Technical Note 33 — February 2016

Issued: June 2006 Amended and reissued: April 2009, December 2009, March 2013, June 2013, February
2016

Guidelines for estimating and reporting

measurement uncertainty of chemical test results
© Copyright National Association of Testing Authorities, Australia 2013
This publication is protected by copyright under the Commonwealth of Australia Copyright
Act 1968.

NATA’s accredited facilities or facilities seeking accreditation may use or copy this
publication or print or email this publication internally for accreditation purposes.

Individuals may store a copy of this publication for private non-commercial use or copy a
reasonable portion of this publication in accordance with the fair dealing provisions in Part
III Division 3 of the Copyright Act 1968.

You must include this copyright notice in its complete form if you make a copy of this
publication.

Apart from these permitted uses, you must not modify, copy, reproduce, republish, frame,
upload to a third party, store in a retrieval system, post, transmit or distribute this content in
any way or any form or by any means without express written authority from NATA.
Technical Note 33 – Guidelines for estimating and reporting MU of chemical test results

Guidelines for estimating and reporting

measurement uncertainty of chemical test results
Introduction
Those making decisions based on test results need to know if the results are sufficiently reliable for the
intended purpose.

Every test or calibration result is subject to some level of error. Estimates of measurement uncertainty (MU)
provide information about how large this error might be. As such, MU is an important part of a reported result
and it may be argued that a result is incomplete unless accompanied by an estimate of MU.

Competent laboratories evaluate and monitor the performance of their test methods and are aware of the
uncertainty associated with the results reported to customers.
1
ISO/IEC 17025 , Section 5.4.6, requires calibration and testing laboratories to have and apply procedures to
estimate the uncertainty of their measurements. Furthermore, Section 5.10.3 of the standard states that test
reports shall include information regarding MU when a customer instructs the facility to provide the
information, when it is relevant to the validity or application of test results, or when it affects compliance to a
specification limit.

This Technical Note is directed primarily to chemical testing laboratories but the general principles also apply
to other areas of testing. Drawing on authoritative references, it aims to define MU, explain why laboratories
should estimate it, then provide general guidance on the estimation and reporting of MU. More detailed,
specific information is provided in the cited references.

What is Measurement Uncertainty?

Measurement uncertainty has been defined as ‘a parameter associated with the result of a measurement
2
that characterises the dispersion of values that could reasonably be attributed to the measurand’ .,or more
lately as a ‘non-negative parameter characterising the dispersion of quantity values being attributed to a
3
measurand, based on the information used’
3
Note 1: The measurand is defined as ‘the particular quantity subject to measurement’ For example, vapour pressure of
a given sample of water at 20°C, mass percentage fat in a sample of cheese, or mass fraction (mg/kg) of a pesticide in a
sample of apples.
4
ASTM describes MU as ‘an estimate of the magnitude of systematic and random measurement errors that
may be reported along with the measurement result’. Whilst this definition perhaps does not properly
embrace the concept of uncertainty it does provide some guidance regarding the factors contributing to MU
and thereby some clues on how it may be estimated.

Both definitions make the important point that MU is associated with a test result not a test method.
However, the MU associated with the results produced by application of a test method is a key performance
characteristic of the method. Some regulators suggest the fitness for purpose of a test method to be judged
5
solely on the MU associated with results .

For a test sample, the true value of the measurand is never known. Considering the ISO definition of MU, the
‘dispersion of values that could reasonably be attributed to the measurand’ implies that the test result,
effectively a range taking the estimated MU into account, will encompass the true result with a particular
probability which depends on the level of confidence provided by the estimate of MU. By convention, in most
instances an expanded uncertainty is estimated, applying a coverage factor to the combined standard
uncertainty to provide a level of confidence of approximately 95%. Laboratories estimating MU in this
manner should be 95% confident that the reported result includes the true value. If required, MU may be
estimated to provide different levels of confidence.

