8466-1
INTERNATIONAL
STANDARD
First edition
1990-03-01
and evaluation
Water quality - Calibration
analytical
methods and estimation
of
Performance
characteristics
Part 1:
Statistical
Qualith de lea
des caractkes
Partie 1: haha
evaluation
rEtalonnage
de Performance
tion s ta tistique
des mhthodes
of
function
danalyse et estimation
lin&aire dh talonnage
Reference number
ISO 8466-1 : 1990 (EI
Contents
Page
Foreword
.. .
Ill
............................................................
Scope ..............................................................
Definitions..
Symbols.........................................................~
Performance
........................................................
..
........................................................
4.1
Choice of working
4.2
Calibration
4.3
Assessment
range ..........................................
and characteristics
of the method
........................
range ..........................................
5.2
Calibration
and characteristics
5.3
Evaluation
......................................................
Annex
Bibliography
of the method
.................................................
Example ............................................................
Choice of working
....................................................
5.1
........................
6
7
7
8
0
ISO 1990
All rights reserved. No patt of this publication may be reproduced or utilized in any form or by any
means, electronie or mechanical, including photocopying and microfilm, without Permission in
writing from the publisher.
International Organization for Standardization
Case postale 56 l CH-1211 Geneve 20 l Switzerland
Printed in Switzerland
Licensed to LAAM /MANUEL SOARES
Foreword
ISO (the International
Organization for Standardization)
is a worldwide
federation of
national Standards bodies (ISO member bedies). The work of preparing International
Standards is normally carried out through ISO technical committees.
Esch member
body interested in a subject for which a technical committee has been established has
the right to be represented
on that committee.
International
organizations,
governmental and non-governmental,
in liaison with ISO, also take patt in the work. ISO
collaborates
closely with the International
Electrotechnical
Commission
(IEC) on all
matters of electrotechnical
standardization.
Draft International
Standards adopted by the technical committees
are circulated to
the member bodies for approval before their acceptance as International
Standards by
the ISO Council. They are approved in accordance with ISO procedures requiring at
least 75 % approval by the member bodies voting.
International
Standard
Water quaiity.
by Technical
Committee
Part 7: Statisticaf
evaluation
Part 2: Calibration
strategy
Part 3: Method
- Part 4: Estimation
basis method.
Annex
of Standard
calibration
ISO/TC
147,
function
functions
addition
of Limit of detection
only.
of an analytical
INTERNATIONAL
STANDARD
and evaluation
of analytical
Water quality - Calibration
methods and estimation
of Performance
characteristics
Part 1:
Statistical
1
evaluation
Scope
function
Originai
MeasUri!g
Sample
;X~~nstructi;
measuring instructions
Measur;
value
r-YG&q
Definitions
2.1
method:
analytical
of procedural,
measuring,
tions (see figure 1).
defini-
method
23
2.2
calibrating
instruction : Describes the approach to
determine the calibration function from information
values, yi,
obtained by measuring given Standard concentrations,
xi. The
slope of the calibration function, b, as a measure of sensitivity
of the analytical method and the Standard deviation of the
which
method, sXO, are figures of merit and characteristics
result from the calibration experiment.
of indepen-
evaluating
instruction : A calculation
guide for the
computation
of concentrations
from the measured values by
the use of the calibration function. Additionally,
the confidence
range permits an objective assessment of the imprecision of the
analytical result[*J.
2.4
measured
of a measuring
initial
System.
25
Yi
Information
value of the Standard concentration Xi calculated from the calibration function.
s2
i
Variante
of the information
analyses of Standard samples,
centration Xi.
fi
2.7 coefficient
Sensitivity
of the method (slope of the calibration line; coefficient
of regression
of the method
sXO: The ratio of
the residual Standard deviation, s,,, to the sensitivity of the
calibration
function,
b. lt is a figure of merit for the performance of the analytical method, and is valid within the working
range (sec equation 13).
2.8 working
2.9
homogeneity
of variances: Homogeneity of variances
of pooled data, such as those resulting from replicate analyses
at different
levels, is confirmed
if these variances are not
significantly
correlated to their appertaining
concentrations.
2.10
sensitivity
of the analytical
The slope of
analytical method, inthe working
range in
Symbols
Xi
Concentration
Subscript
i=
N
the
concentration
levels,
where
by linear
sY2
obtained
by non-
DS2
Differente
Information
Number
Sample.
Mean of information
replicates.
Concentration
sf the
analytical
Sample,
calculated from the i nformation
value Ym
Concentration
of the
analytical
Sample,
calculated from the mean of the information
values 7.
of variances.
value of an analysed
of replicates
values,
resulting
from
XI0
Concentration
of the Standard Sample at the
upper level of the working range (10th Standard
Sample).
Yi,j
jth information
Subscript
Tabled
l-a)
N - 2 degrees of freedom
FV;,f2,1
-a)
Xi.
%O
Standard
deviation
of the method.
