On Fri, 5 Feb 2016 14:11:46 -0500, Denenberg, Ray wrote
> The technical matter: ISO wants the term "precision" replaced by "accuracy".
I have objected to this change and I would like to hear your views.
"Precision" and "Accuracy" are two distinct topics. Over the period of our
discussions here we have, I think, addressed both.
As pointed out "1950" is the expression of a date in year precision.
In measurement, we need recall, "precision" is a term that references actually
a number of distinct features of measurement, in particular:
The market, for example, is flooded with digital devices that read to a large
number of decimal places: tiny fractions of a second, tiny fractions of grams.
Cheap digital readouts to feign precision...
Readability and repeatability or reproducability expresses nothing about accuracy. Central here is the concept of measurement "bias".
"Accuracy is a qualitative term referring to whether there is agreement
between a measurement made on an object and its true (target or reference)
value. Bias is a quantitative term describing the difference between the
average of measurements made on the same object and its true value. In
particular, for a measurement laboratory, bias is the difference (generally
unknown) between a laboratory's average value (over time) for a test item and
the average that would be achieved by the reference laboratory if it undertook
the same measurements on the same test item."
Do we have an agreed upon reference to which we can track our
reports to even speak of accuracy? Accurate to what? When I say a balance
messured something to 10 grams and want to speak of the "accuracy" I need to
appeal to a standardized and trackable reference. In the case of date and
time reporting we can't really speak about "accuracy" but "reliability",
"belief" (even personal belief of the reporter).. at most "consensus".
Precision and bias of carbon14 vs tree ring. Radio carbon years vs seasonal
"Carbon dating is used to work out the age of organic material — in effect,
any living thing. The technique hinges on carbon-14, a radioactive isotope of
the element that, unlike other more stable forms of carbon, decays away at a
steady rate. Organisms capture a certain amount of carbon-14 from the
atmosphere when they are alive. By measuring the ratio of the radio isotope to
non-radioactive carbon, the amount of carbon-14 decay can be worked out,
thereby giving an age for the specimen in question.
But that assumes that the amount of carbon-14 in the atmosphere was constant —
any variation would speed up or slow down the clock. The clock was initially
calibrated by dating objects of known age such as Egyptian mummies and bread
from Pompeii; work that won Willard Libby the 1960 Nobel Prize in Chemistry.
But even he “realized that there probably would be variation”, says
Christopher Bronk Ramsey, a geochronologist at the University of Oxford, UK,
who led the latest work, published today in Science."...
But a lot of our dates are just reported or copied or .. Think of those "known
Here we need to speak more of reliability and certainty of the reported the
date rather than accuracy.
Date: Fri, 26 Nov 2010 09:18:30 +0100
Reply-To: Discussion of the Developing Date/Time Standards
<[log in to unmask]>
That's asking for too much. The source does not matter. All we have done is
to break down "uncertain" to two subjective and personal types:
1) Questionable. I have a date but I'm not sure--- or even that I suspect that
the date is wrong.
2) Unknown. I don't have a date with sufficient precision yet. 198u, for
example, means that I know it was in the 1980s but hope to, at some later
time (state of knowledge),know.
and a data collection type
3) I've not yet collected the data. 198| (to use MARC syntax) means that I've
not yet gotten around to collect the date (incomplete).
198| expresses that the date shall be updated to include its year at some
198u tells us that, at this time, we don't know the year.
(1985)? tell us that we think its 1985 but are not sure.
Pragmatically I would handle 198| and 198u as 1980 decade precision.
(1985)? says something else. Its not even saying with certainty that the
event took place in the 1980s.
We could be crazy and add grade of certainty and data-quality:
- a) Known to be correct (observed, documented etc.)
- b) Likely correct ( p> 50%)
- c) Possibly correct (Might be but not likely)
- d) Likely incorrect (The date is expected to be wrong p ~ 0)
- e) Unknown (certainty unknown).
(1985)? with the grade (a) is equivalent to 1985.
198u says that its known (a) that the date in the 1980s. The 'u' says we know
Going back to by example about the date of the Great Flood and the birthdate
Edward C. Zimmermann, NONMONOTONIC LAB