Page 105 - Getting the Picture Modeling and Simulation in Secondary Computer Science Education

Assessment of Modeling and Simulation
offer six suggestions for assessing computational thinking via programming, among others to make assessment useful to learners, to incorporate creating and examining artifacts, and to have the designer illuminate the whole process (Brennan & Resnick, 2012). These views are corroborated by the findings in our prior study on CS teachers’ pedagogical content knowledge (PCK) of modeling and simulation, where we learned that the interviewed teachers mostly suggest a hands-on approach to learning and that the preferred assessment form for most of them would be a practical assignment lasting several weeks, where student groups would construct models and use them to run simulations and conduct research while extensively documenting the whole process. At the same time, we observed a great diversity in the assessment criteria teachers mentioned, but very few corresponding quality indicators used to judge to what extent these criteria are met (Grgurina et al., 2017).
In the eyes of the students, the assessment defines the actual curriculum,
according to Biggs and Tang (2011). In their constructive alignment network,
the curriculum is stated in the form of clear intended learning objectives (ILO)
specifying the required level of understanding, the teaching methods engage 5 students in doing things nominated by the ILO’s, and the assessment tasks
address these ILO’s. Learning outcomes can be classified using the Structure of the Observed Learning Outcome (SOLO) which describes the learning progress through five levels of understanding. The first three levels — prestructural, unistructural and multistructural — are considered to be quantitative in the sense that prestructural indicates missing the point, unistructural means meeting only a part of the task and multistructural shows a further quantitative increase in what is grasped: “knowing more”. Relational, on the other hand, indicates a qualitative change indicating conceptual restructuring of the components — “deepening understanding”, and extended abstract takes the argument into a new dimension (Biggs & Tang, 2011).
Looking into the use of SOLO taxonomy to assess the novice programmers’ solutions of code writing problems, Whalley et al. (2011) noted that previous research had indicated difficulties in mapping from student code to the SOLO taxonomy “since the mapping process seems very context bound and question specific”. Indeed, Meerbaum-Salant et al. (2013) remarked that while the strength of the SOLO taxonomy lies in the fact that it offers a holistic, rather than a local perspective, “using [it] for various types of activities, simultaneously, is not straightforward”. When they set out to combine the Revised Bloom taxonomy
103