Title: Application of Statistical Techniques to Environmental Radiation Surveillance
Abstract: AN environmental surveillance program, such as that operated by the Division of Radiological Health, Public Health Service, a primary function is the analysis and measure¬ ment of radionuclide concentrations in a variety of samples.Typical media analyzed under this program include air, water, milk, selected food items, and various types of biota.As with any other experimental work, there is an error associated with each determination.The pur¬ pose of this paper is to review some of the statistical techniques employed to estimate and minimize these errors. Sources of ErrorSampling.Since, in many instances, the data obtained by radiochemical assay are used to estimate the radioactive intake of the general population, the samples must be represent ative.Although the selection of samples for surveil¬ lance programs is usually beyond the immediate control of the analytical staff, it is important to realize that invalid sampling can be a major source of error.In the Pasteurized Milk Network of the Public Health Service, collection procedures were designed to obtain reasonably representative samples.In this program, collectors in 62 major U.S. cities submit weekly 1-gallon milk samples, which represent 80-100 percent Mrs. Wall is a statistician, Analytical Services Pro¬ gram, Northeastern Radiological Health Laboratory, Public Health Service, Winchester, Mass. of all fluid milk consumed in that city.Ideally, the sample is collected by drawing from each major milk processing plant a volume propor- tional to its share of the market.Table 1 illustrates the operation of this method for Cincinnati in March 1963.Similar considera¬ tions are applied in the selection of samples for other networks operated by the Public Health Service.