Traceability of a Measurement Result
To educate the Court about the meaning of "traceability" in measurement science.
To obtain an admission that there is a relationship between the calibration certificate (Certificate of Calibration) from the factory and the reference standards used during the factory calibration.
To obtain an admission that there is a relationship between the calibration certificate from the factory and the scientific reliability of the measurement result at time of use.
To introduce the NIST Certificate for the reference standards used at the time of calibration at the factory.
Sample cross-examination on "traceability" - How can every test stand on its own if every measurement result is a comparison to the reference standards used at the time of calibration at the factory? Surely the calibration certificate from the factory is relevant to reliability of the measurement result:
Q. All right. Let's talk about the whole
question of traceability of a measurement result.
Q. You understand what traceability of a
measurement result is? You understand what I'm talking
Q. All right. Now, as I understand it from
reading the Intoxilyzer 8000C training aid, that measurements
in Canada -- that was Exhibit No. 20. We had an excerpt from
the training aid. And in the training aid...
A. December 2013.
Q. ...2013. Page 16 of 238. I will just show
you my own copy to save you some time.
A. Hold on. I have a whole bunch. Let me see if
I can find it.
Q. This is page 16 of 238 in Exhibit 20.
A. Okay. I don't know why there's two copies,
Q. Page 16 of 238.
Q. Third paragraph down:
The units for all the measurements needed in
science technology in every day life are derived
from SI base units, derived units and from units
outside the SI.
A. We covered that the last day.
Q. And above that:
"In 1960, the General Conference on Weights and
Measures adopted the International System of
With the international abbreviation SI?
A. Yes. Right.
Q. Now, how is it that a measurement result on an
Intoxilyzer 8000C can be traceable to the SI units? Why does
A. The instrument is traceable to the standards
that were used to check its calibration.
Q. Really? So you're suggesting that
traceability goes through the control checks by the police
officers, as opposed to the calibration of the instrument by
A. No, that's not what I said. It's traceable to
the standards that were used at the point of calibration.
They are the standards obtained from Guth, or another
supplier, who certifies those standards, and the result is
traceable back to those standards.
Q. So on one of the previous days I showed you a
certificate of calibration for this particular Intoxilyzer
8000C that said that the solutions that were used were from a
standard 1820, and you looked at that document, and I think
we may have a copy of it as an exhibit.
Q. So the traceability of an Intoxilyzer 8000C
goes back through its calibration. Not its control checks,
but through its calibration, right?
Q. Same if it is recalibrated, the traceability
goes through the standards that are used, the reference
standards. And I think "reference standard" is a technical
Q. The reference standards that are used at the
time of its recalibration.
Q. But there's an added difficulty, and that is
that the reference standards that are used during
recalibration could be, say, 50 milligrams per 100 mills, 100
milligrams per 100 mills, 300 milligrams per 100 mills. My
question for you is, how on earth can a measurement result in
an Intoxilyzer 8000C providing a traceable measurement if we
are talking about indications on the instrument of
approximately 130 milligrams per 100 mills or 140 milligrams
per 100 mills, if the calibration of the instrument was based
on standards at 50, 100 and 300?
A. So I believe we covered this during the last
day. That the calibration of the instrument is fixed at the
time of the calibration procedure, and when it's done, either
by a qualified individual or by a manufacture or a designated
person, company that's able to do that?
A. You're doing a single point check of the
calibration of the instrument using the standard of 100
milligrams of alcohol in 100 millimetres of blood every time
you check the instrument to see if it's working, in proper
working order. That standard determines that the
instrument -- the accuracy of the instrument, so basically is
it able to produce an accurate result. It shows that the
instrument is properly working, and that it hasn't lost its
calibration since the last time it was calibrated.
Q. All right. Again, I want to suggest to you
you're speaking policy, rather than science. My question to
you is, how is a measurement of approximately 130 or
approximately 140 milligrams per 100 mills, on an Intoxilyzer
8000C, traceable to an SI unit? I need a scientific answer.
A. So when the instrument is set up, right -- so
we talked about this, again, on one of the previous days,
that you're looking at the response that the instrument
produces in relation to a specific concentration of a
standard. And we know that a standard -- or alcohol when
it's present in the sample chamber will decrease the signal
that's produced. So you would expect that a standard of 50
will reduce the concentration -- will reduce the intensity of
light as it reaches the detector, different from the standard
at 100, which will have a greater decrease in the signal of
the intensity of reaching the detector, and even greater
again at 300. So you're looking at the response of those
standards relative to the standards that are produced. And
the calibration is then set and that is maintained by the
software, and any deviation in the calibration would be
picked up by the software, which would then trigger an
exception message. And the response of the subject test is
then applied to that calibration curve and it's relative
response is a decrease in the signal is relative to the
calibration curve and the standards that are used to generate
the actual number.
