ASTM_D_4855_-_97_2002.pdf

Designation:D 4855 97(Reapproved 2002)Standard Practice forComparing Test Methods1This standard is issued under the fixed designation D 4855;the number immediately following the designation indicates the year oforiginal adoption or,in the case of revision,the 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 practice provides a procedure for evaluating andcomparing test methods under controlled conditions using thesame materials tested during the same time span.The practicedescribes how to obtain and compare estimates on precision,sensitivity,and bias.1.2 This practice covers the following topics:Topic TitleSectionnumberScope1Referenced Documents2Terminology3Significance and Use4Requirements for Materials5Evaluating Test Methods6Sensitivity Criterion7Basic Statistical Design8Experimental Procedure9Procedure for Comparing Precision10Evaluating the Bias Between Test Methods11Procedure for Comparing Sensitivities12Report132.Referenced Documents2.1 ASTM Standards:D 123 Terminology Relating to Textiles2D 2905 Practice for Statements on Number of Specimensfor Textiles2D 2906 Practice for Statements on Precision and Bias forTextiles2E 456 Terminology Relating to Quality and Statistics32.2 ASTM Adjuncts:TEX-PAC4NOTE1Tex-Pac is a group of PC programs on floppy disks,availablethrough ASTM Headquarters,100 Barr Harbor Drive,Conshohocken,PA19428,USA.The calculations for comparing the precision,sensitivity andbias of two test methods can be done using one of these programs andstatements on the relative merits of the two test methods are part of theoutput.3.Terminology3.1 Definitions:3.1.1 accuracy,nof a test method,the degree of agree-ment between the true value of the property being tested(or anaccepted standard value)and the average of many observationsmade according to the test method,preferably by manyobservers.(See also bias and precision.)3.1.1.1 DiscussionIncreased accuracy is associated withdecreased bias relative to the true value;two methods withequal bias relative to the true value have equal accuracy evenif one method is more precise than the other.The true value isthe exact value of the property being tested for the statisticaluniverse being sampled.When the true value is not known orcannot be determined,and an acceptable standard value is notavailable,accuracy cannot be established.No valid inferenceson the accuracy of a method can be drawn from an individualobservation.3.1.2 bias,nin statistics,a constant or systematic error intest results.3.1.2.1 DiscussionBias can exist between the acceptedreference value and a test result obtained from one method,between test results obtained from two methods,or betweentwo test results obtained from a single method,for example,between operators or between laboratories.3.1.3 confidence interval,nthe interval estimate of apopulation parameter computed so that the statement“thepopulation parameter lies in this interval”will be true,on theaverage,in a stated proportion of the times such statements aremade.3.1.4 confidence level,nthe stated proportion of times theconfidence interval is expected to include the populationparameter.3.1.4.1 DiscussionStatisticians generally accept that,inthe absence of special consideration,0.95 or 95%is a realisticconfidence level.If the consequences of incorrectly estimatingthe confidence interval would be grave,then a higher confi-dence level might be considered.If the consequences ofincorrectly estimating the confidence interval are of less thanusual concern,then a lower confidence interval might beconsidered.1This practice is under the jurisdiction ofASTM Committee D13 on Textiles andis the direct responsibility of Subcommittee D13.93 on Statistics.Current edition approved September 10,1997.Published August 1998.Origi-nally published as D 4855 88.Last previous edition D 4855 91.2Annual Book of ASTM Standards,Vol 07.01.3Annual Book of ASTM Standards,Vol 14.02.4PC programs on floppy disks are available through ASTM.For a 312inch diskrequest PCN:12-429040-18,for a 514inch disk request PCN:12-429041-18.1Copyright ASTM International,100 Barr Harbor Drive,PO Box C700,West Conshohocken,PA 19428-2959,United States.3.1.5 confidence limits,nthe two statistics that define theends of a confidence interval.3.1.6 degrees of freedom,nfor a set,the number of valuesthat can be assigned arbitrarily and still get the same value foreach of one or more statistics calculated from the set of data.3.1.6.1 DiscussionFor example,if only an average isspecified for a set of five observations,there are four degrees offreedom since the same average can be obtained with anyvalues substituted for four of the five observations as long asthe fifth value is set to give the correct total.If both the averageand the standard deviation have been specified,there are onlythree degrees of freedom left.3.1.7 error of the first kind,a,nin a sta