6 ~U 117

Desig afita~g:1p~i,~STM International,100 B,arrHarborDriVf

cmp.w., ~7 West Conshohocken, Pennsylvania 19428·295~
An American National Standard

INTERNATIONAL~.LL_ ,
INTVfNArIONAL UnitedStates of America,Thiscopyhas been mac
Stand~rds Worldwide

Stan 8'rM N~~~~M and Quality(STAMEQ)under licen:

Con~tmtiPl!Jteafinlnterlaboratory Study to Determine the
Precision of a Test Method'
This standard is issued under the fixed designation E691; 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.

This standard has been approved Ior use by agencies of the u.s. Department of Defense .

.. 1. Scope 2. Referenced Documents

1.1 This practice describes the techniques for planning, 2.1 ASTM Standards+
conducting, analyzing, and treating the results of an interlabo- E29 Practice for Using Significant Digits in Test Data to
ratory study (lLS) of a test method. The statistical techniques Determine Conformance with Specifications
described in this practice provide adequate information for EI77 Practice for Use of the Terms Precision and Bias in
formulating the precision statement of a test method. ASTM Test Methods
1.2 This practice does not concern itself with the develop- E456 Terminology Relating to Quality and Statistics
ment of test methods but rather with gathering the information E 1169 Practice for Conducting Ruggedness Tests
needed for a test method precision statement after the devel- E 1402 Guide for Sampling Design
E2282 Guide for Defining the Test Result of a Test Method
opment stage has been successfully completed. The data
obtained in the interlaboratory study may indicate, however, 3. Terminology
that further effort is needed to improve the test method.
3.1 Definitions-Terminology E456 provides a more exten-
1.3 Since the primary purpose of this practice is the devel- sive list of terms in Ell standards.
opment of the information needed for a precision statement, the 3.1.1 accuracy, n-the closeness of agreement between a
experimental design in this practice may not be optimum for test result and an accepted reference value. E177
evaluating materials, apparatus, or individual laboratories.
3.1,2 bias, n-the difference, between the expectation of the
104 Field of Application-This practice is concerned exclu- test results and an accepted reference value. E177
sively with test methods which yield a single numerical figure
3.1.3 interlaboratory study, (ILS) in ASTM, n-a designed
as the test result, although the single figure may be the outcome
procedure for obtaining a precision statement for a test method,
of a calculation from a set of measurements.
involving multiple laboratories, each generating replicate test
104.1 This practice does not cover methods in which the
results on one or more materials.
measurement is a categorization; however, for many practical
purposes categorical outcomes can be scored, such as zero-one 3.1.4 observation, n-the process of obtaining information
scoring for binary measurements or as integers, ranks for regarding the presence or absence of an attribute of a test
example, for well-ordered categories and then the test result specimen, or of making a reading on a characteristic or
can be defined as an average, or other summary statistic, of dimension of a test specimen. E2282
several individual scores. 3.1.5 precision, n-the closeness of agreements between
1.5 This standard may involve hazardous materials, independent test results obtained under stipulated conditions.
operations, and equipment. This standard does not purport to EI77
address all of the safety problems associated with its use. It is 3.1.6 repeatability, Il-precision under repeatability
the responsibility of the user of this standard to establish conditions. EI77
appropriate safety and health practices and determine the 3.1.7 repeatability conditions, n-conditions where inde-
applicability of regulatory limitations prior to use. pendent test results are obtained with the same method on
identical test items in the same laboratory by the same operator
using the same equipment within short intervals of time. EI77
I This practice is under the jurisdiction of ASTM Committee Ell on Quality and

Statistics and is the direct responsibility of Subcommittee EI1.20 on Test Method

Evaluation and Quality Control. , For referenced ASTM standards, visit the ASTM website, www.astll1.org, or
Current edition approved Oct. I, 2016. Published October 2016. Originally contact ASTM Customer Service at service@astm.org. For Annual Book oj'ASTM
approved in 1979. Last previous edition approved in 2015 as E691 - IS. 001: Standards volume information, refer to the standard's Document Summary page on
10.1520/E0691·16. the ASTM website.

Copyright © ASTM International. 100 Barr Harbor Drive, PO Box C700. West Conshohocken, PA 19428·2959. United States
 a •UI J
7
CopyrightASTMInternational, 100 Barr Harbor Driv;
West Conshohocken, Pennsylvania 19428-295S
INlVtNA ""HAL U.
Standa,", WO"'wid.nlted States of America, This copy has been mad.
dOT~ by the Directorate for Standards and Quality (STAMEQ)under lim.
<!ijHf1 E1~~~S'fi\11rirnational.
3.1.8 repeatability limit ~,~-the value below which the laboratories tees rm ponent may be
absolute difference betwee t. ~o individual test results obtained close to zero or be negative. In the latter case, the estimate is
under repeatability conditi ay be expected to occur with a set to zero. (See Note 2 and Xl.1.2).
probability of approximat y I .p5 (95 %). El77 3.2.5 cell, n-the intersection of a row and column in a
3.1.9 repeatability stan ill Heviation, (s.), n-the standard two-way classification table, in which the rows represent the
deviation of test result obt n~d under repeatability conditions. laboratories and the columns represent the materials.
E177 3.2.5.1 Discussion-The table holds the test results from an
II
. 3.1.10 reproducibility, -IIPulecision under reproducibility
interlaboratory study, and each cell contains the test results
from a particular laboratory on a particular material (see
conditions. E177
Section 7 and Table 1).
. 3.1.11 reproducibility c n~itfons, n-conditions where test
3.2.6 cell average, .t, n-the average of the test results in a
results are obtained with hJ same method on identical test
particular cell.
items in different laborat ri' I[s~I ith different operators using
different equipment. E177 3.2.7 cell deviation, d, n-the cell average minus the aver-

3.1.12 reproducibility I If (f)'