ISO GUM definitions

Standard uncertainty, u(x ) i

Uncertainty of the result, x , of a measurement expressed as a standard deviation.

i

February 2016 Page 3 of 11

Technical Note 33 – Guidelines for estimating and reporting MU of chemical test results
Combined standard uncertainty, u (y) c

Standard uncertainty of the result, y, of a measurement when the result is obtained from the values of a
number of other quantities, equal to the positive square root of a sum of terms, the terms being the variances
or covariances of these other quantities weighted according to how the measurement result varies with these
quantities.

Coverage factor, k
Numerical factor used as a multiplier of the combined standard uncertainty in order to obtain an expanded
uncertainty.

Expanded uncertainty, U
Quantity defining an interval about the result of a measurement that may be expected to encompass a large
fraction of the distribution of values that could reasonably be attributed to the measurand.
Note 2: The ISO GUM refers to the ISO/IEC Guide to the Expression of Uncertainty of Measurement (Reference 2)
Note 3: An expanded uncertainty is calculated from a combined standard uncertainty and a coverage factor, using
U=kxuc

Why estimate measurement uncertainty?

For facilities seeking accreditation to ISO/IEC 17025, the glib answer to this question is ‘because the
Standard requires you to do so!’ However, there are good scientific reasons for the ISO Standard to stipulate
this requirement. A good estimate of MU is necessary for facilities and their customers to:

 ensure results are fit for purpose,

 ensure results are traceable to international or national standards,
 compare results between laboratories and/or specifications, legal tolerances or regulatory limits,
 make informed decisions, and
 improve test methods.

Traceability is another important property of a test result, particularly if it is to be used for legal or regulatory
purposes. Metrological traceability is defined as ‘property of a measurement result whereby the result can be
related to a reference through a documented unbroken chain of calibrations, each contributing to the
3
measurement uncertainty' The last phrase of this definition emphasises that link between MU and
traceability.

How may facilities estimate MU?

There are a number of different approaches to estimating MU. In deciding upon the approach to use,
laboratories should consider the nature of the test, the purpose of the test, the information available on which
to base an estimate, how test results will be used and the risk associated with decisions based on test
results.

If a well-recognised test method specifies limits to the values of the major sources of MU and the form of
presentation of calculated results, a facility is considered to have satisfied the requirements of ISO/IEC
17025, with respect to the estimation of MU, by following the test method and reporting instructions (ISO/IEC
17025, Section 5.4.6.2, Note 2). Nevertheless, laboratories in this situation are advised to estimate MU since
experience has shown that the limits imposed by standard methods are not always sufficient to ensure
results are fit for purpose. If a facility’s estimate of MU is based on the stated reproducibility for a standard
method then the facility should ensure that the estimate covers all sources of uncertainty pertaining to their
measurements.

The ISO/IEC 17025 Standard (Section 5.4.6.2) also recognises that the nature of some test methods may
preclude vigorous, metrologically and statistically valid calculation of MU. In such cases, laboratories are still
required to make a reasonable estimate of MU based on, for example, professional judgement and
experience, knowledge of method performance and validation data. Such estimates must not give a wrong
impression of MU.

The many different approaches for the estimation of MU may include components derived from what may be
broadly categorised into either ‘bottom-up’ or ‘topdown’ calculations.

February 2016 Page 4 of 11

Technical Note 33 – Guidelines for estimating and reporting MU of chemical test results
The fundamental metrological ‘bottom-up’ approach described in the ISO/IEC Guide to the Expression of
2
Uncertainty in Measurement , commonly called the GUM, combines the uncertainties associated with all
individual operations of an analytical procedure to calculate the combined standard uncertainty and, after
multiplication by a coverage factor, the expanded uncertainty (usually the 95% confidence range). The GUM
adequately covers physical measurements but does not specifically address the particular problems related
to chemical and biological measurements. Whilst it sets down general principles and guidance for estimating
the uncertainty associated with single steps of analytical procedures (weighings, volumes etc.), calculated
concentrations of standard solutions, spiked concentrations or relatively simple tests, the GUM approach is
not easily applied to more complex test procedures. It can be very laborious, even if spreadsheets are used
6
for calculations. Horwitz considers the approach inappropriate for chemical tests because it ignores the fact
that the uncertainties associated with some of the numerous factors influencing test results tend to cancel
out. Furthermore, it is argued that chemical test results are often influenced by factors that overwhelm the
6,7
uncertainties considered in the GUM approach .

