TERMINOLOGY FOR UNCERTAINTY

["Fritz A. Seiler" <faseiler@nmia.com>]

A comment by Fritz Seiler and Joe Alvarez
on the Use of the Term "Uncertainty."

We have just mailed a Letter to the Editor of RISK ANALYSIS,
entitled "On The Use Of The Term 'Uncertainty'."  We consider
the issue important enough to put a shortened version on both
lists
RISKANAL and RADSAFE.

Lately, many risk assessors been using the concept 'uncertainty'
as part of a new terminology involving the terms 'uncertainty' and

'variability', meant to distinguish two particular subsets of what

was traditionally described by the general term 'uncertainty'.
At the same time, international standards were clearly needed to
describe uncertainties of measurements and calculations. A multi-
year effort by the International Organization for Standardization
(ISO) has resulted in an ISO Guide Document, which has found
widespread international acceptance.  It is also used widely in US

industry; has been adopted by the National Institute of Standards
and Technology (NIST) for its own use (1); and was re-issued by
the American National Standards Institute (ANSI) as the American
National Standard, ANSI/NCSL Z540-2-1997, under the title
"American National Standard for Expressing Uncertainty  -  U.S.
Guide to the Expression of Uncertainty in Measurement."  An
overview can be accessed on the NIST website for Constants,
Units, and Uncertainties.  Its URL is:

          http://physics.nist.gov/cuu

These publications make two facts clear: One, these standards
use the term 'uncertainty' in its traditional meaning as a general

term for the total width of stochastic quantities; and Two, this
total width is composed of two complementary contributions:
random and systematic errors (1,2).

As these new standards preempt the term 'uncertainty' for use in
its traditional sense, its continued use in the specialized sense
adopted by some is now contrary to these standards.  However,
the change to this new terminology should not involve any
hardships, nor should it lead to any real difficulties (2).  In
the
interest of better communication, we think that, as risk
assessors,
we would be well advised to make this change, and that it would
be well worth the effort.

REFERENCES

1    B. N. Taylor and C. E. Kuyatt, Guidelines for Evaluating and
     Expressing the Uncertainty of NIST Measurement Results.
     Appendix D: Clarification and Additional Guidance  (NIST
     Technical Note, TN 1297, Gaithersburg, MD, 1994).

2.   F. A. Seiler and J. L. Alvarez, "Toward A New Risk
     Assessment Paradigm: Variability, Uncertainty, and Errors,"
     Technol. J. Franklin Inst. 332A, 221-235 (1995).