How to choose a statistical method ―Basic edition―

Abstract

Statistical processing of data is necessary when planning the implementation of a clinical study, submitting a paper, or presenting results at a conference. In such situations, it is often difficult to determine the statistical method to use. This study outlines a basic method for selecting an appropriate statistical method from the viewpoint of type of analysis (univariate, multivariate), whether the data is paired or unpaired, type of variable (continuous, ordinal, nominal), type of distribution (parametric, non-parametric), number of groups, and number of samples.