n-the value below which
the absolute difference bet ebn two test results obtained under
age of the cell averages.
3.2.8 cell standard deviation, s, n-the standard deviation of
reproducibility conditions !1y be expected to occur with a the test results in a particular cell.
probability of approximat y iO.r5 (95 %). E177 3.2.9 repeatability variance, s/, n-the sample variance of
3.1.13 reproducibility s Ard deviation (SR)' n-the stan- test results obtained under repeatability conditions.

~~~~i~~~is~tion of test res 11@l!btained under reproduci~i~~~ 3.2.9.1 Discussion-This statistic is estimated for a material
as the pooled within-laboratory variances over all of the
3.1.14 ruggedness test, Ta planned experiment in which laboratories in the ILS.
environmental factors or t t co ditions are deliberately varied 3.2.10 reproducibilityvariance, s/, n-the sample variance
in order to evaluate the e .ts bf such variation. E1169 of test results obtained under reproducibility conditions.
3.1.15 test determinatio , I ithe value of a characteristic or 3.2.10.1 Discussion-This statistic is estimated as the sum
dimension of a single test cr' en derived from one or more of the two variance components due to between-laboratories,
observed values. I E2282 s/, and within-laboratories, s/.
3.1.16 test method, n- e itive procedure that produces 3.2.11 standard deviation of the cell averages, SX' n-the
a test result. I E2282 standard deviation of the cell averages for a particular material.
3.1.1 7 test observation, nTs e observation. E2282 3.2.12 variance of the cell averages, s/, n-the sample
3.1.18 test result, n-t . lal[~e of a characteristic obtained variance of the cell averages for a particular material.
by carrying out a specifie test method. E2282 3.2.13 within-laboratory consistency statistic, k, n-the ra-
3.1.19 test specimen, n t lefortion of a test unit needed to tio of the cell standard deviation to the repeatability standard
obtain a single test deter n Iti n. E2282 deviation.
3.2.13.1 Discussion-This statistic is an indicator of how
3.1.20 test unit, n-the It 1 Iuantity of material (containing
one laboratory's cell standard deviation under repeatability
one or more test specime ) [n eded to obtain a test result as conditions compares with the repeatability standard deviation
specified in the test meth I;s~ test result. E2282 estimated from all laboratories for a particular material (see
3.2 Definitions of Term ,I e ijic to This Standard: XI.2.3).
3.2.1 average of the ce ~e ages, X, n-the average of the
cell averages for a particu a I
aterial. 4. Significance and Use
3.2.2 between-Iaborato c nsistency statistic, h, n-the 4.1 ASTM regulations require precision statements in all
ratio of the cell deviation I th. standard deviation of the cell test methods in terms of repeatability and reproducibility. This
averages. I practice may be used in obtaining the needed information as
3.2.2.1 Discussion-Th sta istic is an indicator of how one simply as possible. This information may then be used to
laboratory's cell average I Iares with the average of the prepare a precision statement in accordance with Practice
other laboratories for a p . ,ul material (see X 1.2.2). EI77. Knowledge of the test method precision is useful in
3.2.3 between-Iaborato istf,ndard deviation, sLJ n-the commerce and in technical work when comparing test results
sample standard deviatio tt~butable to differences of test against standard values (such as specification limits) or be-
result means among labor 1 s. h tween data sources (different laboratories, instruments, etc.).
4.1.1 When a test method is applied to a large number of
3.2.4 between-laborato la iance, SL2, n-the sample vari-
portions of a material that are as nearly alike as possible, the
ance component attributab !orifferences of test result means
test results obtained will not all have the same value. A
among laboratories. I measure of the degree of agreement among these test results
3.2.4.1 Discussion-Th st tistic is estimated indirectly describes the precision of the test method for that material.
from the variance of ce I arerages and the repeatability Numerical measures of the variability between such test results
variance. In situations wh. e there is good agreement among provide inverse measures of the precision of the test method.

2
 I
I .~UT~COPYright A5TM International, tOOBarr Har~or Drive
!4Jf4t~!~
~s16onshohocken" Pennsylvania 19428-295~
. .. . . . I s;;,;;;;:;j~ United States of AmerICa, This copy has been mad·
Greater variability implies smaller (that IS, poorer) prebcls~onD' Dat'l_(?911~lstency, , Section
and larger imprecision. j Y tne Irectoratf:IIIljgll&g1WEd8l$lo~lluffi~,~6iAMEQ) under licens 17
... d d d d . . j I> ffI'~ _1!1.ves~igation 18
4.1.2 Precision IS reporte as a stan ar eVlatlOnji't'sm'l\~ M InterflllllOOillroupActions 19
cient of variation (relative standard deviation), variance, or a Glucose I' S Consistency 20
precision limit (a data range indicating no statistically signifi-
Precision Statement Information Section
cant difference between test results). Repeatability and Reproducibility 21
4.1.3 This practice is designed only to estimate the precision
of a test method. However, when accepted reference values are Section
Keywords 22
available for the property levels, the test result data obtained
according to this practice may be used in estimating the bias of Tables Table
the test method. For a discussion of bias estimation and the Glucose in Serum Example 1-4,6-8
Critical Values of Consistency Statistics, hand k 5
relationships between precision, bias, and accuracy, see Prac-
tice EI77. figures Figure
..A
1 Glucose in Serum Example 1-3
4.2 The procedures presented in this practice consist of
three basic steps: planning the interlaboratory study, guiding Appendixes Appendix
Theoretical Considerations Appendix Xl
the testing phase of the study, and analyzing the test result data. Pentosans in Pulp Example Appendix X2
4.2.1 The planning phase includes forming the ILS. task Spreadsheet for E691 Calculations Appendix X3
group, the study design, selection and number of participating
laboratories, selection of test materials, and writing the ILS 5. Concepts of Test Method 'Precision
protocol. A well-developed test method, including a rugged-
5.1 Repeatability and Reproducibility-These two terms
ness test to determine control of test method conditions, is
deal with the variability of test results obtained under specified
essential.
laboratory conditions and represent the two extremes of test
NOTE I-In this practice, the term test method is used both for the actual method precision. Repeatability concerns the variability be-
measurement process and for the written description of the process, while tween independent test results obtained within a single labo-
the term protocol is used for the directions given to the laboratories for
conducting the ILS,
ratory in the shortest practical period of time by a single
operator with a specific set of test apparatus using test
4.2.2 The testing phase includes material preparation and specimens (or test units) taken at random from a single quantity
distribution, liaison with the participating laboratories, and of homogeneous material obtained or prepared for the ILS.
handling of test result data received from the laboratories. Reproducibility deals with the variability between single test
4.2.3 The data analysis utilizes tabular, graphical, and sta- results obtained in different laboratories, each of which has
tistical diagnostic tools for evaluating the consistency of the applied the test method to test specimens (or test units) taken
data so that unusual values may be detected and investigated, at random from a single quantity of homogeneous materiai
and also includes the calculation of the numerical measures of obtained or prepared for the ILS.
precision of the test method pertaining to repeatability and
5.1.1 Repeatability Conditions= The single-operator,
reproducibility.
single-set-of-apparatusrequirement means that for a particular
4.3 The information in this practice is arranged as follows: step in the measurement process the same combination of
Section operator and apparatus is used for every test result and on every
Scope 1 material. Thus, one operator may prepare the test specimens, a
Referenced Documents 2
Terminology 3 second measure the dimensions and a third measure the
Significance and Use 4 breaking force. "Shortest practical period of time" means that
Concepts of Test Method Precision 5 the test results, at least for one material, are obtained in a time
Planning the Interlaboratory Study (ILS) Section not less than in normal testing and not so long as to permit
ILS Membership 6 significantchanges in test material, equipment or environment.
Basic Design 7
Test Method 8
5.1.2 Reproducibility Conditions-The factors that contrib-
Laboratories 9 ute to variability in a single laboratory, such as operator,
Materials 10 equipment used, calibration of the equipment, and environment
Number of Test Results per Material 11
Protocol 12
(for example, temperature, humidity, air pollution) will gener-
ally have different effects in other laboratories, and the vari-
Conducting the TestingPhase of the ILS Section ability among laboratories will be greater.
-, Pilot Run 13
Full Scale Run 14 5.2 Observations, Test Determinations and Test Results-A
Calculation and Display of Statistics Section test method often has three distinct stages: the direct observa-
Calculation of the Statistics 15 tion of dimensions or properties, the arithmetic combination of
Tabular and Graphical Display of Statistics 16 the observed values to obtain a test determination, and the
arithmetic combination of a number of test determinations to
obtain the test result of the test method.
5.2.1 In the simplest of test methods a single direct obser-
vation is both the test determination and the test result. For

3
 II Ofi,p CopyrightASTMInternational, 100 Barr HarDorDriv;
~J~ West Conshohocken, Pennsylvania 19428·2959
~=:Wr;d:~ United States of America,This copy has been mad'
by the Directorate for Standardsand Quality (STAMEQ)under lieer.
~~AI5T'f:t ~~national.
a'i r quire the measurement of the 6.3 Statistician:
lesion, which then becomes the 6.3.1 The test method task group should obtain the assis-
tance of a person familiar with the statistical procedures in this
practice and with the materials being tested in order to ensure
that the requirements outlined in this practice are met in an
efficient and effective manner. This person should also assist
the task group in interpreting the results of the data analysis.
6.3,2 When a person having adequate knowledge of both
the materials and the proper statistical techniques is not
available, the task group should obtain the services of a
statistician who has experience in practical work with data
from materials testing. The task group should provide the
statistician with an opportunity to become familiar with the
statistical procedures of this practice and with both the mate-
rials and the test method involved, The statistician should
become a member of the task group conducting the ILS (task
group members need not be members of ASTM).
6.3,3 The calculations of the statistics (see Section 15) for
each material can be readily done by persons not having
statistical knowledge (see 15.1.3 and 15.4.2).
6.4 Data Analyst-This individual should be someone who
is careful in making calculations and can follow the directions
in Sections IS through 17.
6.5 Laboratory ILS Supervisor-Each laboratory must have
an ILS supervisor to oversee the conduct of the ILS within the
laboratory and to communicate with the ILS Coordinator.The
name of the supervisor should be obtained on the response
form to the "invitation to participate" (see 9.4).

7. Basic Design
7.1 Keep the design as simple as possible in order to obtain
estimates of within- and between-laboratory variabilitythat are
free of secondary effects. The basic design is represented by a
two-way classification table in which the rows represent the
laboratories, the columns represent the materials, and each cell
6. ILS Membership
(that is, the intersection of a row with a column) contains the
test results made by a particular laboratory on a particular
material (see Table I).

8. Test Method
8,I Of prime importance is the existence of a valid, well-
written test method that has been developed in one or more
competent laboratories, and has been subjected to a ruggedness
test prior to the ILS.
8.2 A ruggedness test is a screening procedure for investi-
gating the effects of variations in environmental or other
conditions in order to determine how control of such test
conditions should be specified in the written description of the
method, For example, the temperature of the laboratory or of a
heating device used in the test may have an effect that cannot
be ignored in some cases but may be much less in others. In a
ruggedness test, deliberate variations in temperature would be
introduced to establish the allowable limits on control of
temperature. This subject is discussed more fully in Practice
E1169.
'To facilitate the preparation of e fin I report on the ILS, the task group can 8.3 As a result of carrying out the screening procedure, and
obtain the Research Report format i I fiom ASTM Headquarters. of some experience with the test method in the sponsoring

4
 r
I d OT~ COPYrightASTM [nternational, 100 Barr Harbor Driv,
1 ~~ £fIest Conshohocken, Pennsylvania 19428·295~
~ noHAL United States of America, This copy has been mar
laboratory and one or two other laboratories, a written v N,,",,"sd.e~'lii1s of. t~y Jl~oEedyrrt,,~p.Jd/Jt;A,wiJ-lil1gJ9.assign the work to a
of the test method must have been developed (but not sQYthe QtmIWat~~brt\~(l!tdfJ~ ~l!:i~'E!ij\rli!tI100f.l'Tfi~I~~cisionof a labora-
sarily published as a standard method). This draft ShOljl¢tomA~~t:tn~mmlJr1a~ipate should be recorded on a response form to a
describe the test procedure in terms that can be easily followeci-..writteg igvitatiot:l. The jt:l."itatiQR~QQ'IIQinclude information
in any properly equipped laboratory by competent personnel covering the required time for calibrating the apparatus and for
with knowledge of the materials and the property to be tested, testing all of the materials, and other possible costs, The
The test conditions that affect the test results appreciably response form should include the name, address, and telephone
should have been identified and the proper degree of control of number of the person supervising the ILS work within the
the test conditions specified in the description of the test laboratory, the address and other markings required to ensure
procedure. In addition, the test method should specify how the ILS sample material will be promptly delivered to the ILS
closely (that is, to how many digits) each observation in the test supervisor, answers to brief questions concerning equipment,
method is to be measured. environment, and personnel, including previous use of the test
8.4 The test method shouldspecify the calibration proce- method, upon which the apparent competence of the laboratory
may be judged, and an affirmation that the laboratory under-
dure and the frequency of calibration.
stands what is involved and agrees to carry out its responsi-
bilities with diligence.
9. Laboratories
9.5 The ILS should not be restricted to a group of labora-
9,1 Number of Laboratories:
tories judged to be exceptionally qualified and equipped for the
9.1.1 An ILS should include 30 or more laboratories but this
ILS. Precision estimates for inclusion in a test method should
may not be practical and some ILS have been run with fewer.
be obtained through the efforts of qualified laboratories and
It is important, that enough laboratories be included in the ILS
personnel operating under conditions that will prevail when the
to be a reasonable cross-section of the population of qualified
test method is used in practice.
laboratories; that the loss or poor performance of a few will not
be fatal to the study, and to provide a reasonably satisfactory 10. Materials
estimate of the reproducibility.
