Professional Documents
Culture Documents
Calculation of Measurement Uncertainty: Handbook
Calculation of Measurement Uncertainty: Handbook
for
Calculation of
Measurement Uncertainty
in
Environmental Laboratories
Handbook
for
Calculation of
Measurement Uncertainty
in
Environmental Laboratories
Contents:
1
2
EXAMPLES................................................................................................... 20
6.1
6.2
6.3
REPORTING UNCERTAINTY.................................................................... 30
REFERENCES............................................................................................... 32
10
APPENDICES................................................................................................ 33
Mean value
u(x)
uc
sr
sR
Rw
sRw
Page 1 of 41
CRM
Certified
value
Nominal
value
u(Cref)
bias
u(bias)
Systematic
component
material certification.
Page 2 of 41
2 Introduction
2.1 Scope and field of application
This handbook is written for environmental testing laboratories in the Nordic
countries, in order to give support to the implementation of the concept of
measurement uncertainty for their routine measurements. The aim is to provide a
practical, understandable and common way of measurement uncertainty
calculations, mainly based on already existing quality control and validation data,
according to the European accreditation guideline /12/, the Eurolab Technical
Report No. 1 /3/ and the ISO/DTS 21748 Guide /8/. Nordtest has supported this
project economically in order to promote and enhance harmonisation between
laboratories on the Nordic market.
Practical examples, taken directly from the everyday world of environmental
laboratories, are presented and explained. However, the approach is very general
and should be applicable to most testing laboratories in the chemical field.
The handbook covers all steps in the analytical chain from the arrival of the sample
in the laboratory until the data has been reported. It is important to notice that vital
parts of the total measurement uncertainty are not included, e.g. sampling, sample
transportation and possible gross errors during data storage/retrieval.
The recommendations in this document are primarily for guidance. It is recognised
that while the recommendations presented do form a valid approach to the
evaluation of measurement uncertainty for many purposes, other suitable
approaches may also be adopted see references in Section 9. Especially the
EURACHEM/CITAC-Guide /2/ is useful in cases where sufficient previous data is
not available, and therefore the mathematical analytical approach according to
GUM /1/ with all different steps is to be used.
Basic knowledge in the use of quality control and statistics is required. In order to
make it possible for the reader to follow the calculations, some raw data is given in
appendices
Page 3 of 41
250
Measurement
uncertainty from...
240
Lab 1
230
Lab 2
Ammonium g/l
220
210
200
190
180
Control sample
Rw
170
160
150
Figure
1. Comparing ammonium results from two laboratories, Lab 1 = 199 g/L and Lab6
0
2 = 188 g/L. To the left the error bars are calculated from results on control samples
( 1s) and to the right the error bars are expanded measurement uncertainty.
The customer needs it together with the result to make a correct decision. The
uncertainty of the result is important, e.g. when looking at allowable (legal)
concentration limits
The laboratory to know its own quality of measurement and to improve to the
required quality
Page 4 of 41
The basis for the evaluation is a measurement and statistical approach, where
the different uncertainty sources are estimated and combined into a single
value
Basis for the estimation of measurement uncertainty is the existing knowledge
(no special scientific research should be required from the laboratories).
Existing experimental data should be used (quality control charts, validation,
interlaboratory comparisons, CRM etc.) /12/
Guidelines are given in GUM /1/, further developed in, e.g., EA guidelines
/12/, the Eurachem/Citac guide /2/, in a Eurolab technical report /3/ and in
ISO/DTS 21748 /8/
!
!
