ASTM F465-76(1981)

Recommended Practice for Developing Precision and Accuracy Data on ASTM Method for the Analysis of Meat and Meat Products (Withdrawn 1993)

General Information

• Status: Withdrawn
• ICS: 67.120.10 - Meat and meat products
• Technical Committee: F10 - Livestock, Meat, and Poultry Evaluation Systems

Standards Content (Sample)

-
ASTM Fqb5 76 M O759530 - 0053882 7
~~~ ~~~ -~ ~~
~
-e- 89-05
Designation: F 465 - 76 (Reapproved 1981)
A
Standard Practice for
DEVELOPING PRECISION AND ACCURACY DATA ON
ASTM METHODS FOR THE
ANALYSIS OF MEAT AND MEAT PRODUCTS‘
‘r
This standard is issued under the fixed designation F 465; the number immediately following the designation indicates the
year of original adoption or. in the case of revision, the year of last revision. A number in parentheses indicates the year of last
reapproval. A superscript epsilon (E) indicates an editorial change since the last revision or reapproval.
1. scope
erage of which will approach zero in a long
series of measurements.
1.1 This practice establishes uniform guide-
3.3 bias-a constant or systematic error as
lines for expressing the precision and accuracy
opposed to a random error. It manifests itself
of methods for the analysis of meat and meat
as a persistent positive or negative deviation
products. It includes a procedure for develop-
of the method average from the accepted ref-
ing this information. There is no intent to re-
erence value.
strict qualified groups in their use of other
3.4 precision -the degree of agreement of
techniques.
f
repeated measurements of the same property.
1.2 Statements of precision are restricted
Precision statements in ASTM methods for
to those variables specifically mentioned.
analysis of meat and meat products will be
Task groups are referred to (1, 2, 3).*
derived from the estimated standard deviation
2. Applicable Documents
of a series of measurements and will be ex-
pressed in terms of the repeatability and re-
2.1 ASTM Standards:
producibility of the method.
E 178 Recommended Practice for Dealing
3.5 accuracy - the agreement between an
with Outlying Observations3
experimentally determined value and the ac-
E 180 Recommended Practice for Devel-
cepted reference value.
oping Precision Data on ASTM Meth-
3.6 variance-a measure of the dispersion
ods for Analysis and Testing of Indus-
of a series of results around their average. It is
trial Chemicals4
the sum of the squares of the individual devia-
F 463 Test for Fat in Meat and Meat Prod-
tions from the average of the results, divided
ucts by Ether Extraction5
by the number of results minus one.
3. Definitions 3.7 standard deviation-a measure of the
dispersion of a series of results around their
3.1 error-in a statistical sense, any devia-
average, expressed as the square root of the
tions of an observed value from the true
quantity obtained by summing the squares of
value. When expressed as a fraction or per-
the deviations from the average of the resuits
centage of the value measured, it is called a
and dividing by the number of observations
relative error. All statements of precision or
accuracy should indicate clearly whether they
are expressed in absolute or relative sense.
i
’ This practice is under the jurisdiction of ASTM Com-
3.2 random error - the chance variation
mittee F- 10 on Meat and Meat Products.
Current edition approved Aug. 27, 1976. Published De-
encountered in all experimental work despite
cember 1916.
. the closest possible control of variables. It is
‘The boldface numbers in parentheses refer to the list of
references at the end of this recommended practice.
characterized by the random occurrence of
‘Annual Book of ASTMSiandards, Pari 41.
both positive and negative deviations from the
‘Annual Book of ASTM Standards, Paris 29 and 30.
mean value for the method, the algebraic av- ‘Annual Book of ASTM Standards, Part 46.

F 465
expected value is the average of an infinite
minus one. It is also the square root of the
series of such determinations.
variance and is calculated as follows:
3.12 95 % conjûience level-a term com-
monly used for establishing the probability of
= PXi n-1 - precision statements and means that there are
95 in 100 chances of being correct, and 5 in
where:
100 chances of being wrong, when predicting
s = estimated standard deviation of the se-
that the expected precision (or expected
ries of results,
value) will fa11 within the specified limits or
Xi = each individual value,
range.
= average (arithmetic mean) of all values,
3 -13 repeatability -the precision of a
and
method expressed as the agreement a-ainable
n = number of values.
between independent determinations per-
3.7.1 The following forms of this equation
formed by a single analyst using the same
are more convenient for computation, espe-
apparatus and techniques (see 5.2.6,7.2, and
cialiy when using a calculator:
7.2.12).
3.14 reproducibility - the precision of a
method expressed as the agreement attainable
