The Horwitz function revisited

A statistical test was made of the Horwitz function, an empirical relationship between the reproducibility precision of an analytical method and the concentration of the analyte regardless of the nature of the analyte, matrix, and the method. The large data set (7502 observations) was compiled by Horwitz from collaborative trials (method performance studies) spanning the period 1915 to 1995. The data followed the Horwitz function well down to concentrations of about 10 -8 (10 ppb), but they followed a more stringent specification at lower concentrations. This discrepancy may be due to special circumstances prevailing in collaborative trials at very low concentrations. Deviations of individual observations from the function were in large part accounted for by random variations. No consequential Improvement in precision with time was found.