Page 5 of 41
QC within laboratory, Rw
(Section 4)
Sampling
Value
Analytical
Report
Customer
Reproducibility
between
laboratories sR
(Section7)
Page 6 of 41
Page 7 of 41
Step
Action
Specify Measurand
Quantify u(Rw)
A control sample
B possible steps not
covered by the
control sample
Quantify systematic
components
Convert components
to standard
uncertainty u(x)
1
2
bias
+ sbias 2 + u(Cref ) 2
u(bias) =
3
2
2.2
= + 0.6 2 + 1.5 = 2.05
3
Calculate combined
standard uncertainty,
uc
Calculate expanded
uncertainty,
U = 2 uc
Page 8 of 41
Relative u(x)
Comments
Reproducibility within-laboratory, Rw
Control sample
X
sRw
= 200 g/L
Other components
Systematic components
Reference material
bias
--
Interlaboratory
comparisons
bias
bias = 2.2 %
2.05 %
2.2
+ 0.6 2 + 1.52
3
sbias = 0.6 %
u(Cref) = 1.5%
Recovery test
u(bias)=
bias
--
sR
--
Standard method
sR
--
8.8 %
Combined uncertainty, uc is calculated from the control sample data and bias
estimation uncertainty components from interlaboratory comparisons. The sR can
also be used (see 7.2) if a higher uncertainty estimation is acceptable.
Measurand
Ammonium
Combined Uncertainty uc
Expanded Uncertainty U
2.64 2 = 5.3 5 %
Page 9 of 41
It is of utmost importance that the estimation must cover all steps in the analytical
chain and all types of matrices worst case scenario. The control sample data
should be run in exactly the same way as the samples e.g. if the mean of duplicate
samples is used for ordinary samples, then the mean of duplicate control samples
should be used for the calculations.
It is likewise important to cover long-term variations of some systematic
uncertainty components within the laboratory, e.g. caused by different stock
solutions, new batches of critical reagents, recalibrations of equipment, etc. In
order to have a representative basis for the uncertainty calculations and to reflect
any such variation the number of results should ideally be more than 50 and cover
a time period of approximately one year, but the need differs from method to
method.
Page 10 of 41
two control sample levels (20 g/L and 250 g/L) were used during year 2002. The
results for the manual analysis method are presented in the table below.
Value
Relative u(x)
Comments
Reproducibility within-laboratory, Rw
Control sample 1
X = 20.01 g/L
sRw
Control sample 2
X = 250.3 g/L
sRw
Standard
deviation 0.5
g/L
Standard
deviation 3.7
g/L
--
Other components
2.5 %
1.5 %
From
measurements in
2002, n = 75
From
measurements in
2002, n = 50
s = range / 1.128
Value
Relative u(x)
Comments
Reproducibility within-laboratory, Rw
Variation from
duplicate analysis
2-15 g/L:
> 15 g/L:
sR
5.7 %
3.6 %
n = 43 ( X = 6.50
g/L)
n = 30 ( X = 816
g/L)
Page 11 of 41
At low levels it is often better to use an absolute uncertainty rather than a relative,
as relative numbers tend to become extreme at very low concentrations. In this
example the absolute u(r) becomes 0.37 g/L for the natural sample (mean
concentration 7 g/L) and 0.5 g/L for the control sample in Section 4.2 (mean
concentration 20 g/L).
As the estimate from duplicate analysis gives the repeatability component (sr) only,
it should be combined with the control sample results from Section 4.2 to give a
better estimate of sRw. This way, the repeatability component will be included two
times, but it is normally small in comparison to the between-days variation.
Value
u(x)
Comments
Reproducibility within-laboratory, Rw
Low level
(2-15 g/L)
High level
(> 15 g/L)
sRw
sRw
absolute
u(x) =
0.5 2 + 0.37 2
relative
u(x) =
1.5 2 + 3.6 2
It can be noticed that the sample matrix has some effect on the variation of the
results. The reason for this is not only the matrix, but also the concentration level,
as almost all of the duplicate analyses were performed at a concentration level
below 10 g/L. The quantification limit of the measurement was 2 g/L and the
relative standard deviation usually becomes higher near that limit (cf. Figures 4 and
5 in Section 6.3).
Page 12 of 41
sample. The standard deviation for the results can be estimated from the average
range of the duplicate samples (see Appendix 8), and in this case becomes 0.024.
The control limits at 2s thus lies at 0.048. The average value of the first
determination is 7.53, and the u(x) thus equals 100*0.024/7.53 = 0.32 %.
