In article <C6M4JE.7Bw at eis.calstate.edu>, ltorre at eis.calstate.edu (Lily Anne B.
De La Torre) says:
>>Can someone out there help me distinguish between the scientific terms:
> -Accuracy
> -Reliability
> -Validity
I can give you a quote about "accuracy required" from Wilson, E. Bright Jr.,
in _An Introduction to Scientific Research_, (1990 repub.), Dover, pp. 58-9.
"The accuracy of an experiment depends upon the degree with which
the important variables can be controlled and upon the accuracy of
the measuring apparatus. It is usually very expensive in time and
effort to increase these by a factor of 10. It is therefore seldom
economical to use an accuracy much greater than that necessary for
the aims of the investigation. Some qualification of this advice is,
however, necessary. If it really costs little more to get greater ac-
curacy, it is clearly desirable to do so. Furthermore, good foresight
can often see other applications of the results, perhaps a little in
the future, for which the greater accuracy is needed. Finally, some-
times measurements at higher accuracy bring to light new and unfore-
seen discrepancies of fundamental importance. An example of this is
the Lamb-Rutherford measurement of the hyperfine structure of the
hydrogen spectrum, which showed that the Dirac theory needed modifi-
cation. Generally speaking, however, it is unprofitable to seek high
accuracy with no obvious application in sight, because it so often
happens that when a need for data of this accuracy arises, earlier
measurements are found to have been taken under unsuitable conditions
for use in testing the new hypothesis. Naturally this question is one
where sound judgement is more important than ready-made rules."
I noticed that a subsequent post addresses the other two issues you raise; I
hope this helps.
Regards,
Alex Vrenios, kx9i
PhD Student, CSE Dept.
Arizona State University
azaxv at asuvm.inre.asu.edu
