1、Designation: D 3670 91 (Reapproved 2007)Standard Guide forDetermination of Precision and Bias of Methods ofCommittee D221This standard is issued under the fixed designation D 3670; the number immediately following the designation indicates the year oforiginal adoption or, in the case of revision, th
2、e year of last revision. A number in parentheses indicates the year of last reapproval. Asuperscript epsilon (e) indicates an editorial change since the last revision or reapproval.1. Scope1.1 This standard provides guidance to task groups ofCommittee D22 on Sampling and Analysis of Atmospheres inpl
3、anning and conducting collaborative testing of candidatemethods.1.2 It is intended for use with other ASTM practices for thedetermination of precision and bias.1.3 It is applicable to most manual and automated methodsand to most components of monitoring systems. It is recog-nized that the evaluation
4、 of monitoring systems may providespecial problems. Practice D 3249 should be considered forgeneral guidance in this respect.1.4 It is directly applicable to chemical methods and inprinciple to most physical methods, sampling methods, andcalibration procedures.1.5 The processes described are for the
5、 general validation ofmethods of test. A user has the obligation and responsibility tovalidate any method it uses for a specific application and todemonstrate its own competence in the use of validatedmethods.2. Referenced Documents2.1 ASTM Standards:2D 2777 Practice for Determination of Precision a
6、nd Bias ofApplicable Test Methods of Committee D19 on WaterD 3249 Practice for General Ambient Air Analyzer Proce-duresE 177 Practice for Use of the Terms Precision and Bias inASTM Test MethodsE 180 Practice for Determining the Precision of ASTMMethods for Analysis and Testing of Industrial and Spe-
7、cialty ChemicalsE 691 Practice for Conducting an Interlaboratory Study toDetermine the Precision of a Test MethodE 1169 Guide for Conducting Ruggedness Tests3. Terminology3.1 The terms used in this practice are consistent with thosedefined in Practices D 2777, E 177, E 180, and E 691.3.2 Definitions
8、:3.2.1 accuracythe degree of conformity of a value gen-erated by a specific procedure to the assumed or accepted truevalue. It includes both precision and bias.3.2.2 biasa systematic (nonrandom) deviation of themethod average value or the measured value from an acceptedreference value.3.2.3 candidat
9、e methodan analytical method or measure-ment process being considered for standardization.Amethod isa “candidate” until completion of all phases of the consensusprocess specified by ASTM regulations for a proposal, anemergency standard, or a standard.3.2.4 collaborative testan interlaboratory study
10、of a testmethod wherein the participants analyze or make measure-ments on sub-samples of the same test material. If the testmethod includes the sampling of atmospheres, the participantsshould sample the same test atmosphere, as possible.3.2.5 laboratory biassystematic differences between thetrue val
11、ue and a value reported by a laboratory due to errors ofapplication such as losses, contamination, miscalibration, andfaulty manipulations, for example.3.2.6 method biassystematic departures of the limitingmean from the true value of the parameter measured, caused byphysical or chemical phenomena in
12、herent in the methodology.3.2.7 over-all precisiona value including components ofwithin-laboratory and between-user variability.3.2.8 precisionthe degree of mutual agreement betweenindividual measurements using an analytical method or mea-surement process. In practice, the standard deviation of anen
13、tire array of reviewed and acceptable data is calculated toprovide the value to be stated as the precision of the method.1This guide is under the jurisdiction of ASTM Committee D22 on Air Qualityand is the direct responsibility of Subcommittee D22.01 on Quality Control.Current edition approved March
14、 1, 2007. Published April 2007. Originallyapproved in 1978. Last previous edition approved in 2001 as D 3670 - 91(2001).2For referenced ASTM standards, visit the ASTM website, www.astm.org, orcontact ASTM Customer Service at serviceastm.org. For Annual Book of ASTMStandards volume information, refer
15、 to the standards Document Summary page onthe ASTM website.1Copyright ASTM International, 100 Barr Harbor Drive, PO Box C700, West Conshohocken, PA 19428-2959, United States.3.2.9 ruggedness testa factorial test designed to explorethe sensitivity of the method to variations in the procedure (seeYoud
16、en and Steiner, 1987).33.2.10 single-operator precisiona measure of the replica-tion of repeated measurements obtained by a single operator ona given sample.3.2.10.1 DiscussionOther classifications of precisionwhich are useful in evaluating a method, a measurement, orperformance within a single labo
17、ratory are: multioperatorprecision, single or multi-apparatus precision, and single ormulti-day precision.3.2.10.2 DiscussionThe terms “repeatability” and “repro-ducibility” are not standardized, but have generally come tomean “single-laboratory-operator-material precision” and“multi-laboratory-mult
