Re: what good is "percentage error"?

[PHRP@FHSUVM.FHSU.EDU (Roger A. Pruitt)]

> On Tue, 13 May 1997, Rauber, Joel              Phys wrote:
>
> > Folks,
> >
> > > Mike, you are being too hard on Frank. Don't kill the messenger. Take your
> > > anger out on all of the chemistry departments around the world that grade
> > > this way.
> >
> > You know, I don't mind in the least that the medical lab techs when I was
> > born had been graded in this fashion in the quantitative analysis courses
> > they took.
>
> Really?  I would feel much more comfortable in the hands of someone
> trained to reduce errors in their measurement than in the hands of someone
> trained to make the measurement in a way that gives the result they
> expect.  However, I'm sure that following graduation when they find
> themself in a lab where the right answer is not known in advance that they
> soon learn to overcome this defficiency in their training.
>
> |++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++|
> | Doug Craigen                                                 |
> |                                                              |
> | If you think Physics is no laughing matter, think again .... |
> |    http://cyberspc.mb.ca/~dcc/phys/humor.html                |
> |++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++|
This exchange is an example of the way ideas drift on this list. The
discussion was about percentage error. "Take your anger out on all of the
chemistry departments around the world that grade this way" was originally
my comment, not Joel's, and i was referring to the fact that many of our
students come from a background where percentage error is considered very
important to the grade. Doug, I likewise would feel much more comfortable in
the hands of someone trained to reduce errors in their measurement, but that
was not the point of my comment.

It is hard to break students of wanting to put down a "percentage error" in
physics if the idea that this is important is reinforced elsewhere. John
asked about the importance of giving a "percentage error," and there have
been many excellent comments both pro and con. Some feel very strongly about
it. I just don't happen to like the expression--percentage error. A
"percentage difference" or a "percentage deviation" are, I believe, better
expressions, and I try to get students who insist on giving such a
percentage to use one of these expressions. I do, however, try to guide my
students into understanding the concept of errors in measurement, types,
accuracy, precision, resolution in a measurement, etc.

Many responders emphasized the importance of making a series of
measurements, averaging and finding the "standard" deviation. This is good.
Richard Grandy had such a bad experience that he ended up in philosophy.
However, the TA was wrong and Richard was doing good science. I wish I had
more students like Richard.

Analytical chemistry courses probably all grade on the basis of closeness of
the result to the "true" or "acceptable" value. I personally find no fault
with the fact that they do this, but it does make my job of teaching error
analysis harder. 

Roger
==============================================================
Roger A. Pruitt, PhD
Professor of Physics
Fort Hays State University
Hays, KS  67601
==============================================================