Re: Significant figures - a Modest Proposal

[Robert A Cohen <bbq@ESU.EDU>]

On Wed, 1 Sep 1999, Ludwik Kowalski wrote:

> Robert A Cohen wrote:
>
> > Before this year, I used to separate precision and accuracy.  I used
> > precision to describe the "random" error and accuracy to describe the
> > "systematic" error.  I realized this was simplifying things, but I felt it
> > was sufficient for the students.  Unfortunately, I can't succinctly
> > describe the problem I had with that approach.
>
> The precision can be improved, as much as we wish (in principle)
> by increasing the number of measurements and using the average.
> But the accuracy can not be improved in that way; it can only be
> improved by eliminating biasses.

Thanks for your response.  Is this your opinion or is there a reference
you can refer me to?  The "target" example of low precision and high
accuracy seems to imply that using one imprecise "bullet" would lead to a
result that differs from the "actual" and thus be inaccurate.  Averaging
many imprecise "bullets"  would lead to a result that agrees with the
"actual" and thus would be accurate (assuming no biases).  I could be
wrong...

P.S. The header I received on your message was
> Date: Wed, 01 Sep 1999 21:10:22 -0400
but I received it Tuesday morning around noon.  Is this a mistake on my
end or what?

----------------------------------------------------------
| Robert Cohen               Department of Physics       |
|                            East Stroudsburg University |
| bbq@esu.edu                East Stroudsburg, PA  18301 |
| http://www.esu.edu/~bbq/   (570) 422-3428              |
----------------------------------------------------------