The sources of variance in the results of samples collected during a pilot-scale remediation treatability testing for PCBs were evaluated using basic statistical methods. The relative contributions of soil heterogeneity and chemical testing methods were estimated. The sources of variance were estimated using both inter- and intralaboratory testing and the determination of the total variance of each sample set.
It was found that 67 to 94 percent of the variance in the results was contributed by the sample heterogeneity which can not be readily controlled. The results of treated samples from the pilot-scale treatability testing showed a mean concentration of PCBs of 13.6 ppm with a standard deviation of 10.1. The range of the distribution of concentrations was 1.3 to 35 ppm.
Based on these findings, it is inappropriate to use an “absolute” clean-up standard to evaluate whether the clean-up has been achieved. The use of the mean is more appropriate as it takes into account the variance in the results caused by sample heterogeneity.