Difference, mg/L
0.15
0.10
0.05
0.00
-0.05
-0.10
-0.15
0
10
20
30
40
50
Analysis no.
Figure 3. The difference between oxygen duplicate measurements samples
plotted in an R-chart
However, this only gives the within-day variation (repeatability, sr) for sampling
and measurement, and there will also be a long-term uncertainty component from
the variation in the calibration (here the thiosulphate used for titrating or the
calibration of the oxygen probe, depending on method). For this particular analysis,
the uncertainty component from the long-term variation in calibration is hard to
measure, as no stable reference material or CRM is available for dissolved oxygen.
One method would be to calibrate the same thiosulphate solution several times
during a few days time, and use the variation between the results. Here we choose
to estimate that component by a qualified guess, but laboratories are encouraged to
also try the experimental approach.
The total reproducibility within-laboratory for dissolved oxygen thus becomes:
Page 13 of 41
Value
Relative u(x)
Comments
Reproducibility within-laboratory, Rw
Duplicate
measurements of
natural samples,
difference used in rchart
Estimated variation
from differences in
calibration over time
sR
s = 0.5 %
0.5 %
Page 14 of 41
From
measurements in
2000-2002, n=
50
Estimate, based
on experience
bias sbias
2
u (bias ) =
+
+ u (Cref )
3 ?
and if bias is tested not significant or the result is corrected for bias, the u(bias) is
2
s
u (bias ) = bias + u (Cref ) 2
?
The ? can be n if the data is from the same population, same CRM or same
matrix and 1 if the data is from different CRMs or matrices - see below.
Ref 14 uses bias/2 but here is used bias / 3 in order to get closer to 95 % confidence
limit for the expanded measurement uncertainty
Page 15 of 41
3
4
Quantify systematic
components
Convert components
to standard
uncertainty u (x)
bias sbias
u (bias ) =
+
3 n
2
+ u (Cref ) 2 =
3.48 2.2
+
+ 2.212 = 3.1 %
3 12
Case 2: The results are corrected for bias
s
u (bias ) = bias
n
+ u (Cref ) 2 =
2.2
+ 2.212 = 2.3 %
12
If no t-test is made we recommend Case 1. If a t-test is made Case 2 can be used if
the bias is not significant. However, when there is only a small difference in the
estimation of the u(bias), Case 1 can still be used.
If several CRM:s are used, we will get different values for the bias. The bias
estimation (step 1) will be the average of the bias estimates, sbias will be calculated
from the different bias estimates and step 3 will be the average of confidence
interval of the CRM:s. The uncertainty of the bias estimation will be calculated in
the following way (see also section 5.2).
Quantify systematic
components
Page 16 of 41
Convert components
to standard
uncertainty u (x)
bias sbias
2
u (bias ) =
+
+ u (Cref ) =
1
3
2
1.82 2.4
2
+
+ 1.9 = 3.2 %
3 1
2
u (Cref ) =
sR
9
=
= 2.6
n
12
Page 17 of 41
Quantify systematic
components
Convert components
to standard
uncertainty u (x)
bias sbias
2
u (bias ) =
+
+ u (Cref ) =
3 1
2
3.5 3.3
2
+
+ 2.6 = 5.5 %
3 1
2
5.3 Recovery
Recovery tests, for example the recovery of a standard addition to a sample in the
validation process, can be used to estimate the systematic error. In this way,
validation data can provide a valuable input to the estimation of the uncertainty.
In an experiment the recoveries for an added spike were 95 %, 98 %, 97 %, 96 %,
99 % and 96 % for 6 different sample matrices. The average is 96.8 %
The mean bias is 3.2 % and the standard deviation of the bias is 1.47 %.