An often-stated advantage of the GUM approach is that it provides a clear understanding of the analytical
operations that contribute significantly to MU, allowing the analyst to focus on improving these operations if
required to reduce the MU associated with test results. However, experienced analysts are most likely able
to identify the critical ‘uncertainty’ steps of analytical procedures independent of any estimates of MU.
8
Eurachem/CITAC provides guidance on how the GUM approach may be applied to chemical
measurements. It also includes examples using a ‘top-down’ approach. A ‘top-down’ approach utilises data
from method validation, intra-laboratory QC and/or inter-laboratory studies. It is reasoned that the use of
such data, generated over several months, will maximise the probability of including all potential
contributions to uncertainty in estimates of MU. Numerous references, aimed at providing chemical testing
facilities with practical guidance and examples of simple ‘top-down’ approaches to estimating MU are
9-13
available.

A ‘top-down’ approach is generally more practical than the GUM approach for estimating the MU of chemical
test results. However it is the facility’s decision to use the method most appropriate for their circumstances
and supported by the available data.

The following sub-sections provide examples of how a top-down approach may be applied to estimate MU. A
systematic procedure for obtaining fit-for-purpose estimates of MU is presented in Figure 1 and a practical
example of how MU may be estimated utilising data on bias and precision is shown in Appendix 1.

Estimating measurement uncertainty from reproducibility studies

The standard deviation determined from inter-laboratory studies under reproducibility conditions, s , may be
R
11
used as an approximation of the combined uncertainty associated with a result. This estimate is then
doubled to give the expanded uncertainty (95% Confidence Interval).

u ≈ s , and
C R

U=2s R

where: u = combined standard uncertainty

C

s = standard deviation under reproducibility conditions

R

U = expanded uncertainty, 95% confidence interval

In the absence of data from inter-laboratory studies on a particular method, the reproducibility standard
0.85
14-16
deviation may be estimated from an equation proposed by Horwitz. The Horwitz equation , s = 0.02C ,
R

was empirically derived by plotting reproducibility standard deviation vs concentration, C (in g/g) for more
than 7000 interlaboratory studies. This equation may be used to estimate s and U at different
R

concentrations. It may also be used to check the validity of estimates of s from inter-laboratory studies. The
R

equation predicts that lowering analyte concentration by two orders of magnitude will double the between-
laboratory relative standard deviation, RSD , associated with a test result. However, advances in analytical
R

chemistry with the introduction of techniques such as isotope-dilution mass spectrometry, have provided the
capability to achieve very low limits of quantitation with less uncertainty than predicted by the Horwitz
17
equation. Thompson and Lowthian have reported that laboratories tend to out-perform the Horwitz function
at concentrations below 10 µg/kg.

A facility may estimate MU based on the reproducibility reported for a standard method, however the facility
should first ensure that they are able to achieve the precision stated for the method.

February 2016 Page 5 of 11

Technical Note 33 – Guidelines for estimating and reporting MU of chemical test results
If results are corrected for bias or the bias is small, estimates of MU based on s calculated using the Horwitz
R

equation, are usually conservative overestimates, however such estimates may suffice if there is no
requirement for a more rigorous estimate of the uncertainty associated with results.