9.1.2 Under no circumstances should the final statement 10.1 Material designates anything with a property that can
of precision of a test method be based on acceptable test be measured. Different materials having the same property may
results for each material from fewer than 6 laboratories. be expected to have different property levels, meaning higher
This would require that the ILS begin with 8 or more or lower values of the property. Different dilutions of the same
laboratories in order to allow for attrition. material or compound to be assayed are considered "different
9.1.3 The examples given in this practice include only 8 and materials" for the purpose of this practice. The terminology
7 laboratories, respectively. These numbers are smaller than "different levels of material" may be used, if appropriate.
ordinarily considered acceptable, but they are convenient for 10.2 The number and type of materials to be included in an
illustrating the calculations and treatment of the data. ILS will depend on the range of the levels in the class of
9.2 Any laboratory considered qualified to run the test materials to be tested and likely relation of precision to level
routinely (including laboratories that may not be members of over that range, the number of different types of materials to
ASTM) should be encouraged to participate in the ILS, if the which the test method is to be applied, the difficulty and
preparatory work is not excessive and enough suitably homo- expense involved in obtaining, processing, and distributing
geneous material is available. In order to obtain an adequate samples, the difficulty of, length of time required for, and
number of participating laboratories, advertise the proposed expense of performing the test, the commercial or legal need
ILS in where appropriate (for example, trade magazines, for obtaining a reliable and comprehensive estimate of
meetings, circulars, etc.). precision, and the uncertainty of prior information on any of
these points.
9.3 "Qualified" implies proper laboratory facilities and test- 10.2.1 For example, if it is already known that the precision
ing equipment, competent operators, familiarity with the test is either relatively constant or proportional to the average level
method, a reputation for reliable testing work, and sufficient over the range of values of interest, a smaller number of
time and interest to do a good job. If a laboratory meets all the materials will be needed than if it is merely known that the
other requirements, but has had insufficient experience with the precision is different at different levels. The ruggedness test
test method, the operator in that laboratory should be given an (see 8.2) and the preliminary pilot program (see Section 13)
opportunity to familiarize himself with the test method and help to settle some of these questions, and may often result in
practice its application before the ILS starts. For example, this the saving of considerable time and expense in the full ILS.
experience can be obtained by a pilot run (see Section 13) 10.2.2 An ILS of a test method should include at least three
using one or two trial samples provided by the task group and materials representing different test levels. and for develop-
returning the raw data and the test results to the task group. ment of broadly applicable precision statements, six or more
The importance of this familiarization step cannot be materials should be included in the study.
overemphasized. Many interlaboratory studies have turned 10.2.3 The materials involved in anyone ILS should differ
out to be essentially worthless due to lack of familiarization. primarily only in the level of the property measured by the test
9.4 Obtain written ensurance from each potential participat- method. When it is known, or suspected, that different classes
ing laboratory that it is properly equipped to follow all the of materials will exhibit different levels of precision when

5
 IInT~,CopyrightASTMInternational,100 BarrHarborDriv-

~=:'A=~
~ Pennsylvania19428·295?
WestConshohocken,
UnitedStatesofAmerica.
ti$921tOfilt~ir Standards
Thiscopyhasbeenmac·
andQuality(STAMEQ)
underlim
I

tested by the test method, 0 I n ideration should be giv nfrq0A.5TM In~~~io~bscribe any special circumstances that must be ad-
conducting separate inter1~ la ory studies for each class 0 resse In Imp emen mg t e repeata I ity conditions, such as
material. i the period of time between obtaining the test results for the
10.3 Each material in I S should be mad t b same material; that is, not less than in normal testing and not so
selected to be as homogenso s as possible prior toe its0 subdi- e or Ion~ as to likIley .permltit sigru
siznifi cant c hanzes
anges iIn test matenal,
.
vision into test units or te t p cimens. If the randomization equipment or environment,
and distribution of individ a l't<1stspecimens (rather than test 12.5 Specify the required care, handling, and conditioning "
units) does not conflict wi ~he procedure for preparing the of the materials to be tested. Explain the coding system used in
sample for test, as spec tl I in the test method, greater identifying the materials and the distinction between test units
homogeneity between test n' ts can be achieved by randorniz- and test specimens, where appropriate.
ing test specimens. Then e t st unit would be composed of
the required number of d mized test specimens. (See 12.6 Supply data sheets for each material for recording the
Section II and 14. J for th pa tity of each material needed, raw data as observations are made. Give instructions on the
its preparation and distribu lc . number of significant digits to be recorded, usually one more,
if possible, than required by the test method. Also, supply test
NOTE 2-It may be convenie t 0 se established reference materials, result sheets on which test results can be calculated and
since their homogeneity has be onstrated. reported. In many instances this can be combined with the raw
data sheet. Specify the number of significant digits to be
11. Number of Test Resu reported, usually two more than required by the test method.
11.1 In the design of an sufficient total number of test Request the laboratories send raw data and test result sheets as
results on each material t e specified to obtain a good soon as the testing is completed, and at least weekly if testing
estimate of the measure of e e tability, generally the repeat- will continue over several weeks. For guidance on the number
ability standard deviation. tn ny cases, the standard devia- of significant digits needed for data reporting see Practice E29.
tion in question will be a n I ti n of the property level being 12.7 Request that each laboratory keep a record (or log) of
measured. When this occu ,~h standard deviation should be
determined separately for r eve!. It is generally sound to
limit the number of test e. ul s on each material in each
any special events that arise during any phase of the testing.
This record, to be sent to the ILS coordinator, will provide a
valuable source of information both in dealing with unusual
laboratory to a small nu ti,r, such as three or four. The data and in making improvements in the test method in future
minimum number of test r s lIt] per laboratory will normally revisions.
be three for a chemical tes a ,d three or four for a physical or
12.7.1 Instruct the laboratories to notify the ILS coordinator
optical test. The number m pe as small as two when there is
promptly whenever an error in test procedure arises, so that a
little danger that a test un i I be lost or questionable test
decision can be made as to whether a new set of test units
results obtained, or as man I ~n when test results are apt to
vary considerably. Generall ,lh time and effort invested in an should be sent to the laboratory for a complete retest of the
ILS is better spent on exa i lin more materials across more material.
laboratories than on recordi a. arge number of test results per 12.8 Enclose with the protocol a questionnaire requesting
material within a few labo t, n s. information on specific aspects of the apparatus, reagents,
calibration, or procedure, as well as any other information that
12. Protocol might assist in dealing with data inconsistencies, or ensure the
12.1 In the protocol, cit t, e name, address, and telephone task group that the laboratory complied with the current
number of the person who ~ en designated ILS coordinator requirements of the test method. Also obtain any other infor-
(see 6.2). Urge the labora Ires to call the coordinator when mation that may be needed in preparing the final research
any questions arise as to t 0 duct of the ILS. report on the ILS.

. 12.2 Clearly identify th s Ie ific version of the test method CONDUCTING THE TESTING PHASE
being studied. If the test allows several options in OF THE ILS
apparatus or procedure, t e p otocol should specify which
option or options have bee se e ted for the ILS. Test units and 13. Pilot Run
test data sheets mu~t be ~r vIr for each option. . 13.1 Before investing laboratory time in the full scale ILS,
12.3 When special cali a 10 procedures are required be- it is usually wise to conduct a pilot run with only one, or
fore every determination 9 v ry test result, they should be perhaps two, material(s) to determine whether the test method
described specifically in t te t method. If the test method as well as the protocol and all the ILS procedures are clear, and
specifies calibration only dill r less frequently, the ILS task to serve as a familiarization procedure for those without
group must decide whet rt .require recalibration before sufficient experience with the method (see 9.3). The results of
obtaining each test resul ile doing so will eliminate this pilot run also give the task group an indication of how well
calibration drift and help sir relative independence of the each laboratory will perform in terms of promptness and
test results, changes in cali n may increase the variability following the protocol. Laboratories with poor performance
between test results. should be encouraged and helped to take corrective action.

6
 ASTM International,
Copyright 100 BarrHarborDriv'
Pennsylvania19428·295[
- ~§ Conshohocken,
I ~

13.2 All steps of the procedures described in this pr c~:::~~~~.gi!~i1:~iC9\.~~cDB~~AvabF STATISTICS

should be followed in detail to ensure that these directio \W!tle DirectarateJor StandaJds andQuality(STA~EQ) underlieer
understood, and to disclose any weaknesses in the proto I or ~5. Calculation of the Statlstlcs
the test method. . . rom ASTM ln~~.~IO~verview-The analysis and treatment of the ILS test
results have three purposes, to Idetermine whether the collected
14. Full Scale Run data are adequately consistent to form the basis for a test
14.1 Material Preparation and Distribution: method precision statement, to investigate and act on any data
14.1.1 Sample Preparation and Labelling-Prepare enough considered to be inconsistent, and to obtain the precision
of each material to supply at least 10 % more than needed by statistics on which the precision statement can be based. The
the number of laboratories committed to the ILS. Label each statistical analysis of the data for estimates of the precision
test unit or test specimen with a letter for the material and a statistics is simply a one-way analysis of variance (within- and
sequential number. Thus, for ten laboratories and two test between-laboratories) carried out separately for each level
results for each laboratory the test units for Material B would (material). Since such an analysis can be invalidated by the
be numbered from B 1 to B22, or, if five test specimens per test presence of severe outliers, it is necessary to first examine the
unit are required, the test specimens may be numbered B 1 to consistency of the data. The following paragraphs show, in
BllO. terms of a numerical example, how the entire program is
14.1.2 Randomization-For each material independently, carried out:
allocate the specified number of test units or test specimens to 15.1.1 The calculations are illustrated with test results from
each laboratory, using a random number table, or a suitable an ILS in which the concentration of glucose in serum (see
computerized randomization based on random numbers. See Table 1) was measured at five different concentration levels by
Guide E1402 for a discussion of randomization. eight laboratories. Each laboratory obtained three test results at
14.1.3 Shipping-Ensure that the test units are packaged each concentration level.
properly to arrive in the desired condition. When the material 15.1.2 For extended calculations it is usually necessary to
is sensitive to the conditions to which it is exposed (light, heat, retain extra significant digits in order to ensure that statistically
humidity, etc.), place special directions for opening the pack- important information is not lost in calculation by rounding off
age on a label outside the package. Clearly indicate the name too soon. As a general rule, retain at least two more digits in the
of the person who has been designated as ILS supervisor at the averages than in the reported test results and at least three
laboratory on the address of each package. Follow each significant figures in the standard deviations.
laboratory's instructions for ensuring prompt delivery of the 15.1.3 While the calculations described in this section are
package. arranged for use of a hand calculator, they also can be readily
14.1.4 Follow-up-Once the test units have been shipped,
the ILS coordinator should call each laboratory ILS supervisor
TABLE 1 Glucose in Serum ILS Test Result Data
within a week to ten days to confirm that all test units have
Material
arrived safely. If the task group has decided to intermingle test Laboratory
A B C D E
units from different materials in the order of testing, the testing
41.03 78.28 132.66 193.71 292.78
should not start until all the test units have arrived at the 294.09
41.45 78.18 133.83 193.59
laboratory so they can be tested in the specified order. 41.37 78.49 133.10 193.65 292.89
14.1.5 Replacement Sets of Test Units-As the ILS
2 41.17 77.78 132.92 J90.88 292.27
progresses, a laboratory may discover that the test method was 136.90 200.14 309.40
42.00 80.38
not used properly on some test units. The laboratory ILS 41.15 79.54 136.40 194.30 295.08
supervisor should discuss this with the ILS coordinator, who
3 41.01 79.18 132.61 192.71 295.53
may send a replacement set of test units, replace the misused 40.68 79.72 135.80 193.28 290.14
test units, or do nothing, as may seem desirable. 42.66 80.81 135.36 190.28 292.34

14.2 Checking Progress-From time to time, at intervals 4 39.37 84.08 138.50 195.85 295.19
appropriate to the magnitude of the ILS, the coordinator should 42.37 78.60 148.30 196.36 295.44
call each ILS supervisor to ascertain how the testing is 42.63 81.92 135.69 199.43 296.83

progressing. By comparing the progress of all laboratories, the 5 41.88 78.16 131.90 192.59 293.93
coordinator can determine whether some laboratories are 41.19 79.58 134.14 191.44 292.48
41.32 78.33 133.76 195.12 294.28
lagging considerably behind the others and so advise these
laboratories. 6 43.28 78.66 137.21 195.34 297.74
40.50 79.27 135.14 198.26 296.80
14.3 Data Inspection-The completed data sheets should be 42.28 81.75 137.50 198.13 290.33
examined by the coordinator immediately upon receipt in order
to detect unusual values or other deficiencies that should be 7 41.08 79.76 130.97 194.66 287.29
41.27 81.45 131.59 191.99 293.76
questioned. Replacement sets of test units or of specific test 39.02 77.35 134.92 187.13 289.36
units may be sent when there is missing or obviously erroneous
data. The task group can decide later whether or not the 8 43.36 80.44 135.46 197.56 298.46
42.65 80.80 135.14 195.99 295.28
additional data should be used in the estimation of the precision 41.72 79.80 133.63 200.82 296.12
of the test method.

7
 A nT~') Copyright ASTMInternational, 100 Barr Harbor DrivF
¥ West Conshohocken, Pennsylvania 19428·295~
IItRIfNA1"IONAL United States of America, This copy has been mac
$Mndirds Wo(ll1wl(/~