Uncertainty component from the definition of 100% recovery, u(Crecovery)
Step
Example
Uncertainty of the 100%
Certificate 1.2 % (95 % conf. limit) gives u(conc) =
recovery. Main components 0.6 %
are concentration, u(conc) u(vol) - This value can normally be found in the
of standard and volume
manufacturers specifications, or better use the limits
added u(vol)
specified in your laboratory. Max bias 1 % and
repeatability max 0.5 %
2
1
u (vol ) =
+ 0.5 2 = 0.76 %
3
Calculate u(Crecovery)
Page 18 of 41
Quantify systematic
components
Convert components
to standard
uncertainty u (x)
bias s bias
+
u (bias ) =
+ u (Cre cov ery ) 2 =
3 1
2
3.2 1.47
2
+
+ 1.0 = 2.6 %
1
3
Page 19 of 41
6 Examples
In this chapter, practical examples on how measurement uncertainty can be
calculated using the method of this handbook are given.
Exercis
e
1
2
3
X
s(bias)
Nominal
value
mg/L
154
219
176
Laboratory
result
mg/L
161
210
180
Bias
sR
Number of labs
%
+ 4.5
- 4.1
+2.3
+0.9
4.47
%
7.2
6.6
9.8
7.872
-
23
25
19
22.3
-
Page 20 of 41
1
2
Action
Specify Measurand
Quantify u(Rw)
A control sample
B possible steps not
covered by the
control sample
Quantify Systematic
Components
Convert components to
standard uncertainty
u(x)
u(Rw) = 2.6 %
2
bias
2
u(bias) =
+ sbias + u(Cref )2
3
2
4.3 2.6
2
=
+
+ (1.2) = 2.82%
3 19
Calculate combined
standard uncertainty, uc
Calculate expanded
uncertainty, U = 2 u c
uc =
U = 2*3.84 = 7.68 8 %
Page 21 of 41
1
2
Action
Specify Measurand
Quantify u(Rw)
A control sample
B possible steps not
covered by the
control sample
Quantify systematic
components
u (Cref ) =
Convert components to
standard uncertainty
u(x)
sR
n
7.9
22.3
= 1.67
u(Rw) = 2.6 %
2
bias
2
u(bias) =
+ sbias + u(Cref )2
3
0.9 4.47
2
=
+
+ (1.67) = 4.80
3 1
Calculate combined
standard uncertainty, uc
Calculate expanded
uncertainty, U = 2 u c
Page 22 of 41
uc =
1
2
Action
Specify Measurand
Quantify u(Rw)
A control sample
B possible steps not
covered by the
control sample
Quantify
systematic
components
Convert components to
standard uncertainty
u(x)
The u(Rw) = 8 %
2
bias sbias
u(bias) =
+ u(Cref ) 2
+
n
3
2
5.3 8
2
= +
+ (4.7) = 5.88
3 22
Calculate combined
standard uncertainty, uc
uc = 8 2 + 5.88 2 = 9.9 10 %
Calculate expanded
uncertainty, U = 2 u c
U = 2*10 = 20 %
Page 23 of 41
Quantify u(Rw)
A control sample
B possible steps not
covered by the
control sample
Quantify systematic
components
A: The control sample, which is a stable inhouse material, gives sRw= 8 % at a level of
150 g/kg dry matter. sRw = 8 % is also used
as s when setting the control chart limits.
B: The analysis of the control sample includes
all steps except for drying the sample to
determine the dry weight. The uncertainty
contribution from that step is considered small
and is not accounted for.
Participation in 3 interlaboratory comparisons
with concentration levels similar to the
internal quality control. The bias in the 3
exercises has been 2 %, -12 % and 5 %.
The average bias is 6 %.
The sbias of the three bias estimates is 5.1 %
The sR in the three exercises has been 12 %,
10 % and 11 %, on average sR = 11 % (n=14)
u (Cref ) =
Convert components to
standard uncertainty
u(x)
11
= 2.9 %
14
The u(Rw) is 8 %
2
bias
2
u(bias) =
+ sbias + u (Cref )2
3
2
6 5.1
2
= + + (2.9) = 6.8
1
3
Page 24 of 41
Step
Action
Calculate combined
standard uncertainty, uc
Calculate expanded
uncertainty, U = 2 u c
U = 2*10,5 = 21 20 %
Reproducibility within-laboratory, Rw
Control sample
X = 160 g/kg
dry weight
Other components
u(Rw)
8%
Systematic components
Reference
material
5.3/ 3 = 3.1 %
2.25/150 = 1.5 %
u(bias)
3,12 + 1,52 + 4,72
=5,8
U(Cref) = 9.2 %
9.2/1.96 = 4.7 %
1 Mean bias 6 %. 6/ 3 = 3.5 %
u(bias)
2 The s(bias) is
3,52 + 5,12 + 2,92
5.1 %
5.1 %
=6,8
3 sR (n=14) = 11 11/ 14 = 2.9 %
%
Combined uncertainty, uc, is calculated from internal quality control and the
maximum systematic component - interlaboratory comparisons.
