> I've generally been interested in the discussion but this response is
> getting tiresome. There's nothing irresponsible about the conclusions
that
> others have drawn, especially from those with fairly extensive experience
> with both IBM and the competition.
It depends I think there is something irresponsible about claiming "x is
generally true by experience y" if you can't really say experience "y" is
representative. I think there is something especially wrong with claiming,
as one person did that anecdotal evidence is the only way to evaluate this
system.
> Validity or reliability issues aren't really relevant
If you don't think it's relevant to know if the information you are using to
make a decision is valid or reliable. Feel free. I simply think otherwise.
> and in fact your exacting standard here are rather misplaced
> insofar as it is necessary to answer the question that a listmember asked.
The question that I'm concerning myself with is per the original post was:
"Can anyone give me any reasons why I shouldn't buy a Dell rather than an
IBM?"
I guess our difference of opinion is here is that I don't consider
'experience' and 'reason' to be synonomous in this context. A person may
have an 'experience' with a laptop but said experience may not constitute a
'reason' to buy or not buy. Ergo I consider validitity and representitivy
to be a useful question to raise.
> But a good analyst looks for patterns to develop.
However if you can't say that your sample is representative of the
population. How do you know if your pattern is the exception or the rule?
You have given NO answer to that question.
> Oh, and remember, data never speak for themselves and those who
> place a lot of trust in numbers are either hiding or being hoodwinked
> by their theory.
It's not that data doesn't speak for itself. Just that many people wish to
make it say much more than it actually says (case in point). In fact your
quirky little adage is kind of backwards. It's not about putting a lot of
trust in numbers but knowing which numbers to trust.
Anyway you and Bruce probably don't have to worry. I have better things to
do than teach remedial statistics. For today at least! ;-)
_______________________________________________
Thinkpad mailing list
Thinkpad@stderr.org
http://stderr.org/cgi-bin/mailman/listinfo/thinkpad
Received on Sat Mar 13 13:37:31 2004
This archive was generated by hypermail 2.1.8 : Fri May 26 2006 - 16:02:30 EDT