Answer to Question #11980 Submitted to "Ask the Experts"
The following question was answered by an expert in the appropriate field:
I am looking for guidance regarding the need to maintain actual net results (including negative numbers) for environmental sample data, especially for long-term understanding of impact and/or for use in calculating dose increase caused by an activity. The practice of not allowing negative net results obviously biases all the data in the positive direction. It is my recollection that a National Council on Radiation Protection and Measurements (NCRP) document addressed this, but after some looking I could not find it. I would appreciate help in finding guidance on this topic.
You raise a point that deserves attention, particularly as it relates to analysis of radioactive samples whose activities may be low. Your observation is exactly correct—excluding negative net results from documented data biases such results to conclude falsely that the pool of samples being considered is more radioactive than is the actual case or that net radioactivity is present when in reality no net radioactivity is present. As you are aware, if none of the samples contained radionuclides of interest in excess of that typical of background samples, then when analyses of reasonably large numbers of such samples are made to determine net count rates and associated net activities, equal numbers of negative and positive values would be obtained, and they would be spread around the expected mean value of zero. I have discussed this briefly in at least one other question on this site (see particularly the last couple of paragraphs of the answer to Q10989).
Unfortunately, many individuals, groups, agencies, etc., have adopted the procedure of reporting results in ways that can mislead or, at the least, not provide sufficient information for interested readers to make informed decisions about the significance of the data. Not reporting negative values is one such approach. Another, worse yet, is to report a value only if exceeds predetermined minimum detectable quantity—e.g., minimum detectable activity (MDA). When this is done it is possible to wrongly conclude that no net radioactivity is present when it actually is. Based on the usual counting statistics, it is the predetermined critical level that specifies the level for defining whether net activity is present in a given sample (see the previously cited Q10989 for more on this). This value is sometimes used for reporting purposes but, as we have discussed, this does not provide the full information that would prevail if negative values were also included.
This has been recognized by some groups that have influence in establishing recommended procedures for radioactivity analysis. The NCRP is one organization that has made such recommendations. Their excellent report No. 58, A Handbook of Radioactivity Measurements Procedures, first published in 1968, discusses this near the end of section 7.1.3 in which they note that the concept of MDA is often misused when activities at or below the MDA are reported as zero or less than MDA. The section concludes by saying "Negative values (below background) are as valid a measurement as positive values and should be retained in any series of data."
George Chabot, PhD, CHP
National Council on Radiation Protection and Measurements. A handbook of radioactivity measurements procedures. Bethesda, MD: National Council on Radiation Protection and Measurements; NCRP Report No. 58, 2nd ed.; 1985.