between determinations that are performed in
different laboratories.
= PX n(n - 1)
4. Preliminary Studies
where:
4.1 General-This section covers the pre-
S = estimated standard deviation,
liminary work that should be carried out in a
W = sum of the squares of all of the indi-
few laboratories before undertaking a full in-
vidual values,
terlaboratory evaluation of a method.
(SQ2 = square of the total of the individual
4.2 When a task group is asked to provide
values, and
a specific analytical procedure, there may be
n = number of values.
one or more methods available from the iiter-
NOTE 1 -Care must be taken in using either of
ature or from laboratories already performing
these equations to be sure that a sufficient number
such analyses. In such cases, these methods
of decimal places is carried in the sum of the values
have usually been the subject of considerable
and in the sum of their squares so that serious
rounding errors do not occur. For best resuIts, ali
research, therefore any additional study of
rounding should be postponed until after a value
variables, at this stage, would be a waste of
has been obtained fors. In this recommended prac-
time. It is recommended that such methods be
tice, the standard deviation is obtained from an
analysis of variance of the results of an interlabora-
ASTM format, with full descrip
rewritten in
tory test program (see Section 7).
tions of the equipment and procedure, and be
evaluated in a pilot run by a few laboratories
3.8 coefficient of variation-a measure of
relative precision calculated as the standard on selected materials. Three laboratories and
at least three such materials, using one or two
deviation of a series of values divided by their
average. It is usually multiplied by 100 and analysts performing duplicate determinations
expressed as a percentage. on each of two days, by each method, consti-
tutes a practical plan that can be analyzed by
3.9 range - the absolute value of the alge-
in Sections 6,7, and
the procedures described
braic difference between the highest and the
8. Such a pilot study will confirm the ade-
lowest values in a se€ of data.
quacy of the methods and supply qualitative
3.10 duplicates -paired determinations
indications of relative precision and accuracy.
performed by one analyst at essentially the
4.3 When the method to be evaluated is
same time. This concept also applies to other
new, or represents an extensive modification
such multiple determinations.
of an available method, it is recommended
3.11 95 % confidence interval or confi-
that a study on variables be carried out by at
dence limits-that interval or range of values
around an observed value which will, in 95 % least one laboratory to establish the parame-
ters and conditions to be used in the descrip
of the Eases, include the expected value. The

ASTM F4b5 7b 0757510 0051884 O a
F 465
of independent ,laboratories cannot be re-
tion of the method. This should be followed
cruited, advantage can be taken of a liberal-
by a three-laboratory pilot study before un-
ized definition of collaborating laboratories,
dertaking a full interlaboratory evaluation.
quoted as follows from p. 9 of STP 335 (3):
4.4 Detailed procedures for executing such
preliminary studies are not described in this
Here the term ?collaborating laboratory? has a
more specific meaning than in common usage. For
recommended practice but are available in the
example, a testing process often consists of an inte-
generak?statìstical literature (4).
grated sequence of operations using apparatus, re-
agents, and measuring instruments; and several
5. Planning the Interlaboratory Study more or less independent instaliations may be set up
in the. same area or ?laboratory.? Each such partic-
5.1 General-This section covers the rec-
ipating installation should be considered as a collab-
orating laboratory so far as this procedure is con-
ommendations for the planning of interlabo-
cerned. Similarly, sets of test results obtained with
ratory studies.
different participants or under different conditions
5.2 Variables-The major variables to be
of calibration would in general constitute results
from different collaborating laboratories even
considered are methods, materials or levels,
though they were obtained OIT the Same sets of
laboratories, apparatus, analysts, days, and
equipments.
runs as follows:
This concept makes it possible tQ increase fhe
5.2,1 Methods -The preliminary studies of
available ?laboratorie~? by using two analysts
Section 4 should lead to an agreement on a
(but not more than two) in as many laborato-
single method, which can then be evaluated in
ries as needed to bring the total to the recom-
a full interlaboratory study. If it is necessary
mended minimum of ten. In such cases, the
to evaluate two or more methods, the com-
two analysts must evaluate the method inde-
plete program must be carried out on each
pendently in the fullest sense of the word,
