CV, RSD, and the HorRat - applications and limitations
P. WEHLING (1)
(1) General Mills, Inc., Minneapolis, MN, U.S.A.
Cereal Foods World 55:A6
Relative Standard Deviation, or RSD, (sometimes referred to as CV), is a combination of 2 other statistics: mean and standard deviation. When combining statistics, it is important to be sure that the combination results in a metric which adds additional understanding to the interpretation of the data set. Additionally, all reported statistics should be understood by the audience, and the audience should have a contextual framework to interpret the magnitude of the statistics. RSD was originally used as a way to aid the interpretation of the observed standard deviation of a test method. Subsequently, William Horwitz and Richard Albert developed the Horwitz equation relating RSD of the method to analyte mean concentration. This function has greatly aided the interpretation of validation data, as it gives a prediction of RSD for a given analyte concentration. Calculation of a Horwitz Ratio, or “HorRat” from multi-lab validation studies has been used effectively as a metric to guide interpretation of collab results. Since the Horwitz equation is only applicable to measures of concentration in mass ratio units, the HorRat would not be applicable for non-concentration-based methods, such as pH or color tests. In recent years, biochemical assays have shown startlingly high HorRat values for some analytes. The cause of this is likely due to the fact that the chemical species being detected by these methods are a very small fraction of the total mass of the analyte being quantified. The expected RSD for such analytes would be higher than anticipated by the Horwitz function. Application of the HorRat to these types of analyses should be done with caution.
