Logarithmic Transformations, a Great Help to Statistical Analyses

Abstract

Non-linear relationships in clinical research are often linear after logarithmic transformations. Also, logarithmic transformation normalizes skewed frequency distributions and is used for the analysis of likelihood ratios. Basic knowledge of logarithms is, therefore, convenient for a better understanding of many statistical methods. Almost always natural logarithm (ln), otherwise called Naperian logarithm, is used, i.e., logarithm to the base e. Log is logarithm to the base 10, ln is logarithm to the base e (2.718281828). This chapter is for showing that knowledge of logarithms is helpful for understanding many statistical methods. Basic knowledge of logarithms is convenient for a better understanding of many statistical methods. Odds ratio tests (Chap. 44), log likelihood ratio tests (Chap. 46), Markov modeling (Chap. 55), and many regression models use logarithmic transformations.