~~~~irmo1oi for Standards and Quality (STAMEQ)under lieer

programmed for the com 'r. A spreadsheet can be eas Iyrom ASTM Internation~1. (132.92 + 136.90 - 136.40)
adapted to these calculatio s x _J35.407

15.2 Table of ILS Test fU ts-The test results received 15.4.2 Cell Standard Deviation, s-Calculate the standard
from the laboratories are s al y best arranged in rows and deviation of the test results in each cell using the following
I
columns as in Table I. Eac cbh mn contains the data obtained equation:
from all laboratories for 0 ~ ~ terial, and each row contains
the data from one laborato I 0 all materials. The test results
from one laboratory on one rr jjltnial constitute a cell. Thus, the
s= 2:", (x - .e) 2/(n - 1) (2) ; ,

cell for Laboratory 2 and te ial C contains the test results

The symbols have the same meaning as for Eq 1. Thus for
132.92, 136.90 and 136.40 11 Jli: cell is called C2, by material
Cell C2:
and laboratory. It helps it Iile interpretation of the data to
arrange the materials in 1(, re sing order of the measured [{- 2.487)2 + {1.493)2 + {0.994)2] _ )9.400448 _
values. I s= (3 _ I) - 2 -2.168

15.3 Worksheets-Gener Ity, t facilitates the calculations to While Eq 2 shows the underlying calculation of the cell
prepare a separate calcula ~(n ~orksheet for each material, standard deviation, inexpensive pocket calculators are avail-
using Table 2 as a model t p1 lm~kingappropriate changes ~or able that calculate both the average and the standard deviation
different numbers of labora 0 Ie , and test results per material. directly. Check to be sure the calculator uses (n - I) as the
Enter the test result data fo o~e material (from one column of divisor in Eq 2, not n, and has adequate precision of calcula-
Table 1) on a worksheet.Al o br ter the results of the following tion.
calculations for that mats i'lion the same worksheet, as
illustrated in Table 2. Wor I(~ nly one material at a time. 15.5 Intermediate Statistics:
15.5.1 Average of the Cell Averages, x-Calculate the
15.4 Cell Statistics:
average of all the cell averages for the one material using Eq 3.
15.4.1 Cell Average,x- ': c late the cell average for each
1

p
laboratory using the follow n. equation:
x = 2: xl p (3)
[" I

j F [x/n (I)
I where:
where: x = the average of the cell averages for one material,
x =the average of the t, ~1, results in one cell,
x = the individual cell averages, and
p = the number of laboratories in the ILS.
x =the individual test r s ItI in one cell, and
n =the number of test'i ,s It in one cell. Thus for Material C:
Thus from Table 2 for:tv t Iri I C, Laboratory 2 (that is, for
Cell C2):
x = 1081.1432
8 = 135.1429

TABLE 2 lnterlabcr t ry Study Worksheet for Glucose in Serum Initial Preparation of Test Result Data for Material C
"" I ~st Results, x
Laboratory d h k
5
Number 2 3
1 132,66 133,83 133,10 133,197 0,591 1,946 -{l.73 0,22
2 132,92 136.90 136.40 135.407 2,168 0.264 0,10 0,79
3 132,61 135.80 135.36 134.590 1.729 -{l.553 -{l,21 0,63
4 138.50 148.30 135.69 140.830 6.620 5.687 2,14 2.41
5 131.90 134.14 133.76 133.267 1.199 -1.876 -{l,71 0.44
6 137.21 135.14 137.50 136.617 1.287 1.474 0,55 0.47
7 130,97 131.59 134.92 132.493 2.124 -2.650 -1.00 0,77
8 135.46 135.14 133.63 134.743 0.977 -{l.400 -{l,15 0,36
Average of cell averages, j(' - 135.1429
Standard deviation of cell averages, s. = 2.6559
Repeatability standard deviation, s,= 2,7483
Between-Laboratory standard deviation, SL = 2.1298
Reproducibility standard deviation, SR = 3.4770

where:
x = individual test result (see 15 )
x = cell average(see 15.4,1).
:'. = cell standard deviation (see ,5-1-2),
(SE[
x = average of cell averages
d = cell deviation (see 15.5,2).
ex = standard deviation of cell a
5,S,I).

f~
l
ge (see 15,5.3),
s, = repeatability standard devia f' (s, e 15.6.1).
SL = between-laboratory standa devl tion (see 15.6,2),
SR = reproducibility standard de at n see 15,6,3),
h = between-laboratoryconsisle c (s e 15,7,1). and
k within-laboratory consistenc (I e 5,7.2).

8
 Iifl.Tl..l Copyright ASTMinternational, 100Barr Harbor Drivr
_.Jff1 Qest Conshohocken, Pennsylvania 19428·295c
15.5.2 Cell Deviation, d-For each laboratory calculate le::",:IML United States tWmefiC4 it~~h~ ~PJlWtk"
cell deviation by subtracting the average of the cell avera ~ the Direw&llfafsfyStaodards and Q!lality (SIAMEQ) ~erititer.
from the cell average using the following equation: . ABC D E
d = s_x 4_f 2romASTMInterqatlOilal. 0.21 0.11 0.22 0.02 0.18
_ Q 8.~8 e.t3§ 0.;9 1.78 2.33
3 1.00 0.56 0.63 0.61 0.69
Thus for Cell C2: 4 1.70 1.85 2.41 0.74 0.22
5 0.34 0.52 0.44 0.72 0.24
d = 135.407 - 135.143 = 0.264
0.63 1.03
6 1.32 1.09 0.47
7 1.17 1.38 0.77 1.45 0.84
15.5.3 Standard Deviation of the Cell Averages, s;c-
8 0.77 0.34 0.36 0.94 0.42
Calculate this statistic using the following equation:
A Criticalvalue = 2.06.

s, = ~" d 2/(p - I) (5)

Thus for Material C:

49.376634 15.6.2.1 The data for Material A illustrate the case of

(8 - "1) ."" y' 7.053805 = 2.6559 negative estimate for s/ (see Table 8 for the required statistics
Sx and s; for Material A).
15.6 Precision Statistics-Wh.ile there are other precision
Thus for Material A:
statistics, introduced later in this practice, the fundamental
precision statistics of the ILS are the repeatability standard sr = 0.606(2 -
1.06322/3 = -0.009441,
deviation and the reproducibility standard deviation. The other = 0,
set sr
statistics are calculated from these standard deviations. and set SL = O.
NOTE 3- This situation may occur when the laboratories are in
15.6.1 Repeatability Standard Deviation, sr-Calculate this
excellent agreement, in which case both Sx 2 and s/In in Eq 7 tend to
statistic using the following equation: become estimates of the variance of laboratory averages, and their
difference will fluctuate around zero, causing the estimate SL2 to take on
(6) negative values at times. Because variances cannot be negative (being
s,=~
proportional to a sum of squared deviations from an average), any
negative estimate of the between laboratory variance must be set to zero.
where:
s; = the repeatability standard deviation, 15.6.3 Reproducibility Standard Deviation, sR-Calculate
s = the cell standard deviation (p of them from Eq 2), and this statistic using the following equation:
p = the number of laboratories. s R =~~
V SiTS; (10)
Thus for Material C:
Thus for Material C:
60.425223 _--
s, = 8 = y'7.553153 = 2.7483 SR = y'4.536087+2.748J2 = 3.4770