Estimation of MU from within-laboratory data of bias and precision

The MU associated with a result indicates the accuracy of the result; the smaller the 95% confidence interval
the more accurate the result. There are two components to accuracy; trueness (quantitatively expressed as
bias) and precision. Therefore, a reasonable estimate of MU may be gained from information on the bias and
precision associated with a test result. This information may be gained from participation in suitable inter-
11-13
laboratory and/or intra-laboratory studies and used to estimate MU.

Unfortunately, suitable inter-laboratory studies are not always available and more often than not, laboratories
are required to estimate MU from data generated ‘in-house’.

Ideally, precision and bias should be determined within the same analytical run as the sample(s) analysed.
This is generally impractical and cost-prohibitive for most chemical tests. However, if suitable QC samples
(matrix-matched Certified Reference Materials [CRMs], other suitable Reference Materials [RMs] or spiked
samples) are run within each sample batch to ensure the test method is consistently operating within
acceptable control limits, then the data generated may be used to evaluate both intermediate reproducibility
(within-lab precision under reproducibility conditions) and average bias over a given period of time or number
of sample batches. Precision and bias should be evaluated at the concentration(s) most relevant to the
purpose of the test.
Note 4: Estimates of MU may also be gained from the evaluation of precision and bias during initial validation of a
method.

Considering a simple model for the result of a chemical test conducted under the above conditions:

y = ytrue + b + e (Equation 1)
where: y = observed measurement, uncorrected for bias;
ytrue = true result;
b = bias, defined as y - ytrue , the difference between the mean of a large number of observed
results ( y ) and the true result
e = random error for within-laboratory reproducibility conditions, s . L

Expression for combined standard uncertainty

The combined standard uncertainty of y, may be estimated as:
2 2 2
u (y) = s + u
C L b
(Equation 2)

where: u (y) = combined standard uncertainty of y

C

s L
= standard deviation of results obtained under within-laboratory reproducibility conditions;
u b
= standard uncertainty associated with the measurement of bias.

Estimating bias and the uncertainty associated with the measurement of bias
In practice, the true result is not known and it is necessary to base estimates of bias on the expected result
for the analysis of a CRM, or other suitable sample.

b= y -y (Equation 3)
exp

where yexp is the expected result

2 2 2
u =
b
u( y ) + u(y ) exp
(Equation 4)
where: u( y ) = standard uncertainty of observed result;
u( yexp) = standard uncertainty of expected result.

If the observed result is taken to be the mean result from replicate analyses of a CRM or other suitable
sample, performed over several months, the standard uncertainty may be taken as the standard deviation of
the mean of the observed result.

February 2016 Page 6 of 11

Technical Note 33 – Guidelines for estimating and reporting MU of chemical test results

sL
i.e. u( y ) = (Equation 5)
n

where n = number of replicates performed.

The standard uncertainty of the expected result may be known from the certified value of a CRM, the
characterisation of a secondary reference material, estimated from the uncertainties associated with spiking,
or estimated in some other way, as the particular situation demands.

Once b and u are estimated according to Equations 3 and 4, it is important to test whether or not the bias,
b

taking into account the uncertainty associated with its measurement, is significant. If |b| > t (0.05, n-1)u , b

where t is the Student-t value at n-1 degrees of freedom, then bias is significant and steps must be taken to
account for bias when calculating results or estimating MU. (see below) If the average bias is based on 20
or more measurements, |b| can be simply compared with 2 u in order to check if the average bias is
b

significant.

Estimating combined standard uncertainty and expanded uncertainty

If bias is not significant, the combined standard uncertainty may be calculated from Equation 2.

If u (y) is based on sufficient data the expanded uncertainty, U, may be calculated using a coverage factor of
C

2 to give an approximate level of confidence of 95%.

