Case study: Developing long-term knowledge with Sprego

Abstract

In Hungary, K-12 informatics/computer science education focuses on mostly surfacebased methods. This approach can be observed in the teaching of several topics in the subject, of which we focus on spreadsheet management. This is further emphasized by regulatory documentsthe Hungarian National Core Curriculum and Hungarian Curriculum Frameworks-, where handling algorithms, calling schemata, and problem-solving in general are only assigned to the programming topic. In the process of fulfilling the requirements of the school curricula and the various tool-centered exams, students become familiar with the software interfaces and how to navigate them, instead of developing computational thinking skills and learning how to approach and solve real-world problems. Our educational system is based on a spiral teaching approach; therefore, spreadsheet management is taught throughout several grades in a small number of lessons. Prior research shows that students learning spreadsheet management with surface-approach methods do not build up a reliable knowledge structure. These students cannot solve problems in contexts differing to the ones in which they learned the topic and cannot use their surface navigation abilities in different software environments. Our research group focuses on spreadsheeting with an algorithm-building and problemsolving method at the center of the teaching-learning process. For this purpose, we have developed and introduced the Sprego (Spreadsheet Lego) methodology. Sprego is based on Pólya's four-step concept-based problem-solving approach, and its efficiency has already been proved compared to traditional low-mathability surface-approach methods. In the comparison of the low-and high-mathability approaches, several further questions arise, and amongst them one crucial aspect is how the different methods support the schemaconstruction and knowledge built up in long-term memory. In this paper we discuss this question using a delayed post-test that was carried out one year after the treatment period. We focused on the students' achievement both in the experimental (Sprego) and control (traditional surface-approaches) groups based on the methods used one year prior to the administration of the delayed post-test. The results show that students who learned the spreadsheet management topic with Sprego achieved significantly better scores on the delayed tests than those students who used low-mathability approaches.

The Comparison of Students’ Self-Assessment, Gender, and Programming-Oriented Spreadsheet Skills

“…Previous research proved that teaching spreadsheeting from a programming aspect [27,28,73,74] is much more effective [71,[75][76][77] than the widely accepted tool-centered low-mathability methods. The surface approach methods-focusing on the teaching of how the user interface works, typing spreadsheet tables, browsing wizards and help, coursebooks introducing as many functions as possible, and softwares offering newer and newer functions and features-do not support schema construction.…”

“…On the contrary, most of the students' results based on the three tasks of the test are pre-structural (84.7%), which means that no recognizable pieces of knowledge are present. The students' results demonstrate that the decontextualized, tool-centered, low-mathability approaches that characterize computer science education are inadequate for developing students' problem-solving and computational thinking skills [28,41,42,59,62,[73][74][75][76][77][78]. Furthermore, in complete accordance with the Dunning-Kruger effect [91], the pre-structural level of understanding (missing the point) prevents students from seeing what they do not know.…”

“…One must also note that-contrary to Prensky's unproven claims considering digital generations [52,53]-even digital children are in great need of educational support. To provide this support, we must have expert teachers [26,40,45] who can break away from the widely accepted tool-centered, lowmathability teaching and testing approaches, and switch to high-mathability problemsolving approaches [29,30,57,58,75,76]. Institutional Review Board Statement: Ethical review and approval were waived for this study, due to structure of the test, where only the schools are recorded strictly and occasionally the grades.…”

“…Previously conducted research has proven that teaching spreadsheeting from a programming point of view and using functional languages as introductory programming languages is significantly more effective than either the interface-centered or the imperative or object-oriented "classical" methods [27,28,[71][72][73][74][75][76][77]. Being aware of these results and that school grade levels are often not good indicators of where students are in their cognitive development, the students' self-assessment values are compared to their results in the test.…”

“…However, this is not so, because browsing the interfaces and looking for ideas requires slow thinking [64], whose processes are proved erroneous, leading to serious spreadsheet errors [39,80,84,85]. Instead, with a reasonable cognitive load [29,30,76], the names, the arguments, and the semantics of the functions must be memorized to allow end-users to activate fast thinking, which would lessen spreadsheet errors. Schema construction and applying the built schemata in solving novel problems is widely accepted and applied in teaching mathematics [86].…”