VXO
Coefficient
of Variation
VB (x,
Confidence
2, . . . . ni.
of replicates
Sample.
Number of concentration
ISO 8466, N = IO).
Number
deviation.
obtained
1,2, . . . . N.
Xl
of the cali-
of the
sY1
method:
2.11
Residual Standard
sY
of the method.
.?.
Performance
4.1
Choice
4.1.2
of working
a)
range depends
the practice-related
with
the choice
10
of a
(Y i t j - El*
t:
j=l
S*i =
on
objective
of the variances
range
Esch calibration
experiment
is started
preliminary working rangeL3].
The working
Ili -1
. . .
(11
. . .
(2)
differentes
at
of the calibration.
with the mean
The working
range shall cover, as far as possible, the
application range for water, waste water, and sludge analysis. The most frequently
expected Sample concentration
should lie in the centre of the working range.
10
Y-,J *
c
j=l
.-
Yj
fori
= 1 ori
= 10
yli
b)
feasibilities
of technical
realizability.
PG = 2
PG
forsyo
is
determined
for
the
> sf
F-test
from
. . (3)
1
S*
PG = --$
10
The independence
linearity16t *l.
is verified
by a statistical
test on the
PG is compared
Decision :
4.1.1
Reparation
of the calibration
a)
After establishing
the preliminary
working
range, measured
values of at least five (recommended
N = 10) Standard
samples are determined.
The concentrations,
xj, of these standard samples shall be distributed equidistantly
over the working
range. In Order to check for the homogeneity
of the variances,
ten replicates of each of the lowest and the highest concentrations (x, and x,~) of the working range are determined. Ten information values, )pi jl result from these series of measurements
(see table 1).
Table 1 -
If PG
<
the
Ffl. I f2. I 099
I
differente
between
the
between
the
b) If PG > q, ; j-2;
0,99 the
differente
lSO,8466-1 : 199OE)
4.13
# *]
. Table 2 -
5
B
c
06
I
.-0
051
04
03
+E
8
z-
xj
-vj
x;
lI
I
linear regression
-V2i
Xjuj
--y-----------l----g:ppIIj
1
--
----
.,
----t---l----l
--L----1 t------t-m
I
6
7
,)c
.-----j----~----~--q-----j
5
~--
--L------
-t-
.--~~
--_.-
-----+------
---
12345-6789
Concentration
Figure 2 - Graphical
linearity
check
The differente
of the variances
(4) ..
OS = (N Degrees
: f = 1.
of the non-linear
calibration
function
from equation
. . . (4)
of freedom
are submitted
ferences.
OS* is calculated
PG required
OS*
is calculated
sv2
dif-
from
. . . (5)
S*
Y*
The ten data Sets, consisting of the va2ues of Xi and Yj, are submitted to a linear regression analysis to obtain the coefficients a
and b of the calibration function whsich describe the linear correlation
between
the concentration
x as an independent
variable, and the measured value y as a dependent variable.
The calibration function as well as the characteristics
of the
method should result from data obtained from a working range
x1 to xIo, as received from the measurement and not corrected
for blanks. Generally, no blank value (concentration
x = 0) is
to be included in the calibration experiment and, consequently,
in the least-squares fit of the regression.
The linear calibration
(6)
. . .
Y =U+bX
(6)
N
c
Calibration
is given by equation
Decision :
4.2
function
and characteristics
of the method
(Xj
XI
'
(Yj
i=l
b = p--_Ip--m_-
Y)
. . . (7)
(x; - XI*
Y
M
.
I=
a = r--b%
. . .
(8)
[Y j - (a + bx,)l*
i=l
(9)
N-2
4.3
The confidence
intervals for analytical results, calculated from
the calibration function, are given by the intersections with the
respective hyperbolic paths in figure 3. The estimation of the
confidence intervals are given by equation (12) [71
Assessment
The concentration
a)
of an analyzed
valuey,
Sample is obtained
7,,* = % + VB (.%,
to give ?
-a
. . .
.a
Y -a
.2,,2 =
b
(IO)
(jf- Fl2
or
b)
7, petformed
on
Sample, to give ?
g=-
(Xi -
b*x
.
Z=
7-a
. . .
(11)
(12)
NOTE - If i; = 1, F,,* =
As to the uncertainty
of an analytical result, keep in mind that
the analytical error is a combination
of the uncertainty
of the
determination
of the measured value, and the uncertainty
of
the estimation of the regression coeff icients[*!
