Dr. Margarita Lambova, Assoc. Prof.
University of Economics – Varna
Abstract. Thoughts and considerations are presented about the term “statistical significance” – a common tool for “proving” the compatibility with practice of theoretical models designed on the ground of assumptions.
In accordance with the set goal and based on the theoretical features and logics of the concept, there are revealed problematic issues and possibilities for abusing it when verifying statistical hypotheses. The suggested argumentation enables one to claim that the “prescriptive” verification of statistical hypotheses creates conditions for developing a pseudoscience supported by incorrect conclusions which are made on the ground of “statistically significant” results.
Keywords: statistical significance; statistical hypothesis testing; p-value; significance level; pseudoscience