Reporting Low-level Analytical Data

Abstract

Low level measurements - those near, at, or below limits of measurements (specification, quantification, detection) - should be reported as they appear: positive, negative, or zero, as the best estimate of the measured characteristic, usually concentration. The long-term mean of such poor estimates will often provide a surprisingly good estimate of the "true value" from the operation of the law of large numbers. This manner of reporting data near the limits of measurement can be independent of, and adaptable to, any set of agreed-upon definitions and parameters for the concept, including its uncertainty. This proposal is intended to help to standardize the method of reporting low-level analytical data, not to legitimize the data or any conclusions or actions resulting from their use or interpretation.


South African Journal of Chemistry Vol.53(3) 2000: 206-212