We had to admit that the only compulsory course for our undergraduate students in Engineering Statistics had no impact. We changed our whole approach to mimic as closely as possible the situation of the practicing statistician in the workplace. That was 15 years ago. We were the precursor (or re-inventor?) of flipped pedagogy.
We introduced active learning, with data from the industry, and case studies (including a good part of the theory) with report writing. We completed over the years our own fully hyper-referenced textbook, including Flash animations for the key concepts and historical notes. A website was used to link the class.
Most of the student's work, partly in class, was done in small groups, with much less formal teaching, the professor shifting his role from sage on the stage to guide on the side. We introduced a novel protocol for the evaluation of the course by the students.
The new approach was a success, generating a great interest for Statistics. The level of the course greatly improved and the rate of success for the course reached almost 100%.
We will discuss the advantages and drawbacks of this approach for the students and the professor.
Our work was awarded a PrixPoly1873 for pedagogical efficiency.