Interlaboratory
comparison
n=3
Page 25 of 41
Measurand
PCB
Combined Uncertainty uc
uc =
Expanded Uncertainty U
2*uc = 21 % 20 %
8 + 6,8 = 10.5
2
Conclusion: In this case the calculation of the systematic component gives very
similar results regardless of whether CRM or interlaboratory comparison results are
used. Sometimes interlaboratory comparisons will give considerably higher values,
and it might in such cases be more correct to use the CRM results.
Pb, %
recovery
109.7
125.2
91.8
98.4
98
100.5
105.5
s, %
53.8
12.1
5
3.0
1.7
1.3
1.4
s, %
40
30
20
10
0
0.01
0.1
1
Concentration (log scale)
Page 26 of 41
10
100
15
s 10
%
5
y = 1,06x +
2
R = 0,9798
0
0
10
12
1/concentration
Figure 5: The relationship between measurement uncertainty and the inverted
concentration for Pb in the example (0.1 100 g/L).
The straight-line equation above tells us that the measurement uncertainty equals
1.06 multiplied with 1/concentration plus 1.77. For example, at a concentration of 2
g/L the measurement uncertainty becomes 1.06 * 1/2 + 1.77 = 2.3 %. When
reporting measurement uncertainty to customers, the laboratory can choose
between quoting the formula or calculating the measurement uncertainty for each
value, using the formula. For further reading, see for example /2/.
Page 27 of 41
Cd
g/L
g/L
sr
sR
Synthetic
Lower
33
0.3
0.303
101
3.5 17.0
Synthetic
Higher 34
2.7
2.81
104
1.9 10.7
31
0.572
2.9 14.9
3.07
2.1 10.4
Waste
water
1.00
3.1 27.5
27
Calculate combined
standard uncertainty, uc
uc = sR = 27.5
Calculate expanded
uncertainty, U = 2 u c
U = 2 sR
Measurand
Cd
Page 28 of 41
Combined Uncertainty uc
uc = 27,5 %
Expanded Uncertainty U
2*uc = 55 % 50 %
Variable
pH
Conductivity,
mS/m
Alkalinity
Turbidity, FNU
Ammonium, g/L
Nitrate
Nominal
value
7.64
12.5
Lab %
deviation
-0.037
-2.8
0.673
1.4
146
432
2.3
-9.1
2.2
-1.6
sR
sR
%
0.101
0.40 3.2
No. of
labs
90
86
5
6
0.026
0.1
12.0
16.3
60
44
34
39
3
3
5
6
3.9
14.2
8.8
3.7
Excluded
In Table 4 we find that for conductivity, for instance, the mean value for the results
from 10 interlaboratory comparisons is 12.5 mS/m. The reproducibility relative
standard deviation is 3.2 %, which is an average (or pooled) standard deviation
between the laboratories in the different interlaboratory comparisons, uc=3,2.
If we take the ammonium results, we have a mean nominal value of 146 g/L, and
we find that the reproducibility, sR, is 8.8 %. Thus U = 2*8.8 = 17.6 = 18 % at this
concentration level.
Comment: In Section 3 the expanded uncertainty for ammonium is 5 % using an
automated method in one highly qualified laboratory.
