Re: Question regarding stable systems and correlations

["Carlos H. Méndez" <carmende@ufm.edu.gt>]

Hello Lloyd, thank you very much for your detailed answer, you can't imagine
how many insights (i believe this is the word I'm looking for!) your
comments have given me. I think Steven Prevette calls them Aha moments!.
jeejeje. The good thing is that I have access to Shewharts book, so I will
read it again, since for some reason I missed the point the first time!!!

In your mail you wrote: ------"In this example, if we were interested in the
effect of the variable "# of classes students is taking", we would stratify
or subgroup by that variable before creating the chart. If we did a good job
of subgrouping, and there was an important correlation between exam score
and # of classes, then the control would indicate instability".----------

David Kerridge, in a previous mail,  comments: -----------"I believe that
Walter A
Shewhart, when investigating a process, would draw many charts, looking not
just at the final outcomes, but at a great many other measurable variables.
Once he had established that there was no correlation, he would stop doing
so many measurements.--------------

>From the above, another question pops  into my mind:  If I don't have enough
data to do the sub-grouping, that is make a chart for each level of number
of classes a student is taking (my class is not that big, about 30
students), is it too risky, in
terms of both of Shewhart mistakes, to conclude stabilily with all the
scores blended into only one chart?

Some additional information might be usefull. The chart where all the
students are blended into one is stable, and I have tryed many correlations
with other variables and no correlation seems to
emerge with anything. This are examples of correlations I have tryed: how
much he/she likes the theme of the class, interest in other classes,
tendency to process information with the left or right hemisphere, number of
classes he/she is taking, level of previous education, university where the
student is comming from, tendency to stress and fear, etc.

The fact that I can't find correlations is why I asked the question of this
thread!!! The funny thing is that I used to have correlations about 3 to 4
years ago, but in the past years the correlations seem to have faded away.
Of course I have been making efforts to stabilize and improve the system,
but I wonder if I'm creating and illusion of stability do to bad personal
statistics, jejejeje.

Thank your very much for taking the time to help. Cheers,

Carlos Méndez
Invenia, S.A.
invenia@guate.net
PBX - (502) 336-9669
_
"Being healthy is merely the slowest possible rate at which one
can die."
_
"Mantenerse sano es la forma mas lenta de morirse."