15.6.2 Between Laboratory Variance, SL2, and Standard Thus for Material A:
Deviation sL-Calculate this variance and standard deviation
SR = y' 0 + 1.06322 = 1.0632, thus SR = s,
using the following equations:
15.7 Consistency Statistics, hand k:
s~ = s.~- syn (7)
15.7.1 For each cell, calculate a value of h using the
(8) following equation:
2 2 h = dIs .e ( I I)
If s L is negative, set SL = 0 and SL = O.
Thus for Material C: where:
s~ = 2.65592 - 2.74832/3 = 7.053805 - 2.517718 = 4.536087 h = the between-laboratory consistency statistic,
SL = y'4.536087 = 2.1298 (9)
d = the cell deviation (that is, the deviation of the cell
average from the average of the cell averages, from
15.5.2), and
TABLE 3 Glucose in serum-rr'
S;c = the standard deviation of the cell averages (from
Material
Laboratory 15.5.3).
A B C D E
1 -0.39 -1.36 -0.73 -0.41 -0.46 Thus for Cell C2:
2 -0.13 -0.45 0.10 0.15 1.64
-0.11
0.264
3 0.22 -0.21 -1.01 -0.68
4 -0.10 1.85 2.14
h = 2.6559 = 0.10
0.96 0.49
5 -0.09 -0.99 -0.71 -0.64 -0.34
6 0.83 0.21 0.55 0.97 0.17 Retain two decimal places in the computed values of h.
7 -1.75 -0.16 -1.00 -1.33 -1.62
1.75
15.7.2 For each cell, use the following equation to calculate
8 0.67 -0.15 1.31 0.79
a value of k.
A Criticalvalue = 2.15.
( 12)

9
 OT

'Sl<l<l4S)jlOM orarados uo :ZJ IJ<lJ lOJ sn4l

srjnsor UO!lU[n:J[u:J <l41 ~U!l<llU<l 'IUP<llUW 4:JU<llOJ L'S'I 4~nOl41 '(I'9'S'( WOlJ)
t'S'I U! p<lqp:JS<lp sdois <l41 lU<ld<l}j-s/v!.laJVW .lalfl0 8'S'I JUp<llUW <l41 JO U0!lU!A<lP P U 'U S A1!J!qulU<ld<ll <l41 = "S

''I JO S<lnJUApomduroo <l41 U! sooejd Juw!:><lp OM1 U!U1<l}j pUU '(Z't'S' [

WOlJ) Al01UlOqU[ <lUO lOJ UOnU.A p PlUPUU1S 1J<l:J<l41 = S
6L'0 = £8vL'Z = )/ ':J!lS!lU1S A:JU<llS!stbo, 01UlOqUJ-U!41!M <l41 = 'I
89n: :<ll;}4M

'90'1: = anlell leo!lPO e 'S ~'1: = anlell jll:l!IIJO e

'(ST01: pue vT01: aas) 170lIaO 5u!I:laJJo:l Jalle sanlell paleln:lle:la~ v ,(S' ~'01: pue 17'~'01: aas) 170 la~, 5 uoauoo Jalle sanlell pallljn:lje:la~ v

1:17'0 176'0 89'0 178'0 LL'O 8 6L'O ~8'~ W'O L' '0 sn 8
171)'0 sn 88' ~ 8n H'~ L 1:9'~- cci- 81:'~- 911'( sn- L
80'~ 89'0 88'0 60'~ zn 9 L~'O L6'O 60' ~ ~I'o 88'0 9
171:'0 GL'O 8LO ZS'O 178'0 S 178'0- V9'0- 178'0- 61 rt 60'0- S
1:1:'0 vL'O ZO'~ S8' ~ on 17 6V'0 96'0 69'~ S~'I 01'0- 17
69'0 19'0 ZI'I 9S'O 00'1 8 89'0- 10'1- 80'0- Z~'( ~~'O- 8
88'1: 8n 017'1 68'0 917'0 Z 179'1 91'0 68'0 S '0 81'0- Z
8~'O ZO'O 88'0 I~'O IZ'O I 9V'0- 117'0- 88'0- 9 '~ 68'0- I
3 a 0 8 'V 3 a 0 ~ 'V
fiJolllJOqlll
fiJoleJoqel
lepaleV'j le!JaleV'j I
s'vJf-wnJas u! aso:mlD L 318\f~ s'v'/-wnJas u! asl':) ID 9 318\f~

'Z'8'Z'IX U! 81'IX b31~~IS1 ~::; le:l!liJ:l Jf a41 10 uO!lllln:lle:l JO"

TZ'Z'IX U! S'IX b3 1as ,e
II 11l:l!l!J:l(/ a41 10 U0!leln:llll:l JO"
'S:l!lS!lelS A:lUaIS!SUO:lJf pue u a41 JOI sanlell le:l!liJ:l 10 uO!leln:lle:l JOI SelnWJOI uousin ,Ieq p ~ suouexpep JOI Z' IX UO!paS aas
09'1
09' ~
179'1
179'1
89'1
89'1
8L'I
8L'I
08'1
08'1
68'1
68'1
ZO'Z
ZO'Z
vZ'Z
vZ'Z ~
~t'~
~,
08
6Z
V9'Z
179'1:
09' ~ 89'1 89'1 8L'I 08'1 68'1 ZO'Z 8Z'Z ~~' 8Z 89'1:
09'1 89'1 89'~ 8L' ~ OB'l 68'~ ZO'Z 8Z'Z LG 1:9'1:
09'1 89'1 89'1 8L'~ 08'1 68'1 1:0'Z 8Z'Z 9Z Z9'1:
09'1 89'1 L9'I 8L'I 6L'I 88'~ 10'Z 8Z'Z II~: S1: 19'Z
09'~ 89'1 L9'I 8L'I 6L'I 88'1 10'Z ZZ'Z VZ 09'Z
6S'~ 89'1 L9'I U'I 6L'I 88'1 10'Z ZZ'Z g' 81: 6S'Z
6S' ~ 89'1 L9'I ZL'I 6n 88'1 10'Z IZ'Z S' 1:1: 89'Z
6S'I 89'1 L9'I zn 6L'I 88'1 OO'Z IZ'Z IZ LS'Z
~~'
6S'I 89'1 L9'I U'I 6n L8'I OO'Z IZ'Z d' OZ 9S'1:
6S'~ Z9'I L9'I zn 8n L8'I OO'Z OZ'Z a,' 61 vS'G

d'ut'
6S'I 1:9'1 99'1 GL'I 8n L8'I 66'1 OZ'Z 81 8S'1:
6S'I Z9'I 99'1 IL'I 8n 98'1 66'1 61'Z H IS'Z
8S'I Z9'I 99'1 In LL'I 98'1 86'1 81'Z 91 6V'1:
~~'
8S' ~ Z9'I 99'~ In LL'I 98'1 86'1 LI'Z 1J' SI Lv'Z
OL'I LL'I S8'I L6'I 91'Z 171 VV'Z
8S'~ 19'1 S9'I 9~'

\
8S'~ 19'1 S9' ~ OL'I 9L'I 178'1 96'1 91'Z 81 117'1:
LS'I 09'1 1>9'1 69'1 9L'I vB'l 96'1 vl'Z 1:1 88'Z
LS'~ 09' ~ 1>9'1 69'1 SL'I 88'1 176'1 81'Z u 178'1:
9S'~ 6S'! 89'1 89'1 vn 1:8'1 86'1 ,,'Z g' 01 61:'Z
9S'~ 6S'! 1:9'1 L9'I 8L'I 18'1 Z6'I 6O'Z II, 6 81:'1:
SS' ~ 8S'! Z9'I 99'1 U'I 6n 06'1 90'Z 8 SI'Z
VS'1 LS'I 09'1 S9'I on LL'I L8'I 80'Z L SO'1:
ZS'I
OS'~
Lv'l
ZV' ~
01
SS'!
8S'!
OS'I
1717'1
6S'!
99'1
8S' ~
Lv'l
8
89'1
09'1
9S'I
617'1
L
89'1
S9'I
09'1
ZS'I
9
SL'I
In
99'1
9S'I
S
vB'l
6L'1
8L'I
19'1
17
86'1
Z6'I
Z8'I
L9'I
8
Z,, 'I,
'jr~ 9
S
17
8
Z6'~
vL'1
6V'1
SI'I
6 I d (/
U 'qel rad Sale:l!ldaJ 10 JaqwnN sqllj 10 10 anjell
Jf 10 sanjell le:l!liJO II JaqwnN 18:l!l'JO

laAa1 a:lue:lIJ!LI~~!lr~~i~ w.~~'MJr ~e Il jO sanleA le:)mJo !l 318Vi

mil Japun (b3W~1S) Al!1enopue SPJepU215JOJaleJQllJ~ "'~ ~

,Jew uaaq se4 Ado) sl41 'e)!Jaw~ JOsalelS pal!Un ~~~
jS6l-8lv61 2!U2AlhSuuad 'Ua~)0404SUOJ lsaM A
'AUG JOQJ2Hmg OOl'leuoqeUJalUj W1S~ 14fiuhdoJ I...V'UU ~
 d flT~ Copyright ASTM International, 100 Barr Harbor Driv
~ West Conshohocken, Pennsylvania 19428-295
''''TERNAT/ONAL United Statp5..nf lmerir..a.. JhiuQ.1J.II has hp~n mac
TABLE 8 Glucose in Serum-Precision Statistics S/andifds Worldwide TIJ\T~ {;ONSI~TEN'CY"
NOTE I-This table (with the column for Sji omitted) is a useful forma
by the Directorate for Standards and Quality (STAMEQ) under lieer
for the presentation of the precision of a test method as required by fromlKs-Tf~~~a~nconsistent Results
Section A21 of the Form and Style of ASTM Standards.
17.1 Ct itical I"hluesvI the Consistency 8tatistics- Table 5
Mate-
rial
x s. 5, SR .r R lists critical values of the hand k consistency statistics at the
A 41.5183 0.6061
- '(0632 . 1.0632 . 2~98 2.98
- 0.5 % significance level. The critical values for h (first column)
B 79.6796 1.0027 1.4949 1.5796
"
4.19 4.42 depend on the number of laboratories (p, second column)
C 134.7264 1.7397 1.5434 2.1482 4.33 6.02 participating in the ILS and the critical values for k (columns
D 194.7170 2.5950 2.6251 3.3657 7.35 9.42
E 294.4920 2.6931 3.9350 4.1923 11.02 11.74
headed 2 through 10) depend both on the number of laborato-
ries (P) and on the number of replicate test results (n) per
laboratory per material. The 0.5 % level was chosen based on
16, Tabular and GraphicalDisplay of Statistics the judgment and experience that the 1.0 % resulted in too
many cells being flagged and the 0.1 % level in too few. For
16.1 Material Order-It is often useful to arrange the
further discussion see Appendix Xl.
worksheets in order of increasing values of X, the material
17.1.1 Obtain from Table 5 the appropriate critical values.
averages. This order may facilitate interpretation.
For the glucose in serum example, the respective critical hand
16.2 Tables-From the Table 2 results for each material, k values are 2.15 and 2.06. In Table 3 and Table 4 circle those
prepare tables of hand k as shown in Table 3 and Table 4 for values that exceed the critical values and underline those
the glucose in serum example. values that approach the critical values. On Fig. 1, draw
16.3 Graphs-Prepare bar graphs for hand k with materials horizontal lines for positive and negative values of h. On Fig.
grouped by laboratory as in Fig. I and Fig. 2, respectively. 2, draw a horizontal line for k.
Arrange the laboratories and materials within and between 17.1.2 The hand k graphs and the marked tables give a
each grouping in the same order as used in Table I. Thus the picture of the overall character of the variability of the test
materials will be arranged in order of increasing x from left to method as well as singling out particular laboratories or cells
right, and the laboratories in order of laboratory code number. that should be investigated.

