Computer-aided selection of correct statistical tests in data analysis

Research is the systematic work of using scientific methods and techniques to reveal a truth, to solve a problem and to increase the amount of data available. The exact clarification of the data obtained from the results of research depends on the suitability of the research methods and the correct selection of the statistical tests used during the analysis. The choice of statistical tests should be based on the purpose of the study conducted and the properties of acquired data such as measurement level, variable type, distribution form and homogeneity of variance. The choice of the proper test becomes a complex problem for researchers, when there are multiple variables with different levels of measurement in study. Especially, there are some studies in literature proving that some statistical methods are used inappropriately in medical studies. There are various studies in the literature to help researchers in selecting the proper statistical test. In these studies, it has been suggested to use some questions with decision trees and tables for statistical test selection process. In addition, some statistical package programs can suggest tests to be performed after data entry. In this study, a method for statistical test selection was proposed. Using the proposed test selection algorithm, we aimed to suggest the most suitable statistical test for the research by using measurement level and sample data with the help of computer. The obtained outputs are compared with the literature and the package programs capable to suggest statistical test, and the results are given.