Page 29 of 41
8 Reporting uncertainty
This is an example on what a data report could look like, when measurement
uncertainty has been calculated and is reported together with the data. The
company and accreditation body logotypes are omitted, and the report does not
contain all information normally required for an accredited laboratory. It is
recommended to use either relative or absolute values for the benefit of the
customer.
Analytical Report
Sample identification: P1 P4
Samples received: 14 December 2002
Analysis period: 14 16 December 2002
Results
NH4-N (g/L):
Sample
P1
P2
P3
P4
Result
103
122
12
14
U
5%
5%
10%
10%
Method
23B
23B
23B
23B
TOC (mg/L)
Sample
P1
P2
P3
P4
Result
40
35
10
9
U
4.0
3.5
1.0
0.9
Method
12-3
12-3
12-3
12-3
Signed: Dr Analyst
Page 30 of 41
The laboratory should also prepare a note explaining how the measurement
uncertainty has been calculated for the different parameters. Normally, such an
explanatory note should be communicated to regular customers and other
customers who ask for information. An example is given below:
References:
Guide To The Expression Of Uncertainty In Measurement (GUM)
Quantifying Uncertainty in Analytical Measurement.
EURACHEM/CITAC Guide
Handbook for calculation of measurement uncertainty in
environmental laboratories
Page 31 of 41
9 References
1. Guide To The Expression Of Uncertainty In Measurement (GUM). BIPM,
IEC, IFCC, ISO, IUPAC, IUPAP, OIML. International Organization of
Standardization, Geneva Switzerland, 1st Edition 1993, Corrected and reprinted
1995.
2. Quantifying Uncertainty in Analytical Measurement. EURACHEM/CITAC
Guide, 2nd Edition, 2000
3. Measurement Uncertainty in Testing, Eurolab Technical Report No. 1/2002
4. Interlaboratory comparison test data, personal communication, H. Hovind,
NIVA, Norway.
5. International Vocabulary of Basic and General Terms in Metrology (VIM).
ISO, 1993
6. ISO/IEC 3534-1-2, Statistics Vocabulary and symbols Parts 1-2
7. ISO 5725-1-6:1994, Accuracy (trueness and precision) of measurement
methods and results
8. ISO/DTS 21748:2002, Guide to the use of repeatability, reproducibility and
trueness estimates in measurement uncertainty estimation
9. EN ISO/IEC 17025:2000, General Requirements for the Competence of
Calibration and Testing Laboratories
10. ISO/TR 13530:1997, Water quality Guide to analytical quality control for
water analysis
11. EN ISO 11732:1997, Water quality -- Determination of ammonium nitrogen by
flow analysis (CFA and FIA) and spectrometric detection
12. EA-4/xx: Proposed EA guideline on The Expression of uncertainty in
quantitative testing. Project to be finalised in 2003. (www.europeanaccreditation.org/)
13. ISO8258, First edition, 1991-12-15, Shewhart Control Charts
14. V. J. Barwick, Ellison L.R., Analyst, 1999, 124, 981-990
15. E. Hund, D.L. Massart and J. Smeyers-Verbeke, Operational definitions of
uncertainty. TrAC, 20 (8), 2001
Page 32 of 41
10 Appendices
Action
Specify Measurand
Measurand:
(measurand) in (matrix) by (method) The
customer demand on expanded uncertainty is
_ %.
Quantify u(Rw)
A control sample
B possible steps not
covered by the control
sample
A:
Quantify systematic
components
Convert components to
standard uncertainty u(x)
Calculate combined
standard uncertainty, uc
Calculate expanded
uncertainty, U = 2 u c
B:
Page 33 of 41
Relative u(x)
Comments
Reproducibility within-laboratory, Rw
Control sample
X
sRw
= (conc) (unit)
Other components
Systematic components
Reference material
bias
Interlaboratory
comparison
bias
Recovery test
bias
sR
Standard method
sR
Page 34 of 41
Combined Uncertainty uc
Expanded Uncertainty U
2*uc =
y = m + ( + B) + e
y
method bias
s Rw
where sR2 is the estimated variance under reproducibility conditions and where sL2
is either the estimated variance of B if one method is used by all laboratories or an
estimated variance of B and if several different methods have been used in the
collaborative study and sr2 is the estimated variance of e.