method. In either case, it will be assumed that
interpret as using different sampIes, different
the variables for each method have been ex-
reagents, different apparatus where possible,
plored and that a well-standardized, fully de-
and perform the work on diaerent calendar
tailed procedure has been prepared. The time
days. (In the design in 5.8, laboratories using
and expense required for an extensive preci-
two analysts are designated as A-1, A-2, B-1,
sion study cannot be justified if the prepara-
B-2, etc.). The most desirable laboratories
tion is incomplete.
and analysts are those having previous experi-
5.2.2 Materials or Levels-The number of
ence with the proposed method or with similar
samples distributed should be held to the min-
methods. It is essential that enough experi-
imum needed to evaluate the method ade-
ence be acquired to establish confidencein the
quately, (Increasing the number of samples
will. not significantly increase the degrees of performance of a laboratory before starting
the interlaboratory test series. Such prelimi-
freedom available for predicting the repro-
nary work must be done with samples other
ducibility of the method. This-can be achieved
than those to be used in the formal interlabo-
only by increasing the number of laborato-
ratory test program.
ries.) Some interlaboratory studies can be lim-
5.2.4 Apparatus-The effect of duplicate
ited to a single sample, as in the case of pre-
setups is not often a critical variable in chemi-
paring a specific standard solution. Methods
cal analysis. In instrumental methods, how-
applicable to a single product of high purity
can usually be evaluated with one or two sam- ever, apparatus can become an important fac-
for because the various laboratories may be
ples. When different concentrations of a con-
using different makes or types of equipment,
stituent or values of a chemical property are
involved, the samples should represent the for example, the various colorimeters and
spectrophotometers used in photometric
approximate lower, middle, and top levels of
the expected range. If these vary over a wide methods. In such cases, the effect of appara-
tus becomes confounded with between-labo-
range, the number of levels should be in-
creased and spaced to cover the range. ratory variability, and special care must be
used to avoid misinterpreting the results. Of
5.2.3 Laboratories -To obtain a reliable
precision estimate, it is recommended that the course, if- enough laboratories have instru-
ments of each type, the apparatus can be
inferlaboratory study include approximately
made a planned variable in the study.
ten qualified laboratories. When this number
.
\, 124 .
ASTM FL165 76 0757530 005Löä5 2
F 465
4CTb
and should immediately request a replace-
5.2.5 Analysts-The use of a single analyst
ment.
in each laboratory (as described in 5.2.3) is
5.4.2 The most important requirement is
adequate to provide the information needed
that the subsamples be represenfative and as
for calculating the repeatability and reproduc-
homogeneous as possible. This is to be em-
ibility of the method as covered in this recom-
phasized in preparing samples involving the
mended practice. It is essential that all ana-
comparison of a test with a reference method.
lysts complete the entire interlaboratory test
If collaborating laboratories are involved in
program. With regard to analyst qualifica-
evaluating a single method, a high degree of
tions, an analyst who is proficient in the
homogeneity should be assured. In such cases,
method should be selected.
it is recommended that the person responsible
5.2.6 Days -The repeatability of the
for preparing and distributing samples con-
method (see 3.13) shall be evaluated in terms
duct analyses on representative lots to deter-
of independent determinations by the same
analyst. To achieve this, all scheduled deter- mine that the sampling error is acceptably
low.
minations must be performed on each of two
5.4.3 Instability of any type may impose
days (see 5.8 and 7.2).
other restrictions on the execution of a
NOTE 2-As used in this recommended practice,
the term “days” represents replication of a set of planned program. It is the responsibility of the
determinations performed on any day other than
task group chairman to incIude in the plans for
that on which the first set was run. It may become a
the interlaboratory study specific instructions
s stematic variable to the extent that it 1s desirable
tgat a given laboratory run the entire set on an- on selecting, preparing, storing, and handling
other. Although this may introduce a bias for that
of the standard samples.
laboratory, there appears to be little chance that
5.4.4 The samples distributed for the for-
such a bias would be common to all laboratories.
mal interlaboratory test program should not
When preiminary studies suggest that instability
may result in an overall systematic “days” effect
...