Impact van de gebruikte statistische methode op een kleine dataset

Abstract

As a researcher you have to decide in advance what statistical method you use to analyse your data. When dealing with small sample sizes, decisions in statistical method can have a major impact on the results. In this thesis six exact datasets were generated to investigate the differences in results. Each dataset consisted of four groups. The datasets differ in effect size (small/medium/large) and power (0.4 & 0.8) and were analysed using: ANOVA, Contrast test, F-bar test, Non parametric (K&W-test), bootstrap and Bayesian statistics. The results showed that the method used could be of great impact on the final conclusion. Especially the methods where an informed hypothesis was tested, stood out in a positive way. Based on this study we recommend that if you have information about your data in advance, you should try to implement it in the chosen statistical method before testing.