[Examples of pitfalls in statistical analysis--1. Summarizing your data before statistical analysis].

Abstract

When conducting studies, it is important to summarize the gathered data before beginning calculations for statistical analysis. If the methods used to summarize the data are inadequate for purposes of the analysis you want to perform, many misleading or unreasonable discussions that are normally avoidable arise. If there is any uncertainty in the way data are handled, for example, confusion as to whether comparison of mean (parametric analysis) or comparison of median (nonparametric analysis) is used, or whether the data are paired or unpaired, it is impossible to make good charts that show us the features of the data at a glance. In this article, I will address several frequently asked questions about summarizing data, and discuss making suitable charts by presenting a typical example.