Discussion: View Thread

Robert, Robert, Robert,
I could almost agree with your positions on what you claim I said.
But whether I agree or not is irrelevant because you are attacking stuff
that I never said in the first place --- nor the second place -- nor the
third place. Is there is a full moon where you are?

You declared that my assumption of a normal distribution was nonsense but
you have yet to provide any justification as to why the distribution of
appraisal results across a population of 15 or more employees could not
possibly be a normal distribution. Please do so.

If it helps, presume that the results choices are: Far Exceeds, Exceeds,
Meets, Nearly Meets and Unacceptable.

And please be intellectually honest enough to recall that you did not say
"may be nonsense" but stated "nonsense" period.

Secondly, please understand that I said nothing about sampling anything;
random, highly selective or otherwise. Likewise I said nothing about
selection or hiring. The topic was performance appraisals (of incumbents).

Meanwhile, reflect on the following:
[RB] "You told me that we don't need to test assumptions or something to
that effect in one of your private remarks. [...]"
--- and consider whether "something to that effect" justifies even posting
your message. Wouldn't it be better to take issue with what I said than with
what you want to talk about?

I actually said, "Wherever did you get the idea the assumption needs to be
justified?" Please notice that the sentence is interrogative. I did not
assert
that the assumption does not need to be justified I only sought more
insight into why you say the (expletive deleted) things you do.

[RB]> I'm sorry, but you suggested a method that has no justification. It's
> that simple.

Interesting opinion. But, pending your elegant proof, still opinion.

On 7 May 00, at 14:54, Jack Ring wrote:

> Robert, Robert, Robert,
> I could almost agree with your positions on what you claim I said.
> But whether I agree or not is irrelevant because you are attacking stuff
> that I never said in the first place --- nor the second place -- nor the
> third place. Is there is a full moon where you are?

This is what you said:

"First, presume that any organizational component with 15
or more employees should exhibit a performance profile
approximating a normal distribution. Any group skewed
higher simply is not being sufficiently challenged and any
skewed lower are not getting the competency development
(quality of management) they deserve."

I commented that the assumption was faulty, hence the
method was faulty.

>
> You declared that my assumption of a normal distribution was nonsense but
> you have yet to provide any justification as to why the distribution of
> appraisal results across a population of 15 or more employees could not
> possibly be a normal distribution. Please do so.

I didn't say it couldn't be. I said you can't count on it, hence any
method assuming it and not testing for it (assuming it) is bound to
be faulty when it isn't (and often it isn't in a selected group).

To that you replied in your own ever-polite style:

From: "Jack Ring" <jring@amug.org>
Subject: Re: [MG-ED-DV] Appraisals and
Compensation
Date sent: Fri, 28 Apr 2000 14:40:36 -0700

"Wherever did you get the idea the assumption needs to be
justified?

Get over the bad habit of believing whatever you hear just
because it sounds like your voice saying it."

>
> Interesting opinion. But, pending your elegant proof, still opinion.

Jack, I've learned in the past that nothing will prove anything to you
that you don't currently believe. It's hard to NOT notice your
avoidance of dealing with the many other erudite (and polite) people
who commented on the thread you started.



Visit the work911.com supersite at http://www.work911.com
for work related articles, or to find almost anything including
book reviews and suggestions, discussion lists and more.

Dear Rob and Jack

This is a terrific debate, entertaining, illuminating, instructive
and just the sort thing to warm us up in the morning over here in
foggy England.

The reason that I find the exchanges so useful is that it helps me to
comprehend the way that people, very intelligent people, can and do
conceptualise statistical issues.

Most engineers, chemists, managers, social scientists and the like
that I meet do have some difficulty with these concepts.

The reason that the Normal dist' is so useful is that arithmetic
averages of measurements from any distribution that is
observed, will themselves be approximately Normally dist' provided
that the number of items from which the average is calculated is
big enough. So, for example, a persons average 'score' of some
measurement over a period of time, say X, (averaged over several, say
n, measurements) will behave as a 'Normal' dist provided n is big
enough. We could plot a bargraph of many Xs and the bell shape will
appear, provided no change has occurred over time.

So all distributions 'gravitate' or 'evolve' to 'Normal' in that
sense.

We could also plot the individual Xs on a time plot, and if the Xs
go up, it may indicate improvement in X over time. We can tell if
this improvement is statistically significant. We can take into
account assessor error, we can take into account the reliability of
the scoring system. Similarly we can systematically monitor the
assessors performance, compare it to others, itself, benchmark it.

In fact if you can measure something we can find a statistically
valid way to monitor it over time. We don't actually have to use the
Normal dist' for this.

The fundamental issue seems to be that of measuring differences,
either between some average or some target, and the individual or
group results over time. This in the face of measurement error
(assessor variance). The issues of the measurement score itself ,
and it's frequency, are also critical.

Why don't you get together to try out a few investigations?

The end product could be a way to prove incorrect use of the bell
curve when it happens, and provide ways of finding bias
in assessments and procedures, valid ways of spotting 'good' and
'bad' performance, 'good' and 'bad' schemes etc.

Bearing in mind your opposite viewpoints, this would undoubtedly be
rigorous and could be very useful for the whole field.

Regards
Dave Stewardson
ISRU {Industrial Statistics Research Unit}
MMME {Department of Mechanical, Materials and Manufacturing Engineering}
Stephenson Building
University of Newcastle upon Tyne
Tyne & Wear
England
GB - NE1 7RU
TEL 00 44 191 222 6244
FAX 00 44 191 222 7365

Robert,
Haven't gotten to the polite ones yet, many of whom presented interesting
opportunity for dialog.
Yes, I have been working the wrong opportunity.
Thought the howls might stem from some intellectual need but now discern
that you simply must be sitting on a cocklebur.
Good luck.
[...]>