2.5 -

- ------------------- ----------------------. 2.15

2

1.5 -

1
u
t;
'.19.., .5 -

-]Jrl )r r' l-lr

III
~
c o - J1
~
'iii
c
l j_ r
8 I
-.5
.c

-1

-1.5 -

-2
- -------------------------------------------
-2~5
-2.5 -
MAT: ABCDE ABCDE ABCDE ABCDE ABCDE ABCDE ABCDE ABCDE
LAB: 2 3 4 5 6 7 8
FIG. 1 Glucose in Serum: h-Materlals within Laboratories

II
 ~,"I
~ Copyright ASTMInternational, laD Barr Harbor Driv,
West Conshohocken, Pennsylvania 19428·295(
:DfJO:~'nited States of America. This copy has been ma(
3 by the Directorate for Standards and Quality (STAMEQ)under lieer.
from ASTM International.

2.5 -

- ---------- -------- ----------------------- 2.06

2

~
~ii 1.5 -
I_
s
~
I

l_! [j_ j _ _II J

MAT: ABCOE ABCDE ABCDE ABCDE ABCDE ABCDE ABCDE ABCDE
LAB: 1 2 3 4 5 6 7 8
FIG.2 Glucose In Serum: k-Materials within Laboratories

17.2 Plots by Laborato I order to evaluate the differ- 18. Investigation

ences between laboratories u e he following guidelines. 18.1 Clerical and Sampling Errors-Examine the labora-
17.2.1 h Graph-There r I tTree general patterns in these tory report for each flagged cell. Try to locate where each test
plots. In one, all laboratorie aVFboth positive and negative h result in the flagged cell begins to deviate from the others. Is it
values among the materi s.IIi the second, the individual in the original observations? Are the data rounded prema-
laboratories tend to be ei r positive or negative for all turely? Are the calculations correct? Then, look for signs of
materials and the number I e~tive laboratories equals the mislabeling of test units such that the test result for one
number of positive laborat i , eore or less. Neither of these material was reported as belonging to another material. Check
patterns is unusual or requir s n estigation, although they may these errors with the laboratories: do not assume them to be so.
tell something about the nat r 10. the test method variability. In
the third pattern, one labor y, with all h values positive (or 18.2 Procedural Errors:
negative), is opposed to al tie other laboratories, with sub- 18.2.1 Study the laboratory reports again looking for devia-
stantially all the h values n a iv (or positive). Such a pattern tions from either the test method or the protocol. For instance,
calls for an investigation 0 t at laboratory.
I
variations in the number of significant digits reported in the test
17.2.1.1 Another kind 0 ~att rn to look for occurs within results may be a sign of incorrect rounding, or that the
one laboratory, in which th n alues for low property levels equipment in one laboratory is different from the rest. Also,
p
are of one sign, and for hig 10 erty levels are of the opposite study the event log for special comments relating to the flagged
cells.
sign. If the values are e e e, this behavior should be
investigated. I 19. Task Group Actions
17.2.2 k Graph-Here th· ri ary pattern to look for is that
I
• I
f
of one laboratory having lar e alues (or very small k values) 19.1 General-If the investigation disclosed no clerical,
for all or most of the m t ri~ls. High k values represent sampling or procedural errors, the unusual data should be
within-laboratory imprecisi n Vi ry small k values may indi- retained, and the precision statistics based on them should be
cate a very insensitive me e ent scale or other measure- published. If, on the other hand, a cause was found during the
ment problem. investigation, the task group has several options to consider. If

[2
 r) Amb)t~pyright ASTM lnternational,100 BarrHarorb '
Dm
dl~ West Conshohocken, Pennsylvania 19428· 295 (
I

the laboratory clearly and seriously deviated from the ~k

method, the test results for that laboratory must be remo OOMaras Worluw,a.
rn~~~m~m~~~~~SIbe~t>RMATION
• • • .'

from the ILS calculations. However, despite the danger of It, the oolc'to~f'~~Mwane~aMf~~~W8~liter, .
recalcitrant laboratory having prior knowledge, it may be. S
1 m A,
appropriate to ask the laboratory to retest one or more mater al~
following the correct procedure, and then include the new se"("-nas
e~liN.
1
er 10 Ge~ral-Once
~8•
ffiaJ IClentl ••
the task group has concluded which
InconsIstent to
•• .
action. and action
een ta en, the statistics of 15.4 through 15.6 are recalcu-
of test results in the ILS calculations. Of course, if the data lated (see also 20.1.5). Using the corrected statistics, calculate
have changed, recalculation of the hand k values must be made for each material the 95 % repeatability and reproducibility
and the data consistency examined again. limits (see Practice E 177) according to the following Eq 13 and
Eq 14:
19.2 Exception-When a large number of laboratories have
participated in the ILS and no cause for some unusual cell r=2,Ss, (J 3)
values have been found during the investigation, it may be R = 2.S s R ( I 4)

appropriate to delete a cell from the study if all of the other 21.2 Prepare a table for the corrected precision statistics as
laboratories are in substantial' agreement. The number of shown in Table 8.
laboratories that can be considered large enough to support
21.3 Variationof Precision Statistics with Property Level:
deletion of data without an identified cause cannot be stated
21.3.1 Quite often the values of s, and SR will be found to
exactly. Any action which results in discarding more than five
vary with the values of the property level X. This type of
percent of the ILS data likely will lead to the presentation of
response can be seen in Fig. 3, that is based on Table 8. The
precision data that the test method cannot deliver in routine
manner in which the statistics vary with the property level
application.
should be shown in presenting the precision information in the
19.3 Test Method Vagueness-One of the important things precision statement of the test method. The statistician should
to be on the alert for during a laboratory investigation is for recommend the most appropriate relationship to present, using
vagueness in the test method standard that permits a wide range Practice E 177 as a guide.
of interpretation leading to loss of precision. Particular ele-
21.4 Precision Statement-Table 8 (with the column for Sf
ments to check are lack of measurement tolerances, diversity of
omitted) is a useful format for the presentation of the precision
apparatus and insufficient direction for operator technique.
statement of the test method as required by Section A21 of the
These problems can be the basis for a revision of the standard.
"Form and Style of ASTM Standards (Bluebook)." Having
obtained the required precision information in accordance with
20. Glucose ILS Consistency, this practice, the final form of the precision statement may be
20.1 Glucose in Serum-The ILS is described in 15.1.1. prepared in accordance with Practice E 177.
20.1.1 h Statistic-The overall impression given by Fig. 21.5 Conclusion-The precision statistics obtained by an
and Table 3 is one of reasonable consistency for variation ILS such as described in this practice must not be treated as
among laboratories. Only Laboratory 4 stands out with large exact mathematical quantities which are applicable to all
values for Materials Band C. circumstances and uses. The small number of laboratories and
20.1.2 k Statistic-Laboratories 2 and 4 stand out in Fig. 2 of materials included in the usual ILS guarantees that there will
and Table 4. be times when differences greater than predicted by the ILS
20.1.3 Cells and Test Results-Cells C4 and E2 should be results will arise, sometimes with considerably greater or
investigated. A look at Table I reveals that the second test smaller frequency than the 95 % probability limit would imply.
results of 148.30 in C4 and of 309.40 in E2 are the particular The repeatability limit and the reproducibility limit should be
values to be investigated. considered as general guides, and the associated probability of
95 % as only a rough indicator of what can be expected. If
20.1.4 Action-If the data from Laboratory 4 were typed,
the result 148.30 in Cell C4 could have been a typographical more precise information is needed in specific circumstances,
those laboratories directly involved in a material comparison
error. We have no way of knowing this today, many years after
must conduct interlaboratory studies specifically aimed at the
this study was made. We will suppose, however, that the task
material of interest."
group did indeed call the laboratory and did find that the
number should have been 138.30. However, let us suppose that 22. Keywords
for Cell E2 the task group could find no explanation of the
22.1 precision; repeatability; reproducibility; repeatability
apparently high value of 309.40. In such a case they should
limit; reproducibility limit
retain the value.
20.1.5 Recalculation-Table 6 and Table 7 show the recal- 4 Following the ASTM Research Report format guide. prepare a research report
culated consistency statistics resulting from correcting Cell C4. on the ILS to be filed at ASTM Headquarters,

13
 5

4. -
o
4 - •
3., -
o

3 -
•
2. -
o
2 -
1. - 0
• •

1 - Q

.5 -
o 50 100 150 200 250 300 350 400
Averages
FIG. 3 Glueo e n erum: Standard Deviations of ReproduelbJllty (0) and Repeatability (0) Versus Average

APPENDIXES

(Nonmandatory Information)