Q. And so there's a calibration curve that is
created. That calibration curve is created using - and this
is at the time of calibration at the factory - using the
reference standards. A calibration curve is created
electronically that relates the response of the instrument to
A. Correct. Yes. And a check of that is done as
either a standalone calibration check or as part of a breath
Q. And, and on the previous day I asked you
questions about how on earth it was possible to check the
configuration of a curve by using one data point.
A. I seem to recall that. I mean, that was about
a month ago.
Q. All right. So let's go back to the reference
standards themselves, because we have the Centre of Forensic
Sciences in the training aid saying that measurement relates
to SI units. We have the Canada Weights and Measures Act
that says that our measurements relate to SI units. That's
section 4(1) of the Weights and Measures Act, just for His
Honour's benefit. But both documents refer to something
called the General Conference on Weights and Measures. The
French name I won't try to pronounce, but its initials are
CGPM, right? And I'm going to pass up a bundle. My friend
received a copy of... no, that?
A. Okay, that's -- can I put this one away for
Q. Yes, please.
THE COURT: You can put it up here if you want.
A. Are you sure?
THE COURT: Yes.
A. Okay. Sorry.
THE COURT: You're going to need a little more
MR. BISS: Q. So from a scientist perspective,
whenever you are expressing a measurement in terms of an SI
unit, the definition for each element of -- each SI unit, and
I understand there are six or seven of them. The definitions
for each SI unit are defined by the International Brochure
for SI Units. So if you can just turn to -- Your Honour, I
put a tab on the copy I sent up to you so you can find it
faster. Page 95 describes what the International Bureau of
Weights and Measures is. It has its headquarters near Paris.
Now, these are the same people who keep the kilogram standard
and who define what a metre is, for example?
Q. And their job is to establish fundamental
standards and scales for the measurement of the principal
physical quantities and maintain the international prototypes
for each of the units of the measure that we use around the
world, in particular in Canada, right?
A. I would assume so, if we are a member,
organization of that society.
Q. Well, in terms of membership in that society,
tab 7, the indication there is Canada became a member in
1907, and the representatives currently is Dr. James McLaren
from the NRC, in the International Bureau of -- Bureau
International Des Poids Et Mesures. So quantities and units
are defined in the SI brochure. That's tab 3 of this bundle,
Q. Okay. Now, at page 103, that's the second
tab, Your Honour, for your benefit. It says:
"The value of a quantity is generally expressed as
the product of a number and a unit."
Now, what's the difference between a number and a
unit? Can you explain that from a scientific perspective?
A. Well, the unit is the reflection of what is
being measured and the number is the value that's actually
Q. So a unit would be something like kilometres
per hour or....
A. Milligrams of alcohol in 100 millilitres of
blood, such as in this case.
Q. That's a unit?
Q. And the number is the number that's associated
Q. And in the second paragraph on that page, it
"In order to establish a system of units, such as
the International System of Units, the SI, is it
necessary first to establish a system of
quantities, including a set of equations defining
the relations between those quantities."
Does that make sense?
Q. So an SI unit is defined by the CGPM in the SI
A. You're using a lot of acronyms and a lot of
A. I'm going to suggest to you that a lot of this
is probably beyond my scope of being able to deal with and to
be able to put it into any kind of context with respect to
Q. All right. But you would agree with me that
from a scientific perspective there is an International
System of Units?
Q. When we were talking about milligrams per 100
mills, we are talking about milligrams as being a fraction of
Q. We are talking about mills as being a fraction
of a litre?
Q. And a litre is defined as a cubic decimetre,
A. I think we discussed that on the previous day,
Q. Right. So the question is, and the point is,
that all measurements in Canada need to be traceable to
reference standards, and the original reference standards are
established by an international organization called the CGPM?
A. I would assume so, yes.
Q. I mean, you don't know of any other place that
establishes the metre that we use in Canada, the kilogram
that we use in Canada?
A. You would have to talk to an organization that
is responsible for that. That's beyond my area of being able
to speak to it.
Q. But there is a concept in science called
traceability of a measurement result. Every measurement
result needs to be traceable back to the original SI units,
Q. So here is my question, a very, very general
question. How can it be the case that every subject test
stands on its own? Because if a subject test is a
measurement result, then surely it is a measurement result
that has to be expressed as a comparison back to the
reference standards, the SI units.
A. Well, in the case of a calibration, there
isn't a calibration standard for 130 milligrams of alcohol in
100 millimetres of blood. So when the calibration curve is
set up, it is extrapolated between the two points, between
100 and the 300, I think you said...
A. ...or between the 100 and the 50.
Q. In the creation of the calibration curve?
Q. And that's why linearity of response is so
A. Or nominal in the area in some cases.
Q. But ultimately what we are hoping to achieve
by designing or by setting up, by calibrating an instrument,
is to take something which is inherently nonlinear and turn
it into a linear response, so that if a given sample is
provided into the sample chamber of the instrument, that we
get an indication that comes out the other end that we can
rely upon, right?
