A complex concept about statistical inf...

Carranza and Kuzniak (2008) analysed the negative effects of reducing probability to a purely frequentist notion on the students’ perception of the methods to learn. Probability is incomplete and can not be fully understood unless it is linked to statistical inference. Statistical inference usually is reduced to very simple methods and, more recently, attempts have increased to follow the ideas of informal inference, which reduces it even to a mere significance test based on simulation and a narrow interpretation of probability as relative frequencies. This informal approach is favoured by its simplicity. Nevertheless, there still is a need for innovative studies on teaching and learning probability and statistical inference from a broader perspective, which would also make it possible to integrate Bayesian methods into inference, partly based on the Bayesian formula. We develop the different views (Bayesian and classical inference) in parallel – so that students can understand either of them better. To support this challenging task, we integrate philosophical considerations, use paradoxes, and make extensive use of computers. We build on encouraging feedback from our courses to date. In Hungary, we are currently in the process of reforming the curriculum, in which we will try to find ways to implement our ideas sustainably, especially among in-service teachers. Of course, our reflections will also be revised with regard to the feedback of the experimental teachers involved.