From question to conclusion : the effect of statistical analysis choice

Abstract

Many people assume that as long as the data are collected well and the results are significant, reliable, and valid, then the original research question was answered. The problem with this assumption is that a data set can yield different results and conclusions depending on the type of hypotheses that are made and the statistical analyses that are chosen. This paper provides a demonstration of what can happen when the same data set is analyzed using different statistical methods (ANOVA, Regression, and MANOVA). Slightly different results were obtained that led to different conclusions about the variables in question.