Q. And the way that we do that is by a
calibration of the instrument that takes place at a point in
time earlier than the date of the subject tests.
A. Correct. Yes.
Q. So every subject test does not stand on its
own. Every subject test is really a comparison with the
reference standards that were used when the instrument was
A. I suppose you could look at it that way, yes.
Q. Well, that's the whole concept of traceability
in science. Every measurement result needs to be traceable
back to reference standards. That's, I would suggest to you,
a basic concept in science.
A. Yes. So in this case, this breath test is
traceable back to the standards that were use at the time of
Q. And for His Honour's benefit, I would also
suggest it's a matter of Canadian law because of section 4(1)
of the Weights and Measures Act. So the point is, though,
sir, that measurements need to be traceable. Now, back a
long, long time ago, I sent off to you, and I think I sent
copies to my friend and to the police, a copy of a document
from NIST. What is, what is NIST? It's a U.S. entity?
A. That's correct. It's an organization that's
responsible for... I think I have that document.
THE COURT: Which document?
MR. BISS: The supplementary materials for
A. Okay. Oh, there we go.
Q. So what is, what is NIST?
A. It's an organization in the United States
that's responsibility for traceability of standards.
Q. And NIST has answers in their supplementary
materials of what methodological traceability is, right?
Q. All right. So how do -- what answer do they
give with respect to what traceability is?
A. Well, that's defined in the first paragraph
Q. Yes. And it suggests that it's the property
of a measurement result whereby the result can be related to
a reference through a documented unbroken chain of
calibration, each contributing to the measurement
uncertainty. Do you agree with that definition?
Q. So traceability is not something that we think
of in terms of an instrument. It's not something we think of
in terms of a standard. Traceability is something we think
of in terms of a measurement result that is traceable.
Q. And that's contained in the second section
there under (i)(a)(2). Only measurement results are
Q. (I)(a)(3), it's not organizations that are
traceable, it's a measurement result that is traceable?
Q. Right? So once again, I want to suggest to
you that there's a problem with your approach in saying that
every test stands on its own. My respectful submission is
that every test has to be connected through an unbroken chain
of comparisons, traceable back to the realizations of the SI
A. That is correct. However, each test is what
it is at the time of that particular test. And, again,
another test in time has no impact on that particular test
with respect to --
Q. That isn't what I asked you. I'm not asking
you about causation. I'm asking you about the traceability
of the measurement in science. I want to suggest to you that
if you don't know that the measurement result is traceable,
having a policy that says that every test stands on its own
doesn't tell us whether the measurement result is reliable or
not if you don't know anything about its traceability.
A. We know that the instrument is traceable back
to the standards that were used at the time of calibration.
Proper working order of the instrument is determined at the
time of the test.
A. And the response of that instrument to a
standard is evaluated at that time.
A. And the subject test is compared to the
original standards that were used at the time of calibration.
In this case we know that there was a standard that was
A. ...shortly prior to this case -- or, sorry,
just prior to the breath test.
A. And a response was obtained and a numerical
value was obtained, so there was a decrease in the signal,
A. So then you have a subject test that's
provided that produced a signal that decreased the
sensitivity of the -- decreased the intensity of the light
that reached the detector, so we know that that result is
higher than that standard that was measured at that
Q. All right. But my point is, that in assessing
the traceability of a measurement, one has to make reference
to the calibration. You've been talking about the
calibration by the factory and you've been talking about the
reference standards that were used at the factory, right?
Q. So in determining the traceability of the
measurement result, one has to look back to does the
instrument have a calibration certificate and are there
reference standards that were used at the time of the
calibration certificate. And, fortunately with the
instrument that's before the Court, we have that.
A. Sorry, did you say "unfortunate"?
Q. I said "fortunately".
A. Oh, sorry, okay.
Q. Fortunately we have that. We know that
there's a calibration certificate from the instrument from
2009 that's been made an exhibit in this matter, and we know
that they used NIST reference standard 1828, which we know is
a reference standard, and you looked at the documentation of
NIST standard 1828 on one of the previous days. But the
point is, that that relationship, that traceability
relationship is relevant to an assessment of the reliability
of the results of the measurement results on the day of the
subject tests, right?
A. The result that's obtained is a comparison to
Q. All right.
MR. BISS: Your Honour, could the bundle be made
an exhibit, please, and the NIST document?
THE COURT: Yes. What number are we at?
THE REGISTRAR: Forty-seven.
THE COURT: Okay. So 47 is the International
Measurement Standard binder, and then the NIST
traceability document, N-I-S-T, traceability
document be the next exhibit.
Reference: To build International Measurement Standards binder see bipm.org and download the following:
International Measurement Science Materials from Conférence Générale des Poids et Mesures, (CGPM) and its Bureau International des Poids et Mesures (BIPM)
Tab 1 The Metre Convention
Tab 2 BIPM Member States
Tab 3 SI Brochure
Tab 4 SI Summary
Tab 5 BIPM Governance