Abstract:

The application and advance on methodologies is a common objective in recent teaching research studies. The development of new teaching strategies is mainly focused on getting that the students not only increase their knowledge, but also improve their professional capabilities. Nowadays, teaching Statistics benefits from the powerful computation of large amounts of data in software environments. In particular, the Simulation-Based Learning becomes easy to implement in teaching programs of the Statistics subject. The free software R is an useful tool for simulation studies, mainly due to its effective graphical component.
In the statistical framework, understanding the theoretical concepts associated with statistical hypothesis testing is in general a difficult task for the students. Frequently, the students do not capture in a proper way the statistical reasoning related to the criterion to solve the tests. The aim of this work is to design a simulation-based methodology to teach the statistical criterion of the hypothesis tests by using simple parametric examples. The suggested methodology is based on the graphical analysis of the so-called type I and type II errors of the tests obtained from simulated data with R. Allowing the students to experiment by running self-made simulations, they can better understand the difference between the errors, the impossibility of controlling both errors at the same time, and finally the statistical reasoning which leads to the conclusion of the test.

Keywords:

Teaching Statistics, Hypothesis Testing, Simulation-based Learning.