Equation
brackets
statistical
VB(g) is
X1,2-
7
0
a /
1
I
I
I
- 44
vl3m
I
I
I
I
I
I
I
I
I
I
X
10
Working range
Figure 3 - Working
range x, to xlo, calibration line with confidence band and a Single analytical
appertaining
confidence interval
Licensed to LAAM /MANUEL SOARES
ISO Store order #:913756/Downloaded:2008-05-21
Single user licence only, copying and networking prohibited
5.1
. . .
of working
range
sXO =2
Choice
5.1.1
Testing
the hornogeneity
v xo = s~xloo
(13)
analytical
expressed
(3)
= 2,9
4,67 x 10-6
F(9,9;
0,991 = 5,35
The photometric
determination
of nitrite is used to demonstrate
the calibration and the subsequent estimation of the statistical
characteristics
of the method and their influence on the final
results of the evaluation.
xi
Y*l,l
Yi,2
Yi,3
Y-1,4
Y-lt5
0,05
0,140
0,143
0,143
0,146
0,lO
0,281
0,15
0,405
0,20
0,535
0,25
0,662
0,30
0,789
0,35
0,916
0,40
1,058
0,45
1,173
0,50
1,303
1,302
1,300
1,304
10 =N
from equation
(14)
Example
is calculated
13,54 x IO-6
S2
PG=s=
. . .
sf the variancesl)
According
to the approach outlined in 4.1 .l, the variances ~2
of the information
values obtained from the Standard concen:
trations at the lower or upper limit of the working range respectively, were determined (sec table 4).
a working
of NO;
Yfj
i,
Y-1,7
i,
0,144
0,145
0,344
'
1,300
1,296
1,295
R,a
&,!3
Yi, 10
0,146
0,145
0,148
1,301
1,296
1,306
1
j
I
/
I
)
I
1
I
1
1
Table 4 -
sf variance.
Object:
nitrite
i
1
I
10
xi
Y.
Yi,2
Yi,3
Yi, 10
Ext.*)
Ext.*)
Ext.*)
Ext. *)
Y*
EX:, )
Yi, 9
Ex%
Y*
Ex;.:)
yi,8
mgll
Ext.,
Ext. *)
Ext. *)
mg& 12
0,05
0,140
0,143
0,143
0,146
0,144
0,145
0,144
0,146
0,145
0,148
4,67. IO-6
0,50
1 1,303
1,302
1 1,300
1 1,304
1 1,300
1,295
1 1,301
1 1,296
1 1,306
Yi, 5
yi,6
1,296
s?
11356 . lW6 1
5.1.2
Testing
A non-linear
linearity
regression
The Standard
equation ( 13)
function
may be derived
0,005
sy
b2,575
5.2
Calibration
xo
5.3
0,002o
as a
= 0,73
0,275
Single determination
0,641 - 0,018
%,2
+
is calculated
x 100
expressed
Evaluation
5.3.1
of the method
from
= 0,002 0
Sxo
V xo = x x loo=
procedure)
and characteristics
is calculated
= 0,005 2 mg/l.
sY*
As both residual Standard deviations
of the variances DS* [see equation
calculated. The non-linear calibration
a significantly better adjustment, e.g.
linear.
method
The coefficient
of Variation of the method,
percentage,
is given by equation (14)
of the
[*l as
Y = 0,013 5 + 2,62 x - 0,081 8 ~2, giving a residual standard deviation of sY2 = 0,005 2 mg/l.
sYl
deviation
2,575
0,002 0 x 2,31
+ -
from
(2,575)*
x 0,206 25
b =
(Xi
33
(yi
i=l
jd
= 2,575
tan be expected
with a confidence
(X i-
c
i=l
32
5.3.2
a, (calculated
blank) is calculated
from
Replicate
Calculation
The residual
equation (9).
Standard
deviation,
sY, is
calculated
of the analytical
0,635
%,2
from
analysis
- 0,018
2,575
(0,635 - 0,726 2)*
(2,575)*
= 0,005 2 [Ext.l
N-2
The equation
Y = 0,018
gave
result:
(Y i - yi)*
sy =
method
x 0,206 25
(4)
+ 2,575 2 x
Table 5 -
tan be expected
with a confidence
1 1 1 $i 1 ,66,
X = 0,275 mg /I
0,20
0,535
7 = 0,726 (Extinction)
0,30
0,789
0,35
0,916
0,40
1,058
dV*1
mg0
T--t-/
r
9
0,45
1,173
10
0,50
1,303
--
xi
c I 2f75l
i=l
7,262
I
l
Annex A
(informative)
Bibliography
111 VONDERHEID, C., DAMMAN, V. , DRR, w., FUNK, W. and KRUTZ, H., Statistical
assessment
and comparison
of analytical
procedures.
of quantitative
ZEEW,
analytical
procedures,
BI
for statistkaI
2nd Edition,
in Chemical analysis,
a Ienergie
atomique,
water,
series -
Presentation
of a Computer
program,
(1966), Leipzig.
quality,
statistical analysis.
evaluation, 3rd Edition, Springer Vdag, (1971), Berlin, Heidelberg, New York.
statistics,
method
New York.
BI
for the
Heidelberg,
characteristics
151 GRAF, U., HENNING, H.J. and STANGE, K., Formulae and tables of mathematical
Berlin,
a nd Performance
methods
to standardization,
data, lnterscience
An approach