18、i-operator-single material precision,”respectively. Such usage is maintained in the text of thispractice.3.2.10.3 DiscussionFurther classifications of bias whichare useful in evaluating performance are: operator bias, appa-ratus bias, and day bias.4. Summary of Guide4.1 Data supporting a statement o
19、f single-operator repeat-ability is the entrance requirement for any candidate method tobe considered for standardization by Committee D22. The taskgroup to which a candidate method is assigned will review itfor adequacy in this respect, and conduct further tests asnecessary to evaluate its precisio
20、n and bias, as technicallyfeasible. A method may be accepted as a proposed method,provided the repeatability is known or has been ascertained andprovided all other criteria for acceptance have been met.Independent tests by at least three laboratories shall be requiredto substantiate the repeatabilit
21、y of a method before it attains thestatus of a standard method. Collaborative testing by at leastfive laboratories to estimate the interlaboratory bias and, ifapplicable to evaluate the methods inherent bias with respectto the “true” value is needed for all standard methods and mustbe accomplished w
22、ithin 5 years of its initial issuance as astandard, if such testing has not already been done. Failure tosubject such methods to appropriate collaborative testing,constitutes valid grounds for disallowing its reapproval as astandard.4.2 Procedures that may be used in collecting the requireddata are
23、given with particular emphasis upon the applicabilityto analysis of atmospheres. Documentation requirements areestablished. Terms that are useful in expressing statements ofprecision and bias are presented.5. Significance and Use5.1 The objective of this standard is to provide guidelines toCommittee
24、 D22 for the evaluation of the precision and bias, orboth, of ASTM standard methods and practices at the time oftheir development. Such an evaluation is necessary to assurethat a cross section of interested laboratories could perform thetest and achieve satisfactory results, using the method aswritt
25、en. It also provides guidance to the user as to what levelsof precision and accuracy may be expected in such usage.5.2 The write-up of the method describes the media forwhich the test method is believed to be appropriate. Thecollaborative test corroborates the write-up within the limita-tions of the
26、 test design. A collaborative test can only userepresentative media so that universal applicability cannot beimplied from the results.5.3 The fundamental assumption of the collaborative test isthat the media tested, the concentrations used, and the partici-pating laboratories are representative and
27、provide a fair evalu-ation of the scope and applicability of the test method aswritten.6. General Policy6.1 This section describes the general policy to be followedby Committee D22, its subcommittees, and task groups in thedevelopment of ASTM standard methods and practices. Theobjective of Committee
28、 D22 is to develop fully evaluatedstandard methods and practices as far as possible. In caseswhere this is not expedient, proposed methods, as defined in6.2, may be developed. In each case, an appropriate task groupshall have the responsibility to critically examine the method orpractice, conduct ev
29、aluation tests by round robins or othertechniques including ruggedness tests, and to recommend it, ifmeritorious, for subcommittee balloting. No method or practiceshall be released and recommended for balloting unless theprecision or accuracy requirements, or both, as set forth in thefollowing, have
30、 been satisfied.6.1.1 Collaborative testing by D22 is the preferred methodof validation. Data obtained by collaborative testing by othersmay be used in lieu of D22 testing, provided that such testingwas equivalent to ASTM approved procedures. In either case,a copy of the test procedures and data mus
31、t be filed in aresearch file maintained at ASTM for such purposes.6.2 Proposed MethodA proposed method is one that hasfound favorable usage in a specific laboratory, or has been usedby several laboratories, but has not yet been standardized. Ineach case, the test method is submitted by its proponent
32、s toCommittee D22 for standardization.6.2.1 The minimum requirement for balloting of a proposedmethod shall be the inclusion in it of a single laboratorysstatement of single-operator precision, together with support-ing experimental data. Test methods meeting this requirementwill be referred to a Ta
33、sk Group, following procedures estab-lished by Committee D22.6.2.2 The experimental data needed to support a proposalmust reflect a test of the method as a whole, that is, sampling,apparatus, reagents and, calibration, and must use a procedurethat is essentially identical to that described in the pr
