Re: Whenever you see a post by Luis A. Afonso, read THIS first
- From: "Old Mac User" <chendrixstats@xxxxxxxx>
- Date: 12 Jul 2006 06:19:56 -0700
Reef Fish...
First, welcome home from your travels. I hope you had a good time.
You asked for comments:
Any comment from anyone is welcome, but I specifically ask, if
there's ANYONE who doesn't understand what I had explained to
Afonso (in the RF quotes above) to show WHY he was wrong in
his repeated assertion:
I've followed that horrific series of posts without comment. What
could I possibly
say that would penetrate that very thick wall? I considered writing
something to try to improve the situation, but decided doing that would
only make it worse. Your patience seems to be endless. Well, at least
more enduring than mine.
The fundamental problem here seems to be that LA doesn't understand the
concept of a sample drawn from a population. Therefore he doesn't grasp
the concept of the variance of a sample of data (sample variance) as
compared to the variance of the population.. Without that basic concept
he can't go anywhere with statistics. I have the impression that he is
attending a course in statistics, and is blessing us with the
revelations he learns (if "learns" is the right word...??) in the
classroom. If so, I'd be interested in seeing his test scores.
The concept of "variance of a sample" or "sample variance" should be
fully understood by anyone who posts on or reads this board. But let
me add this thought. I've taught several thousand practicing engineers,
scientists, physicians, and others and I assure you that the concept of
sample variance (or the standard deviation of a sample) is not
immediately obvious even to people who have graduate degrees in those
fields. Some years ago I came to the following practice. When I teach
basic statistics I use a plastic bag that contains about 250 chips each
bearing a random "normal" number. The numbers are printed on paper and
pasted on heavy stock paper, then cut into chips. I explain that this
bag "is a process". It has an average and it has a standard deviation.
We could learn the values of those if we had access to all of the
numbers in the bag. But we don't. So we take a sample of chips (sample
of data) from the bag and calculate the sample average and the sample
standard deviation. We do that. Then the question is "what if we take
another sample of chips... the same number of chips as before? Will
that sample have the same average and the same standard deviation as on
the first trial?" All will agree that the average and standard
deviation will be different. The next question is "why" will they be
different. Some will quickly give the right answer... because the
numbers in the sample will be different. I work on this concept until I
am satisfied that every student understands that the sample averages
are varying "around" the population average and that the sample
standard deviations are varying "around" the population standard
deviation. From there we go into the distribution of the sample
averages etc. People have called me as much as 18 years after seeing
this demonstration and tell me "I still remember the bag of chips".
(Later I use a second bag of chips... a different color... to deal with
the concept of detecting a difference of averges, the t-ratio, etc.)
Here's what drove me to use this demonstration. In some fields of
science and engineering the word "sample" implies a physical sample...
like a sample of a chemical in a bottle or a sample of soil. Saying
"the sample standard deviation" immediately causes some people to think
in terms of "the standard deviation of the physical sample in the
bottle". Problem: They won't say that's what they are thinking. Once
they've gotten on that track it's almost impossible to back up and fix
the damage.
If LA had seen a demonstration of this sort, he surely would realize
that sample variances vary "around" the population variance... hence
some are larger and some are smaller than the population variance. But
of course he would have to actually "see it" in the true sense of the
word.
One parting thought on this long and difficult episode.
"When you try to cultivate the minds of men, you often turn up clods."
OMU
Reef Fish wrote:
http://tinyurl.com/guq98
Summary:
1) he has no concept that population variance is a function of the
distribution and exists independent of sampling,
in which case he's thick as a brick; or
2) he's a troll. Given the quantity and quality of explanations, and
his penchant for misrepresentating or ignoring those explanations
combined with the ongoing abuse, I vote for #2.
That was quoting Paul Sanchez's opinion.
This was Reef Fish Bob's latest theory:
Given your TWO alternative theories, MY current theory is the THIRD:
3. Luis A. Afonso is more than one person. At least TWO, perhaps
three or more.
< detailed analysis in the tinyurl >
The TROLL must be a THIRD person (or the improbable machine
bot). What identified the current TROLL is exactly the reason:
Nobody could possibly
be that dense - we're way past neutronium here.
As Sherlock Holmes said to Watson, "Once you eliminate the
impossible, ... whatever remains, however improbable, must be
the truth. ..."
-- Reef Fish Bob.
.
- References:
- Whenever you see a post by Luis A. Afonso, read THIS first
- From: Reef Fish
- Whenever you see a post by Luis A. Afonso, read THIS first
- Prev by Date: Whenever you see a post by Luis A. Afonso, read THIS first
- Next by Date: Re: Whenever you see a post by Luis A. Afonso, read THIS first
- Previous by thread: Whenever you see a post by Luis A. Afonso, read THIS first
- Next by thread: Re: Whenever you see a post by Luis A. Afonso, read THIS first
- Index(es):
Relevant Pages
|
Loading