Xl. THEORETICAL CONSIDERATIONS

X1.I Underlying Assump 0 s of ILS the average of all ILS test results for a given material, and
whose standard deviation is designated by s L' (The effect of a
X 1.1.1 Within-Laborato Vi riability-«The cell standard
single outlying laboratory on this assumption will be less if
deviation is a measure of I e w~thin-Iaboratory variability of there are enough laboratories.) For the ILS calculations, SL is
each individual laboratory. 11 laboratories are assumed to estimated from the standard deviation of the cell averages, S;"
have essentially the same I v II f variability when following
and the repeatability standard deviation, s, as follows:
the specified repeatability b~'
ions. This assumption is not ,
always fulfilled. However, horter the period of time in ., .., s;
s-=s- -- n (X 1.1)
which the test results for a p '!c lar material are to be obtained
L .e

by the laboratories the mor Ii e y the validity of this assump- The term s/ is the observed variance of the cell averages
tion. Therefore, the Iaborat cJll variances can generally be from the ILS data. The term s/In estimates the variance of cell
pooled by averaging the squ r s ~f the cell standard deviations. averages of n test results assuming that there is no difference
The square root of this av a eJwithin-Iaboratory variance is among lab means. Thus s/ indirectly estimates the variability
the repeatability standard d hti10ns.: of cell averages due to laboratory differences. When s/
Xl.l.2 Between-Laborat I Vrriability: calculates to less than zero, SL is taken equal to zero (see Note
2).
X 1.1.2.1 Variabilityof b rqtory Means-The test results
obtained on a particular m 6,ia} at any particular laboratory XI.l.2.2 Reproducibility Standard Deviation-The variance
are considered part of a po u atron having a normal distribu- among individual test results obtained in different laboratories
tion with a standard dev loti equal to the repeatability is the sum of the within-laboratory variance and the between-
standard deviation but wit a mfan that may be different for laboratory variance of the laboratory means. Thus, the repro-
ducibility variance is given by Eq X1.2 as follows:
each laboratory. The laborat leans are also assumed to vary
according to a normal distri 0, whose mean is estimated by (XU)

14
 Another calculation previously used for estimatin
reproducibility standard deviation that does not explicit
Sv as seen in previous versions of this practice, i
following.
, Substituting Eq Xl.l into Eq Xl.2 produces Eq Xl.3:
, values obtained are given in Table S. The appropriate Microsoft
(X 1.3) Excel function for such values of tis TINV(O.OOS, p-2).

Simplifying and taking the square root gives Eq X 1.4 as X 1.2.3 Within-Laboratory Consistency:
follows: X 1.2.3.1 The consistency statistic, k, is an indicator of how
one laboratory's within-laboratory variability, under repeat-
(X 1.4) ability conditions, on a particular material, compares with all of
the laboratories combined. Values of k larger than 1 indicate
When SR calculates to less than s., SR is set equal to s.: greater within-laboratory variability than the average for all
,Xl.2 ConsistencyStatistics laboratories. Since such variation among laboratories is
expected, critical values of k have been calculated to aid in the
X 1.2.1 Critical Values-The derivation of the equations for
decision of whether the cell standard deviation of one labora-
calculating critical values of hand k are given in X1.2.2 and
tory is sufficiently different from the rest of the laboratories as
XI.2.3. In each case critical values were calculated at three
significance levels, I %, O.S %, and 0.1 %. Of these three only to require investigation.
the O.S % critical values were chosen for flagging as described X1.2.3.2 A valid test for determining whether a particular
in Section 17. This choice is based on the judgment from cell variance is inconsistent relative to the variances of the
experience that the I % values are too sensitive (flag too many) other laboratories is to calculate the F-ratio of the one cell
and the 0.1 % values are not sensitive enough for flagging variance to the pooled variance of all the other laboratories-
adequately in the analysis of ILS data. excluding the variance being tested. This is shown in Eq X 1.10
X 1.2.2 Between-Laboratory Consistency: as follows:
X 1.2.2.1 The consistency statistic h is an indicator of how S2
F= C (XI.IO)
one laboratory's cell average, for a particular material, com-
pares with the average of the other laboratories. The critical [ Li;'C s~]
p-I
values for the comparison are calculated with an equation
derived from an unpaired t-test as given by Eq XI.S as follows: where:
(z - x*) s/ = cell variance of cell being compared,
t= • '
S'Yl+[I/II'-I))
(Xl.5)
Li;tc s/ = summation of all variances except the one being
compared,
where: P = the number of laboratories.
= observed Student's t value, The consistency statistic k is defined by Eq XU I and the
-& = cell average being compared,
repeatability variance by Eq X1.12 as follows:
x* average of all cell averages except the one being
* compared, k = s Ls r (Xl.lI)
s" = standard deviation of all the cell averages except the
one being compared, and (XI.12)
p = number of laboratories in the ILS.
In this relationship t has p-2 degrees of freedom. Three Combining Eq XI.I 0, Eq X 1.11, and Eq X 1.12 results in Eq
further equations are required in order to express h in terms of XI.13 as follows:
t from Eq X1.5. These follow as Eq X1.6, Eq Xl.7, and Eq
X1.8:
- p - I
(XI.13)
k-~
x* = (px - xcl/(p - I) (X 1.6) I+-F-
., ip - I)s~ p(x - xy
(s,)-= (p _ 2) [(p - I) (p - 2)] (X 1.7) X1.2.3.3 The upper critical values of k were calculated by
d Xc - x Eq X1.13 where F values are the upper 0.995th percentiles of
h=-=-- (XI.8) the F distribution with n - I and (p - I) (n - I) degrees of
s.r s.r
freedom for selected combinations of numbers of test results
Each of these equations is derived by simple algebraic (n) and number of laboratories (P). The values obtained are
operations from the definitions of symbols accompanying Eq given in Table S. The Microsoft Excel function for such values
X I.S and Table 2. Combining them with Eq X I.S results in Eq of F is FINV(O.OOS,(n-I),(n-I)*(p-I)).
XI.9 as follows:

15
 It nT~) Copyright ASTMInternational, 100 Barr Harbor Drivr
~ West Conshohocken, Pennsylvania 19428·2959
~ ~tn
SIimC1ardli WorldwIde
United States of America, This copy has been maC'
by the Directorate for Standards and Quality (STAMEQ) under lieer
X2. PENTOSAN~~~~~~n~fI\~,LE • '.11 ,.,. ; ,

X2.1 Pentosans in Pul S( ven laboratories tested nine Laboratory 7 test results are less than half the values obtained

Table X2.1. l
materials, obtaining three t.! I results per material as shown in

X2.2 h Statistic-At firs ~ n e no one laboratory is singled

by the other laboratories, while for Material I the test results
are about 10 % higher than the rest. This variation with
property level should be explored.

out for attention by Fig. X'. 0t

Table X2.2. For Material A, X2.S Action-Note that Laboratory 7 reported test results to
Laboratory 7 is different an Or Material C, Laboratory I. On three significant digits for all property levels while all the other
further inspection of the la 1< Ia~bry plot for Laboratory 7 we laboratories reported to two decimal places. This difference in
note the first five materia ~ brf negative and the last four reporting would have been a good place to start the inquiry of
positive. Keeping in mind t a t~e first five materials are close this laboratory. It might be the indication of apparatus
together in property level w i e the last two are much higher in differences, or perhaps a sign that the laboratory may have
property level, one can set t at Laboratory 7 has a different disregarded other requirements of the test method or interlabo-
response to property level )a I tile other laboratories. Labora- ratory protocol. The apparently poor within-laboratory preci-
tory 6 shows the reverse re p n e, but not as strongly. sion of Laboratory I, if determined to be due to improper test
equipment or poor maintenance of test environment might have
X2.3 k Statistic-From FIg. X2.2 and Table X2,3, It IS required omitting this laboratory's data from the analysis, but
obvious that Laboratory 1 ~i ferent from the rest with five with so few laboratories in the ILS and no physical evidence,
materials greatly exceedin I ~nt:1one near, the critical value the task group should retain this laboratory's data in the
I
line. In addition, Laborator I i different for Material H, analysis.
X2.4 Cells and Test Re f ts Both Laboratories 1 and 7 X2.6 Quite often the values of s; and SR will be found to
should be investigated in d~Il !h. Examination of the cell data
for Laboratory I in Tabl( I' 2,1 suggests special attention
vary with the values of the property level. This type of response
can be seen in Fig. X2.3 that is based on Table X2.4.
should be given to test resu ts i2 1'nA, 3 in B, 2 in C, 3 in D, 3
in E, and 1 in G, but there pt ,ea s to be an overall problem in X2.7 PrecisionStatement-Table X2.4 (with the column for
within-laboratory variabilit Pn the other hand, Laboratory 7 s.. omitted) is a useful format for the presentation of the
9
has a different problem in 0 a reeing with the other labora- precision statement of the test method as required by Section
tories at the two extremes 0 pperty level. For Material A, the A21 of the "Form and Style of ASTM Standards (Bluebook)."