U (95% confidence interval) = k u (y)

C

= 2 u (y)
C
(Equation 6)
8
Note 5: The Eurachem/CITAC Guide , Section 8.3, recommends that for most purposes k is set at 2. The Guide states
that this value may be insufficient where the combined uncertainty is based on statistical observations with relatively few
degrees of freedom (less than about six) and further recommends that in such cases k be set equal to the two-tailed
value of the Student’s t for the number of degrees of freedom associated with the observations and the level of
confidence required (normally 95%). It should be noted that about 19 degrees of freedom are required for k to be less
than 2.1.
If bias is significant and based on reliable data such as the analysis of a CRM, then the measurement result
2
should always be corrected for bias. Where excessive bias is detected, action should be taken to
11
investigate, and if possible, eliminate the cause of the bias. Such actions should at least reduce bias to
acceptable levels before a facility proceeds with analyses of test samples.

If results are not corrected for significant bias, the MU must be enlarged to ensure that the true result is
encompassed by the reported confidence interval.

A number of approaches have been described for taking significant bias into account when estimating MU.
18
O’Donnell and Hibbert critically evaluated different approaches and concluded the best estimate of the
expanded uncertainty to be;

U = k u (y) + |b run| , where b run is the run bias

C

Note 6: The symbols used in the above equation are consistent with those used in this Technical Note, rather than those
used by O’Donnell and Hibbert.
O’Donnell and Hibbert specify run bias in their recommended equation for enlarging the estimate of MU, and
provide explanation for using this approach. However, a reasonable estimate of MU is obtained if average
total bias, evaluated as described above, is used in place of run bias.

U = k u (y) + |b|
C
(Equation 7)

Failure to include the uncertainty of the estimation of bias in estimates of combined uncertainty or failure to
correct for significant bias or enlarge the uncertainty of the result to account for it, invalidates both the test
result and the estimated MU.

Laboratories should ensure that MU is estimated (at least) at the concentrations most relevant to the purpose
of the test, for example regulatory, legal or specification levels.

Reality checks for estimates of MU

Although ISO/IEC 17025 has been in place since 1999, it is fair to say that the estimation of MU is still not a
highly developed capability of many chemical testing laboratories. The widely different estimates reported by
February 2016 Page 7 of 11
Technical Note 33 – Guidelines for estimating and reporting MU of chemical test results
laboratories using essentially the same methodology in Proficiency Tests provides evidence that further work
is needed. In some cases, the reported MU has been clearly over-estimated or under-estimated. Such
mistakes may be identified if laboratories critically evaluate their estimates and carry out simple reality
checks to ensure they are consistent with their own knowledge and experience. Such reality checks include
comparison with collaborative trial statistics, comparison with inter-laboratory reproducibility and examination
19
of proficiency trial results.

Reporting Measurement Uncertainty

8
The Eurachem/CITAC Guide states that ‘unless otherwise required, the result x should be stated together
with the expanded uncertainty U, calculated using a coverage factor, k = 2.’

The recommended form for reporting a result is:

(Result): x ± U (units).

The coverage factor used to calculate U should be stated, for example;

'DDT: 3.52 ± 0.14 mg/kg
The reported uncertainty is an expanded uncertainty calculated using a coverage factor of 2 which
gives a level of confidence of approximately 95%’.

The value and its uncertainty should both be reported in the same units. Results reported as x ± (a
percentage of x) are not recommended.

The value of a result and its uncertainty should not be reported with an excessive number of significant
figures. It is rarely, if ever, necessary to report a chemical test result with the uncertainty stated to more than
2 significant figures. Results should be rounded to be consistent with their uncertainty.

It is recognised that some Laboratory Information Management Systems may not be able to readily report
MU in the recommended format. Other presentations are acceptable providing the essential elements of the
recommended format are clearly covered in the test report.

