Page 120 - Getting the Picture Modeling and Simulation in Secondary Computer Science Education

118
Chapter 5
5.6 Discussion and Conclusion
We designed and investigated an assessment instrument for the assessment of the intended learning outcomes for Computational Science. The design process was all but straightforward due to the fact that some ILO’s of modeling are at the core of CS (e.g. implementation of the model), while others are not often seen in a CS classroom (e.g. experiment). Even for implementation, which comes down to programming, it was not easy to find related work addressing assessment of programs at just the right level of abstraction. The same holds true for validation: while there is plentiful literature on validation of computational models, we could not find any focusing on the assessment of validation in a formal learning setting.
The project documentation and program code proved to be sufficient sources for assessment. However, when possible, we suggest to let students present their projects in the classroom too, and we encourage the teachers using this assessment instrument to take into account their observations of students at work when assessing their projects, as suggested by a number of teachers participating in one of our previous studies (Grgurina et al., 2017). Indeed, the teacher who cooperated in this research noted that, while assessing his students’ work, he constantly thought of his impressions from the classroom and wanted to take these impressions into account. This might be especially important for students who perform poorly when verbalizing their thoughts, as witnessed with many HAVO students in the parts of the assignment requiring textual descriptions such as stating the case and research question. We saw that none of these students achieved extended abstract level, while in the 12th grade VWO one teacher found four instances of student groups reaching it. The other teacher, however — while assessing the same projects — found none and said, “it was difficult to see clearly where the boundary lies between relational and extended abstract levels.” Therefore, it could be argued that the extended abstract level is unobtainable for HAVO students, which would signify a situation similar to the one described by Castro and Fisler who found no instances of extended abstract level in their students’ work (2017). Meerbaum-Salant et al. (2013) did not consider it at all and designed assessment with only the three intermediate SOLO categories to monitor novices’ learning of CS concepts. An issue to consider here is the question, what level of understanding is intended for the HAVO students, as opposed to the VWO students, and with what purpose are the students learning about Computational Science. The HAVO students are following education stressing a