The nature and transfer effects of computational thinking in STEM

Abstract

Computational thinking (CT) has attracted much interest as a problem-solving skill and debatably ranks in importance with the time-honoured literacy skills of reading, writing, and arithmetic. However, the nature of CT is confounded by the diverse conceptualization in the literature. More so, it is unclear whether CT skills could be acquired from learning to program, and what constitutes the pedagogical ingredients in supporting learners. This thesis is a compilation of five connected studies to gain insight into the nature of CT, discover the centrality of abstraction in CT, investigate the efficacy of project-first approach to programming in technology deprived environment, explore the nature of CT in everyday problem-solving and then examine whether programming enhances CT. Study 1 is a systematic review (n = 81) that overviews the diverse ways in which CT has been operationalised in empirical studies. Findings include the propensity of researchers to operationalise CT as a composite of programming concepts or framed from the assessment-based models. Contentiously, a sizeable number of studies neither established the meaning of CT nor distinguish between CT and programming. Study 2 builds on the finding that abstraction is a consistent dimension across CT frameworks but is incoherently operationalized. The systematic review (n = 96) of the nature of abstraction identified four independent but related abstractive processes – discover, extract, create, and assemble (DECA) that enhance the simplification of problem-solving. Study 3 considered the prior finding that programming is the overwhelming approach to learning CT and explored how students from technology-deprived classrooms develop their programming skills using constructionist-based project-first approach and whether the presence of different programming concepts is associated with programming ability. Study 4 is an inquiry into the nature of CT problem-solving. Although most educators attribute problem-solving as the underlying target for CT, the nature of CT solvable problems is unknown. Through a qualitative multiple case study, study 4 explored how CT is demonstrated in everyday problem-solving. Findings show that CT practices are latent in everyday problem solving, focusing on simplification trumps formal proof of algorithmic correctness, and (re)structuring of problems increases the applicability of CT in everyday problem-solving. Study 5 is an experimental study (n = 59) that investigated if learning with the DECA model enhances students’ programming knowledge and CT skill. Subsidiary questions explored whether students’ age, gender, computer proficiency and prior programming experience affect the acquisition of CT. No significant difference was found in the CT skill and programming knowledge of the groups at the posttest. However, within-group paired t-tests showed that DECA approach resulted in significant CT improvement in the experimental group but not in the control group across the pretest-posttest axis. Keywords: computational thinking, problem solving, algorithms, programming, abstraction, decomposition, generalisation, constructionism.