Laserfiche WebLink
ASBESTOS and OTHER FIBERS by PCM:METHOD 7400,Issue 2,dated 15 August 1994-Page 10 of 15 <br /> a group of laboratories,there is a 90%probability that the mean would fall between 0.01 and 0.08 fiber/ <br /> mL.These limits should be reported in any comparison of results between laboratories. <br /> Note that the S,of 0.45 used to derive Figure 1 is used as an estimate for a random group of laboratories. <br /> If several laboratories belonging to a quality assurance group can show that their interlaboratory S,is <br /> smaller,then it is more correct to use that smaller Sr.However,the estimated S,of 0.45 is to be used in <br /> the absence of such information.Note also that it has been found that S,can be higher for certain types <br /> of samples,such as asbestos cement[9]. <br /> Quite often the estimated airborne concentration from an asbestos analysis is used to compare to a <br /> regulatory standard.For instance,if one is trying to show compliance with an 0.5 fiber/mL standard <br /> using a single sample on which 100 fibers have been counted,then Figure 1 indicates that the 0.5 <br /> fiber/mL standard must be 213%higher than the measured air concentration.This indicates that if one <br /> measures a fiber concentration of 0.16 fiber/mL(100 fibers counted),then the mean fiber count by a <br /> group of laboratories(of which the compliance laboratory might be one) has a 95%chance of being <br /> less than 0.5 fibers/mL;i.e.,0.16+2.13 x 0.16=0.5. <br /> It can be seen from Figure 1 that the Poisson component of the variability is not very important unless <br /> the number of fibers counted is small.Therefore,a further approximation is to simply use+213%and <br /> —49%as the upper and lower confidence values of the mean for a 100-fiber count. <br /> 90%Confidence Interval on Mean Count <br /> [subjective component(0.45)+Poisson component] <br /> 500 <br /> 400 <br /> D <br /> 0 <br /> V <br /> v <br /> E 300 <br /> M <br /> Ln <br /> v W"--95%probability mean count is below this level <br /> c <br /> '7n 200 <br /> 0 <br /> v <br /> a, 100 <br /> c <br /> v <br /> V <br /> i <br /> UJ <br /> a 0 <br /> 95%probability mean count is above this level <br /> —100 <br /> 0 10 20 30 40 50 60 70 80 90 100 <br /> Number of fibers counted in a single sample <br /> Figure 1.Interlaboratory precision of fiber counts. <br /> NIOSH Manual of Analytical Methods(NMAM),Fourth Edition <br />