Page 157 - Getting the Picture Modeling and Simulation in Secondary Computer Science Education

as correct too, and have not performed additional testing or formal verification to examine their correctness as doing so was beyond the scope of this research project.
In our final study (chapter 6), we explored students’ understanding and difficulties while working on Computational Science assignments using the teaching materials we developed ourselves. We focused on their understanding and difficulties concerning verification and validation of the models they develop. We characterized their 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.
Looking at the difficulties the students encountered when verifying and validating their models, we found that:
• Not all students explicitly engaged in research necessary to build their models. In the process of abstraction, some students employed erroneous perspectives. During the implementation of the model, the students reported making omissions.
• When testing the model, a systematic approach to generating outcomes was employed only by some of the groups and was lacking when observing and interpreting the outcomes.
• When reflecting upon their models, some students were satisfied
with a clearly unrealistic model, and some appeared not to 7 understand the essence of modeling altogether.
To discuss our findings, we note that students’ difficulties with validation of their models have — to our best knowledge — barely been explored in the context of computer science education. When Louca et al. (2011) asked their students to construct computational models, they only assessed the surface structure of the implemented models, and only labeled them as correct or incorrect. In mathematics education, difficulties with validation of models were explored more extensively (Eraslan & Kant, 2015), as was satisfaction with unrealistic models that was reported in students engaging in mathematical modeling too (Edo et al., 2013; Maaß, 2006). The scientific value of our approach stems from the distinguishing characteristic of our studies: they took place in a CS class where
General Conclusions and Discussion
155