The Use of Transformations in the Analysis of Field Data

Statistics is used in biological experiments to summarize and evaluate observations in order to test a principle or hypothesis. However, if misused these statistical procedures can be misleading and cause researchers to reach conclusions that are not justified. Statistics have been widely used in agricultural research to identify and understand the distribution of organisms in space and to determine the effects of treatments on yield, quality and pest control. Confidence limits of statistical tests are based on populations that have a normal distribution. However, not all biological data follow this distribution. Once the population distribution has been identified, specific transformations of the data are essential in order to accurately interpret field results. The purpose of this report is to describe a statistical technique to identify population distributions most commonly associated with biological data and to suggest the appropriate transformation to use on the data prior to running the analysis of variance and establishing confidence limits.