Page 145 - Getting the Picture Modeling and Simulation in Secondary Computer Science Education
P. 145
Students’ Understanding and Difficulties
they employ? — we characterized the students’ understanding in terms of these elements:
• Construct the model upon assumptions resulting from research and abstraction process.
• Test the model: generate, observe and interpret outcomes.
• Reflection after the model is built: plausibility, accuracy, credibility
and satisfaction.
In answering our second research question — What difficulties do the students
encounter when verifying and validating their models? — we observed a number of issues and problems:
• Performing the research necessary to build the models does not always happen. Erroneous perspectives are sometimes employed in the process of abstraction. Omissions are being made during the implementation of the model.
• When testing the model, a systematic approach to generating outcomes is seldom employed and is lacking when observing and interpreting the outcomes.
• Finally, during the reflection upon the models, there is satisfaction with a clearly unrealistic model, and even cases of not understanding the essence of modeling altogether.
6.5.2 Reflection on the Findings 6 Next to the validation aspects we observed, it is interesting to mention what
we did not observe.
A category of validation techniques we did not observe in this study has to
do with numerical aspects of modeling. None of the students reported extreme conditions test — making sure the models’ outputs were plausible in extreme conditions (Sargent, 2013). Also, none of them used historical data to not only calibrate the model, but also to check whether the model behaves as the real system (Sargent, 2013). No-one performed predictive validation, i.e. used the model to forecast the outcomes and then compared those forecasted outcomes to the behavior of the real system (Sargent, 2013) either. Finally, no student used statistical tests or other appropriate techniques to objectively interpret the outcomes their tests. We could speculate about the reasons students did not engage in these techniques. It is plausible to think that it was difficult for them to get sufficient real data. Furthermore, their understanding of the phenomena they modeled was rather limited, which is not strange considering the position
143