Comment
For samples that are more inhomogeneous and have big variations in matrix the
estimation of the measurement uncertainty of the method can become too low. The
treatment of this item is beyond the scope in this handbook if not included in
control sample or duplicate analyses of real samples.
Page 35 of 41
Exercise
Nominal
value
mg/L
1999 1
2
2000 1
2
2001 1
2
Laboratory
result
mg/L
81
73
264
210
110
140
83
75
269
213
112
144
X
S
u(Cref) =
sR
n
8.8
34
Bias
sR
2.4
2.7
1.9
1.4
1.8
2.9
+ 2.18
0.58
8.8
-
Number
of labs
34
-
= 1.5%
A t-test shows that the bias (+2.18 %) is not significant (t = 0.6). However, in order
not to complicate the calculations when the bias is small, t-test are normally not
performed.
Page 36 of 41
Sam ple
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
X1
7.46
9.01
3.6
6.48
14.49
10.84
4.61
2.6
2.8
5.84
2.12
2.3
2.52
3.71
7.43
8.83
9.12
8.24
2.62
3.33
2.69
12.09
4.24
10.49
3.68
9.37
2.22
6.1
2.96
14.02
4.24
5.1
2.78
8.52
12.82
3.17
11.28
14.31
4.01
3.27
9.98
12.56
3.35
X2
7.25
9.17
3.1
6.48
14.12
9.89
5
2.42
2.62
6.19
2.5
2.11
2.89
3.71
7.43
8.51
8.79
7.9
2.78
3.33
2.69
12.09
4.24
10.64
3.52
9.37
2.06
6.1
2.86
13.7
3.62
4.61
2.62
6.81
14.05
2.4
11.43
13.82
4.48
3.58
10.29
13.66
2.88
M ean:
7.355
9.090
3.350
6.480
14.305
10.365
4.805
2.510
2.710
6.015
2.310
2.205
2.705
3.710
7.430
8.670
8.955
8.070
2.700
3.330
2.690
12.090
4.240
10.565
3.600
9.370
2.140
6.100
2.910
13.860
3.930
4.855
2.700
7.665
13.435
2.785
11.355
14.065
4.245
3.425
10.135
13.110
3.115
6.499
0.210
-0.160
0.500
0.000
0.370
0.950
-0.390
0.180
0.180
-0.350
-0.380
0.190
-0.370
0.000
0.000
0.320
0.330
0.340
-0.160
0.000
0.000
0.000
0.000
-0.150
0.160
0.000
0.160
0.000
0.100
0.320
0.620
0.490
0.160
1.710
-1.230
0.770
-0.150
0.490
-0.470
-0.310
-0.310
-1.100
0.470
s(r) % = range(mean)/1.128 =
2.855
1.760
14.925
0.000
2.587
9.165
8.117
7.171
6.642
5.819
16.450
8.617
13.678
0.000
0.000
3.691
3.685
4.213
5.926
0.000
0.000
0.000
0.000
1.420
4.444
0.000
7.477
0.000
3.436
2.309
15.776
10.093
5.926
22.309
9.155
27.648
1.321
3.484
11.072
9.051
3.059
8.391
15.088
6.4363
5.71
= m ean range (% )
%
Page 37 of 41
X1
Sample
1 37.62
2 16.18
3 28.82
4 4490
5 135.7
6 62.56
7 158.9
8 16540
9 31.26
10 58.49
11 740.5
12 130.3
13 29.35
14 1372
15 36.55
16 22.57
17 34.75
18 92.93
19 40.6
20 80.36
21 15.76
22 78.22
23 48.89
24 17.65