TABLE X2.1 Pentosans in Pulp-ILS Test Results

Laboratory A II B c D E F G H I
1 0.44 Q.96 1.23 1.25 1.98 4.12 5.94 10.70 17.13
0.49 1.88 1.25 1.92 4.16 5.37 10.74 16.56
~.92
0.44 1.24 1.42 1.80 4.16 5.37 10.83 16.56

2 0.41
0.41
J::: 1.12
1.12
1.25
1.25
1.99
1.94
4.10
4.11
5.26
5.26
10.D7
10.05
16.08
16.04
16.13

r
0.41 1.12 1.26 1.95 4.10 5.26 9.82

3 0.51 .92 1.11 1.35 2.05 4.11 5.16 10.01 16.01

0.51 .93 1.13 1.35 2.08 4.16 5.16 10.17 15.96
0.51
, .92 1.11 1.35 2.03 4.16 5.21 10.17 16.06

4 0.40 1.15 1.29 2.05 4.20 5.20 10.98 16.65

~.96
0.38 .94 1.13 1.29 2.04 4.20 5.20 10.67 16.91

5
0.37

0.49 r 4
1.13

0.98
1.29

1.23
2.04

1.94
4.22

4.61
5.20

5.00
10.52

10.48
16.75

15.71

r
0.49 (j:~; 0.98 1.23 1.96 4.63 5.00 10.27 15.45
0.49 0.98 1.23 1.96 4.53 4.96 10.38 15.66

6 0.43 .88 1.11 1.31 2.01 3.93 4.85 9.57 15.05

0.41 (j.92 1.12 1.30 1.99 3.92 4.87 9.57 14.73
0.40 .88 1.11 1.31 1.98 3.84 4.91 9.62 15.04

7 0.186 0.~66 1.05 1.13 1.98 4.21 5.27 11.5 18.8

0.171 O.~OO 0.962 1.15 1.93 4.18 5.32 10.8 18.2
0.153 O. 31 0.927 1.15 1.98 4.16 5.10 11.5 18.1

16
 Laboratory
A B C Y H
0.63 0.36
0.46 0.35 2.05 from ASTM-' A~national.
-0.39 Q :ii e.75 -0.25
2 0.05 -1.14 -0.05 Q~~
3 0.93 0.88 -0.07 1.35 -0.18 -0.04 0.50 -0.32
4 -0.19 1.40 0.05 1.16 0.12 0.07 0.57 0.38
5 0.75 -).28 -0.94 -0.51 1.97 -0.91 -0.04 -0.69
6 0.08 0.21 -0.09 0.23 -1.37 -1.42 -1.45 -1.30
7 -2.08 -0.41 -0.94 -1.85 -0.33 om 0.21 1.54 1.84

A Critical value = 2.05.

TABLE X2.3 Pentosans In Pulp-J<A

Material
Laboratory
A B C D E F G H
1.93 2.24 '2.61 2.62 2.32 0.71 2.47 0.34 1.53
2 0.00 0.18 0.00 0.15 0.67 0.18 0.00 0.72 0.21
3 0.00 0.18 0.08 0.00 0.64 0.89 0.22 0.48 0.23
4 1.02 0.36 0.08 0.00 0.15 0.36 0.00 1.21 0.61
5 0.00 0.36 0.00 0.00 0.29 1.63 0.17 0.54 0.64
6 1.02 0.72 0.04 0.15 0.39 1.52 0.23 0.15 0.84
7 1.10 1.07 0.44 0.31 0.73 0.77 0.87 2.09 1.76
A Critical value = 2.03.

TABLE X2.4 Pentosans in Pulp-Precision Statistics

Material x S" S, SR R
A 0.4048 0.1131 0.0150 0.1137 0.04 0.32
B 0.8841 0.0447 0.0322 0.0519 0.09 0.14
C 1.1281 0.1571 0.1429 0.1957 0.40 0.55
D 1.2686 0.0676 0.0375 0.0742 0.11 0.21
E 1.9809 0.0538 0.0396 0.0628 0.11 0.18
F 4.1814 0.2071 0.0325 0.2088 0.09 0.58
G 5.1843 0.2172 0.1330 0.2428 0.37 0.68
H 10.4010 0.5630 0.1936 0.5848 0.54 1.64
I 16.3610 1.0901 0.2156 1.1042 0.60 3.09

17
 d01..) COPYright ASTM Jnternanonat, 10(1Barr Harbor Oriv,'
~t!_1
INTm ..A
West Conshohocken, Pennsylvania 19428·2951.
riONAl. U ' , '
"'ted States of America, Tnls copy nas been mao
Slo".a,a. W''''aw,",
2.5
, uy the Directorate for Standards and Quality (STAMEQ)under licen.
2 - _flQ..11l6~ntlDtematirwl.__ --.;'" --- ---- --- --- -- 2.05

1.5

1
u
:g
VI
:g .5
~
~
e
; 0 -
'iii
c
8I
-.5
~
-1

-1.5

, ------- -2.05
-2

-2.5
MAT: B I DEFGHI ABCDEFGHI ABCDEFGHI ABCDEFGHI ABCDEFGHI ABCDEFGHI ABCDEFGHI
LAB: 1 2 3 4 5 6 7
FIG. X2.1 Pentosans in Pulp: h-Materials within Laboratories

18
 1LII"'..... --:-C~Wight ASTMInternational, 100 Barr Harbor Driv

3
.::::::!UJ~ West Conshohocken, Pennsylvania 19428·295
INnRHATWNAL
Siandaras WorlOWldA
United States of America. This COpy has been mar
by the Directorate for Standards and Quality (STAMEQ) under licen
from ASTM International.
2.5 -

2
- -- - ---------------------------------------------------- - 2.03

..

MAT: ABCOEFGHI. ABCOEFGHI ABCDEFGHI ABCDEFGHI ABCOEFGHI ABCDEFGHI ABCOEFGHI

LAB: 2 3 4 5 6 7
FIG. X2.2 Pentosans in Pulp: k-Materials within Laboratories

19
 a
~.
Copyright ASTM lnternational, 100 Barr Harbor DriVE
7 W~st Conshohocken" Pennsylvania 19428· 295S
~-~OIted States of America, This copy has been mad
12 ) under licen.
" from ASTMlntern~tional. o
, 1 -

.,

0
0
0
• •
0 • •
000
0 - •••• •
I
I
I

I
I o 2 4 6 8 10 12 14 16 18 20
I AveMiges
FIG. X2.3 Pent I s In Pulp: Standard Deviations of Reproducibility (0) and Repeatability (.) Versus Average

20
 Cell Statistic 'Formula for Lab 1
Average E3=AVERAGE(B3:D3)
Standard deviation .\ from AS~ ~e~"rn:a:tio:na:\'~_...i.·F
_..
F.3=STDEV(B3:D3)
G3=E3-E$12
Deviation from material average
Between-laboratory consistency statistic -~-- H H3=G3/E$13
Within-laboratory consistency statistic I 13=F3/E$14

.s'

ASTM International takes no position respecting the validity of any patent rights asserted in connection with any item mentioned
in this standard. Users of this standard are expressly advised that determination of the validity of any such patent rights, and the risk
of infringement of such rights, are entirely their own responsibility.

This standard is subject to revision at any time by the responsible technical committee and must be reviewed every five years and
if not revised, either reapproved or withdrawn. Yourcomments are invited either for revision of this standard or for additional standards
and should be addressed to ASTM International Headquarters. Yourcomments will receive careful consideration at a meeting of the
responsible technical committee, which you may attend. If you feel that your comments have not received a fair hearing you should
make your views known to the ASTM Committee on Standards, at the address shown below.

This standard is copyrighted by ASTM International, 100Barr Harbor Drive, PO Box C700, West Conshohocken, PA 19428-2959,
United States. Individual reprints (single or multiple copies) of this standard may be obtained by contacting ASTM at the above
address or at 610-832-9585 (phone), 610-832-9555 (fax), or service@astm.org (e-mail); or through the ASTM website
(www.astm.org). Permission rights to photocopy the standard may also be secured from the Copyright Clearance Center, 222
Rosewood Drive, Danvers, MA 01923, Tel: (978) 646-2600; http://www.copyright.com/

22
 ,--'_._-
cO IE~~T~,
COPYrightASTM International, 100Barr HarborDrivf
I ~±illl!_ West Conshohocken, Pennsylvania 19428·2959
X3. SPREADSHEETJo~'it:~~;;~itJe~i\M~erica, This copy has been mad·
1j' by the Directorate for Standards and Quality (STAMEQ)under licen.
X3.1 ILS Calculations Using a Microsoft Excel Spread- from fs~~~rJatlMi~, listed in Table X3.1 below. For the other labs
sheet .. __~ formulas can be dragged down through the rows.
X3.1.1 A spreadsheet for performing Practice E691 calcu- X3.1.4 Calculation of Statistics for the Material-These
lations is shown in Fig. X3.1 and uses the Glucose in Serum calculations are made in CeIJs EI2 - EI6 in the spreadsheet
Material C example in Section 15. This ILS had eight labora- example below, and comprise the folJowing:
tories with each lab conducting three replicate test results on
each of five materials.
• EI2 Average of ceIJ averages, x
• E13 Standard deviation of the ceIJ averages, s,
X3.1.2 Data Entry-Test results are entered in ceIJs B3 - • EI4 Repeatability standard deviation, Sr .,
DIO, one row for each lab. The number of labs is entered in • E 15 Between-laboratory standard deviation, sL
B12 and the number of replicates per lab is entered in B 13, as • EI6 Reproducibility standard deviation, SR
these values are used in subsequent calculations.
X3.1.2.1 Differentnumbers of labs and replicates per lab are
accommodatedby expanding or contracting the area of the test
result entry. This shouldbe done by inserting or deleting the
middle rows and columns of the data entry area to preserve the
calculation formulas.
X3.1.3 Calculation of" Cell Statistics-Five ceIJ statistics
are calculated for each laboratory and appear in Columns E -
I for this example. The s~!l1bols,columns, and Excel formulas

A B C 0 E F G H I
;
1 Test Results x Cell Statistics
2 Lab " 1 2 3 x s d h k
3 1 132.66 133.83 133.10 133.197 0.591 -1.946 -0.73 0.22
4 2 132.92 136.90 136.40 135.407 2.168 0.264 0.10 0.79
5 3 132.61 135.80 135.36 134.590 1.729 -0.553 -0.21 0.63
6 4 138.50 148,30 135.69 140.830 6.620 5.687 2.14 2.41
7 5 131.90 134.14 133.76 133.267 1.199 -1.876 -0.71 0.44
8 6 137.21 135.14 137.50 136.617 1.287 1.474 0.55 0.47
9 7 130.97 131.59 134.92 132.493 2.124 -2.650 -1.00 0.77
10 8 135.46 135.14 133.63 134.743 0.977 -0.400 -0.15 0.36
11
12 # Labs 8 x 135.1429 E12=AVERAGE{E3 :EIOl
13 # Reps 3 Sy 2.6559 E13=STDEV(E3:EIOl
14 s; 2.7483 E14=SQRT(SUMS_illF3:FIO_}lBl~
E15= IF«E13" 2>(E14" 2)/B13),SQRT(E13" 2-
15 SL 2.1299 (E14"Z)lB13J,Ol
16 Sa 3.4770 E16=SQRT((E14"2)+(E15/\2))
17
FIG. X3.1 Spreadsheet Example for ILS Calculations

.•

21