34、oposal.Any significant deviations between the procedure used togather the data and the proposed procedure shall be clearlyidentified.6.2.3 If such data are missing or inadequate, but the methoditself is considered by consensus of Committee D22 to beworthy of further study, a task group may be assign
35、ed toconduct experimental studies or enlist the services of at least3Youden, W. J. and Steiner, G. H., Statistical Manual of the Association ofOffcial Analytical Chemists, AOAC International, 481 North Frederick Ave., Suite500, Gaithersburg, MD 20877-2417, 1987.D 3670 91 (2007)2one competent laborat
36、ory to obtain the data upon which to basea statement of single-operator precision.6.3 Standard MethodInitial AcceptanceA method thathas found favorable acceptance and for which the within-laboratory repeatability has been verified by a multilaboratorytest program, shall be examined by the task group
37、 for compli-ance with the following requirements.6.3.1 An initial minimum requirement for establishing astandard method is a statement of within-laboratory precisionbased on data from three laboratories similar to that describedin 6.2.1-6.2.3.6.3.2 If the method purports to measure the concentration
38、 ofa substance, an investigation of the bias of the method bycomparison with a standard must be made by at least onelaboratory and the results included in an accuracy statement.6.3.3 A standard can only be carried under the provisions of6.3 for five years. Conditions for reapproval are specified in6
39、.4.6.4 Standard MethodReapprovalAstandard method maybe retained if it has found extensive use and between-laboratory precision data have been obtained. Before doing acollaborative study, a ruggedness test should be performed byat least one laboratory. (see Guide E 1169)6.4.1 The minimum requirement
40、for retaining a standardmethod shall be a statement of the between-laboratory preci-sion of the method as established in a collaborative testincluding at least five participants.6.4.2 If a bias statement is appropriate for the method, thedata supporting the statement should be obtained by at leasttw
41、o laboratories. At least one such test shall include theintroduction of potential interferences.6.5 In all testing, the minimum number of participantsshould be exceeded to the extent possible. The statistical powerof collaborative testing is greatly enhanced as such numbersare increased. The possibi
42、lity of invalidation of a test byoutliers or missing data is also minimized.7. Sample Requirements7.1 The precision and bias of test methods are typicallyevaluated by the data obtained in the measurement of testsamples. The extent to which such measurements can be madeis dependent upon the availabil
43、ity of test samples of adequatestability and homogeneity. The scope of interest of CommitteeD22 is wide, ranging from contaminants at the parts-per-billionlevel up to several percent. Particulate concentrations exist atsimilar concentration ranges and measurements of radioactivityextend the level ev
44、en lower. The variety of substances ofinterest range from simple inorganic constituents to complexorganic molecules. Accordingly, it is not possible to set forthrigid sample specifications, but only to delineate guidelines fortest sample preparation. Each method should be tested withactual samples f
45、or which it is applicable, or as close asimulation as possible. The degree of evaluation will, ofcourse, depend on the simulation achieved, and the statementsof precision and accuracy must define the test conditions.7.2 The ideal test sample is the actual atmosphere for whichthe method is intended.
46、However the use of such offerscomplications because the composition may not be known atthe moment of test and furthermore may undergo changeduring the tests. Because actual atmospheric samples cannot becollected and stabilized for long periods of time, two proce-dures are acceptable. Reproducibility
47、 and repeatability may beevaluated by simultaneous measurement by participating labo-ratories sampling the same atmosphere at substantially thesame time. Alternatively, comparison of a candidate methodwith a standard method of known precision and bias willconstitute an acceptable technique for evalu
48、ation of precisionand accuracy. Such measurements made by several laboratoriesmay be statistically treated to evaluate the reproducibility ofthe candidate method. In this latter case, the measurementsneed not be made at the same place and time by the collabo-rating laboratories.7.3 A test sample or
49、series of test samples that are stableduring the period required to perform a limited series ofmeasurements are adequate for evaluation of single-operatorprecision to satisfy the requirements for consideration as aproposed method. Three levels of concentration are recom-mended, with such levels sufficiently well established todetermine whether, and to what extent, the repeatability isdependent or independent of concentration level.7.4 A series of test samples of at least three concentrationlevels, and available in sufficient number, is required for use bycollaborat
