The motivational value for students of problem-based immersion in the process of data collection,<br>data analysis and interpretation, is accepted by many. However, the culture of instruction<br>through technique-based courses is still used at the tertiary level in many universities. The<br>coverage of topics seems to trump guidance through the process of data analysis. In this paper, I<br>suggest how to complement a problem-based experiential presentation of statistical methods with<br>a presentation of the abstract structures necessary for future applications. A series of problembased<br>courses might fail to highlight the general and transferable concepts and principles that<br>help to bring coherence to the toolbox of statistical techniques. To overcome this shortcoming<br>one can present the logical structure - that is definitions, strategies, theoretical frameworks and<br>justifications - to unify the collection of problem-specific methods, but only after extensive<br>immersion in practical problems. Once students have experienced the effectiveness of the<br>practical statistical approach, they may be better prepared to absorb the abstract generalizations.
The CAUSE Research Group is supported in part by a member initiative grant from the American Statistical Association’s Section on Statistics and Data Science Education