Ratio data and the quantification of drug effects

Abstract

Analysis of data which have been arranged in some form of a ratio is frequently employed by several disciplines which use pharmacological agents as independent variables. When the denominator term of the ratio is also a measured variable, it has been shown that correlations are induced between variables sharing the common denominator. Therefore, when data from drug conditions are calculated as ratios or percentages of data from non-drug controls, statistical analyses may produce misleading results due to the mathematical properties of ratios. Using data from a pharmacological experiment, comparisons were made between analysis of data in ratio form and analysis of data where the denominator variable is taken as a concomitant variable. The results indicate that analyses using ratios decrease the sensitivity of the statistical test. Alternatives to the use of ratios are discussed.