Re: what good is "percentage error"?

[gastineau@badgerden.com (John E. Gastineau)]

On 13 May 97 at 11:10, Richard W. Tarara wrote:

> While we've beaten the 'error' part to death (without, I think, a
> clear winner), let me go back to the percentage part.  
>
> An example of why getting students to think in terms of percentage
> differences (uncertainties, errors) is important can be illustrated
> using a simple experiment.  In the first lab of the year, I have
> students determining densities and hence the material of several
> objects.  One of those objects is a long (50-60 cm) copper wire with
> a diameter in the 1-2

snippage...

I see no problem at all with discussing the relative uncertainty of a 
measurement--that's quite an interesting thing. My beef has been with 
the measurement minus accepted over accepted kind of calculation. 
Your copper wire experiment is quite instructional and does not 
include the problems we have been beating on for the last day. 

It's good that you raise this, however, as it is important to draw 
the distinction between what I call the relative uncertainty, or 
sigma(x)/x, and the descrepancy calculation which refers to an 
outside authority without regard to the precision of the measurment. 
(By sigma(x) I mean an estimate of the uncertainty of a measurement, 
which might be the standard deviation (but it might not be, as 
well.))

JEG

==================
John E. Gastineau                (304) 296-1966
900 B Ridgeway Ave.              gastineau@badgerden.com
Morgantown, WV 26505
www.badgerden.com/~gastineau