Programming tools are being used in education to teach computer science to children as young as 5 years old. This research aims to explore young children’s approaches to programming in two tools with contrasting programming interfaces, ScratchJr and Lightbot, and considers the impact of programming approaches on developing computational thinking. A study was conducted using two versions of a Lightbot-style game, either using a ScratchJr-like or Lightbot style programming interface. A test of non-verbal reasoning was used to perform a matched assignment of 40, 6 and 7-year-olds to the two conditions. Each child then played their version of the game for 30 minutes. The results showed that both groups had similar overall performance, but as expected, the children using the ScratchJr-like interface performed more program manipulation or ‘tinkering’. The most interesting finding was that non-verbal reasoning was a predictor of program manipulation, but only for the ScratchJr-like condition. Children approached the ScratchJr-like program differently depending on prior ability. More research is required to establish how children use programming tools and how these approaches influence computational thinking.

The recent shift in compulsory education from ICT-focused computing curricula to informatics, digital literacy and computer science, has resulted in children being taught computing using block-based programming tools such as Scratch, with teaching that is often limited by school resources and teacher expertise. Even without these limitations, Scratch users often produce code with ‘code smells’ such as duplicate blocks and long scripts which impact how they understand and debug projects. These code smells can be removed using procedural abstraction, an important concept in computer science rarely taught to this age group. This article describes the design of a novel educational block-based programming game, Pirate Plunder, which concentrates on how procedural abstraction is introduced and reinforced. The article then reports an extended evaluation to measure the game’s efficacy with children aged 10 and 11, finding that children who played the game were then able to use procedural abstraction in Scratch. The article then uses game analytics to explore why the game was effective and gives three recommendations for educational game design based on this research: using learning trajectories and restrictive success conditions to introduce complex content, increasing learner investment through customisable avatars and suggestions for improving the evaluations of educational games.

Scratch users often struggle to detect and correct 'code smells' (bad programming practices) such as duplicated blocks and large scripts, which can make programs difficult to understand and debug. These 'smells' can be caused by a lack of abstraction, a skill that plays a key role in computer science and computational thinking. We created Pirate Plunder, a novel educational block-based programming game, that aims to teach children to reduce smells by reusing code in Scratch. This work describes an experimental study designed to measure the efficacy of Pirate Plunder with children aged 10 and 11. The findings were that children who played the game were then able to use custom blocks (procedures) to reuse code in Scratch, compared to non-programming and programming control groups.

Computation is a fundamental part of our world, with today's children growing up surrounded by technology. This has led governments and policymakers to introduce computer science into primary and secondary education (age 5 to 16). These developments have been driven by 'computational thinking': the idea that the problemsolving skills used in computer science are useful in other disciplines. They have resulted in a wide range of programming tools designed for novices, of which Scratch, a block-based visual programming environment, is the most popular. Yet, so far, both computer science education and claims of computational thinking as a universal skill have failed to live up to their potential. This thesis begins by reviewing the literature on computer science in primary education and computational thinking. It then describes a study that aimed to reproduce findings that programming improves story sequencing, a non-computational skill, in young children (age 5 and 6) using a programming game. The results showed an overall improvement for both the intervention and control group. In addition, it highlighted issues with teaching programming to young children. The thesis then refocuses on teaching older children (age 9 to 11) the computer science skill of abstraction and the idea that it can be used to refactor code to remove 'code smells' (bad programming practices). Code smells indicate an underlying problem in a program, such as code duplication, and are common in Scratch projects. A study is then reported that establishes that primary school children can recognise the benefits of abstraction when asked to alter Scratch projects that contain it. The thesis then describes the design and development of Pirate Plunder, a novel educational block-based programming game designed to teach children to use abstraction in Scratch, using custom blocks (parameterised procedures) and cloning (instances of sprites). Two studies are reported in the subsequent chapters. The first investigates the value of a debugging-first approach in Pirate Plunder, finding that it was not always beneficial. The second measures for improvements in using abstraction in Scratch, finding that children who played the game were then able to use custom blocks to reduce duplication code smells in a Scratch project. In addition, Pirate Plunder players improved on a computational thinking assessment compared to the non-programming control group. The final chapter discusses the original contributions of the thesis, the implications of these and future direction. Declaration I hereby declare that: 1. I have not been enrolled for another award of the University, or other academic or professional organisation, whilst undertaking my research degree. 2. None of the material contained in the thesis has been used in any other submission for an academic award. 3. I am aware of and understand the University's policy on plagiarism and certify that this thesis is my own work. The use of all published or other sources of material consulted have been properly and fully...