For further information regarding MU, contact your client coordinator or the Sector Manager, Life Sciences
Neil Shepherd, in NATA’s Melbourne office on (03) 9274 8200, email:
st
[email protected], 1 Floor, 2-6 Railway Parade, CAMBERWELL VIC 3124

References
1. ISO/IEC 17025 (2005) General requirements for the Competence of Calibration and Testing
Laboratories
2. ISO/IEC Guide 98-3 (2008) Uncertainty of Measurement-Part 3: Guide to the Expression of
Uncertainty in Measurement (GUM 1995)
3. ISO/IEC Guide 99 (2007), International Vocabulary of Metrology-Basic and general concepts and
associated terms, (VIM3)
4. ASTM International (2005) Form and Style of ASTM Standards
5. EC Directive 2001/22/EC
6. Horwitz, W (2003), J. of AOAC International, 86, 109–111;
7. Eurolab (2002) Measurement Uncertainty in Testing, Technical Report No. 1/2002
8. Eurachem/CITAC (2002), Eurachem/CITAC Guide, Quantifying Uncertainty in Analytical
Measurement, 2nd Edition: www.eurachem.ul.pt
9. www.measurementuncertainty.org
10. Barwick V.J. and Ellison S. L. R. (2000), Protocol for uncertainty evaluation from validation data, VAM,
Report No. LGC/VAM/1998/088
11. ISO (2004), Guidance for the use of Repeatability, Reproducibility and Trueness Estimates in
Measurement Uncertainty Estimation, ISO/TS 21748
12. Magnusson B., Naykki T., Hovind H. and Krysell M. (2003), Handbook for Calculation of Measurement
Uncertainty in Environmental Laboratories, NORDTEST Report TR537
13. Eurolab (2007) Measurement uncertainty revisited: Alternative approaches to uncertainty evaluation
Technical Report 1/2007, www.eurolab.org
14. Horwitz W. Kamps L. R. and Boyer K. W (1980), J. Assoc. Off. Anal. Chem., 63, 1344–1354
15. Horwitz W. (1982), Anal. Chem. 54, 67A–76A

February 2016 Page 8 of 11

Technical Note 33 – Guidelines for estimating and reporting MU of chemical test results
16. Boyer K. W., Horwitz W. and Albert R. (1985), Anal. Chem., 57, 454–459
17. Thompson M and Lowthian P J (1997), Journal of AOAC International, 80(3), 676-679
18. O’Donnell G. E. and Hibbert D. B. (2005) Analyst, (130), 721–729
19. Royal Society Chemistry (2003), Is my Uncertainty Estimate Realistic? Analytical Method Committee
Technical Brief No. 15

February 2016 Page 9 of 11

Technical Note 33 – Guidelines for estimating and reporting MU of chemical test results

Appendix 1:
Estimation of MU from within-laboratory data on bias and precision
Scenario
A facility has validated a method for the determination of residues of the pesticide, chlorpyrifos in tomatoes
and has applied the method on customer samples submitted over a period of eighteen weeks. A matrix-
matched certified reference material (MX 001) obtained from the NMI was used during the validation process
and henceforth once a fortnight to supplement spiked samples routinely used for within-batch QC. The
certified value for chlorpyrifos in the CRM is 0.489 ±0.031 mg/kg (95% confidence range).

The facility initially estimated the MU of chlorpyrifos results based solely on validation data but now wishes to
update its estimate, taking into account the QC data generated during routine analysis of customer samples.

The facility analysed 7 replicates of the CRM during method validation and a further 9 replicates during
routine analyses. The following data was generated via these processes;
Validation data: mean 0.391 mg/kg, standard deviation ( s r ) 0.051 mg/kg, CV 13.0%
QC data: mean 0.385 mg/kg, standard deviation ( s R ) 0.082 mg/kg, CV 21.3%

Control charts plotted for both the CRM and spiked samples run as QC samples demonstrate the test
method to be under good control.

The average result was consistent for both validation and QC analyses, however, as expected, the results of
the QC tests conducted under intra-laboratory reproducibility conditions were less precise than those for
method validation conducted under near repeatability conditions. Both the % recovery (negative bias) and
precision were considered acceptable for the test.