25 36.56
26 51.89
27 197.5
28 70.32
29 29.99
30 31.9
X2
36.85
37.235
16.56
16.370
28.65
28.735
4413
4451.500
124.7
130.200
62.25
62.405
159.2
159.050
16080
16310.000
30.12
30.690
60.11
59.300
796.2
768.350
126.9
128.600
29.19
29.270
1388
1380.000
44.74
40.645
23.37
22.970
33.15
33.950
94.01
93.470
42.23
41.415
86.36
83.360
18.54
17.150
73.76
75.990
50.91
49.900
16.72
17.185
35.3
35.930
52.2
52.045
206.5
202.000
69.22
69.770
30.62
30.305
32.36
32.130
Mean: 816.331
0.770
-0.380
0.170
77.000
11.000
0.310
-0.300
460.000
1.140
-1.620
-55.700
3.400
0.160
-16.000
-8.190
-0.800
1.600
-1.080
-1.630
-6.000
-2.780
4.460
-2.020
0.930
1.260
-0.310
-9.000
1.100
-0.630
-0.460
s(r) % = range(mean)/1.128 =
Page 38 of 41
2.068
2.321
0.592
1.730
8.449
0.497
0.189
2.820
3.715
2.732
7.249
2.644
0.547
1.159
20.150
3.483
4.713
1.155
3.936
7.198
16.210
5.869
4.048
5.412
3.507
0.596
4.455
1.577
2.079
1.432
4.0843
3.62
m e a n
R e s . 2
m g /L
8 .9 1
9 .0 1
8 .9 0
9 .1 2
8 .6 4
8 .5 1
8 .8 1
8 .0 0
7 .0 8
7 .0 1
7 .1 6
6 .7 8
6 .5 3
4 .6 8
5 .3 3
7 .4 0
7 .6 3
5 .8 8
6 .0 6
6 .3 3
5 .9 0
6 .2 7
6 .0 2
9 .1 1
9 .1 4
8 .4 4
8 .7 1
8 .0 9
7 .5 8
6 .3 2
6 .4 4
7 .3 4
7 .3 1
8 .0 3
8 .2 9
9 .2 9
9 .0 8
9 .3 6
9 .3 7
9 .2 5
8 .4 9
8 .2 8
8 .3 1
8 .1 5
8 .0 3
7 .4 0
7 .4 9
4 .4 9
4 .4 4
4 .2 7
m e a n ra n g e :
r a n g e /1 .1 2 8 :
R a n g e
m g /L
0 .0 1
0 .0 2
0 .0 0
0 .0 1
0 .0 4
0 .0 9
0 .0 0
0 .0 2
0 .0 3
0 .0 3
0 .0 3
0 .0 1
0 .0 2
0 .0 0
0 .0 5
0 .0 2
0 .0 1
0 .0 0
0 .0 3
0 .0 0
0 .0 0
0 .0 3
0 .0 0
0 .0 2
0 .0 4
0 .0 6
0 .0 2
0 .0 0
0 .0 2
0 .0 2
0 .0 1
0 .0 9
0 .0 3
0 .0 3
0 .0 9
0 .0 6
0 .0 1
0 .0 1
0 .0 1
0 .0 7
0 .0 2
0 .0 1
0 .0 6
0 .0 6
0 .0 2
0 .0 2
0 .0 0
0 .0 3
0 .0 1
0 .0 2
0 .0 2 6
0 .0 2 4
Page 39 of 41
Page 40 of 41
Res. 1
218.90
206.46
221.18
215.00
194.96
218.65
223.86
215.58
196.26
210.89
228.40
206.73
207.00
224.49
201.09
218.83
216.69
206.36
215.21
Res. 2
Average
214.77
216.84
220.83
213.65
210.18
215.68
206.50
210.75
218.03
206.50
216.55
217.60
212.19
218.03
213.01
214.30
214.93
205.60
206.89
208.89
222.73
225.57
229.03
217.88
208.47
207.74
213.66
219.08
214.07
207.58
223.13
220.98
218.22
217.46
227.96
217.16
226.18
220.70
Average: 214.84
s:
5.58
s%:
2.60
Factor ,
d2
1.128
1.693
2.059
2.326
2.534
2.704
2.847
2.970
3.078
s=
range
d2
For comparison
Rectangular
interval
95 % conf. limit.
3.464
3.92
Page 41 of 41