It was considered important to estimate the MU associated with results close to 0.5 mg/kg, the maximum
residue limit (MRL) for chlorpyrifos in tomato. No doubt with the MRL in mind, the CRM produced by the NMI
has a certified value for chlorpyrifos closely matching the MRL.

Estimation of MU
Data
The mean of all results is the best estimate of the likely batch-to-batch result. This value (the average of all
validation and QC replicate measurements) is the average observed result, y = 0.388 mg/kg; mean
recovery = 79.3%

The standard deviation of the QC results best reflects the imprecision of results produced under normal test
conditions; s L = s R = 0.082 mg/kg

The expected result, y exp , is the certified value of the CRM; 0.489 ± 0.031 mg/kg
The standard uncertainty of the expected result, u y exp  is obtained by halving the expanded uncertainty
(95% confidence range) stated for the CRM;
 
u y exp = 0.031/2 = 0.0155 mg/kg.

Estimation of bias and the uncertainty of bias

Bias ( b ) = y - y exp (Equation 3) = 0.388 – 0.489 = - 0.101 mg/kg
This is the average bias and the best available estimate of the bias that might apply to any individual test
result.

Uncertainty of bias (from equation 4), ub = 

u y  u y exp 
2 2

Uncertainty of the observed result (Equation 5), 

uy =
sL
= 0.082/3 = 0.027 mg/kg
n

February 2016 Page 10 of 11

Technical Note 33 – Guidelines for estimating and reporting MU of chemical test results

ub = 
u y  u y exp  =
2 2
0.0272  0.01552 = 0.031 mg/kg

Is the bias significant?

Is b > t (0.05, n-1) u b ?
Here, n = 9 since the uncertainty of the observed result is based on 9 replicate tests
From t tables (critical values for two-tailed student t-tests), t (0.05, 8) = 2.306
Is 0.101 > 2.306 x 0.031? Is 0.101 > 0.071? Answer YES, so bias IS significant

It is necessary for the facility to correct results for bias (preferable alternative) or enlarge their estimate of MU
to account for uncorrected significant bias.

Estimating combined standard uncertainty and expanded uncertainty

(i) results corrected for bias
If the facility decided to correct results at or about the MRL for the negative bias, the correction could be
achieved by simply adding the bias to the measured value, assuming the bias to be constant over a narrow
concentration range. There is uncertainty associated with this correction, and it is therefore necessary to
include the uncertainty of bias in the combined standard uncertainty, irrespective of whether or not results
are corrected for bias.

Assuming a raw test result = 0.35 mg/kg.

Corrected result = 0.45 mg/kg

Combined standard uncertainty, (from Equation 2)

uC  y  = s L 2  ub 2 = 0.0822  0.0312 = 0.088 mg/kg
Expanded uncertainty (Equation 6);
U (95% confidence interval) = 2 u C  y  = 0.18 mg/kg

The facility would report the result as 0.45 ± 0.18 mg/kg, noting that the reported uncertainty is an expanded
uncertainty calculated using a coverage factor of 2 to give a level of confidence of approximately 95%.

(ii) results not corrected for bias

Expanded uncertainty (Equation 7);
U  2u c ( y)  b

= 0.18 + 0.10

= 0.28 mg/kg

The facility would report the result as 0.35 ± 0.28 mg/kg.

When the expanded uncertainty is enlarged in this manner to account for uncorrected significant bias, the
reported result encompasses a wide 95% confidence range that is wider than justified on the lower side of
the measured value, since the uncertainty allows for both negative and positive bias although only negative
bias is present.

The result corrected for bias provides the customer with a better estimate of the true result and a better basis
for decision-making, although as expected when the ‘true’ value is close to a limit, the result is equivocal with
respect to compliance with the MRL.

AMENDMENT TABLE

The table below provides a summary of changes made to the document with this issue.
Section Amendment
All Minor Formatting Changes

February 2016 Page 11 of 11