From Virtual to Physical Problem Solving in Coding: A Comparison on Various Multi-Modal Coding Tools for Children Using the Framework of Problem Solving

Introduction

Educators and researchers have identified the benefits of learning computer programming (a.k.a. coding) at a young age. Papert (1980) envisioned that learning coding could help children overcome fear of math, and help them to learn actively. This insight has been later emphasized and proven by other researchers. Clements et al. found that children who learned programming outperformed those who did not in reflectivity, divergent thinking, and metacognitive abilities (Clements, 1986). Strand et al. (1986) reported that programming facilitated collaboration among students, improved their social skills, and encouraged greater focus on their work. As the inventors of Scratch (Maloney et al., 2010) and ScratchJr (Flannery et al., 2013), Resnick et al. (2009) argued that through coding, children can develop “computational thinking” (Wing, 2006), which could cultivate creativity and important problem-solving strategies. Solomon (2005) suggested that “computer programming can be a useful, creative, and thoroughly entertaining second language for students at all levels.” More recently, Wong et al. (2015) investigated the impact of coding education at primary schools in Hong Kong, and reported an improvement in the overall performance of the students in mathematics, and the development of soft skills, after learning coding.

The revealed benefits of coding education have motivated many research efforts on new coding tools for children as an alternative to the conventional textual programming environment. The concept of visual programming (Bragg & Driskill, 1994) uses graphical symbols to represent coding concepts and allows children to compose computer programs by piecing graphic icons together. Although textual coding has its own benefits of low viscosity and high expandability (Green & Petre, 1996), visual programming, especially block building, has significant advantages over textual mode for children just entering the venue of computing. It allows children to focus more on learning the computational concepts rather than the complex syntax. As the tangible user interfaces (TUIs) emerged, researchers have also developed tangible block-based programming tools (e.g., Horn & Jacob, 2007). Studies show that graphical-user-interface-based visual programming can achieve better independence in study and learning outcome, while tangible programming is easier to use, more inviting, and more supportive for collaboration (Bers & Horn, 2009; Horn, Solovey, Crouser & Jacob, 2009; Resnick et al. 2009).

In spite of many research efforts on children’s programming education, existing work has mostly emphasized comparing user experience (e.g., the usability of coding interface) with different interface modalities. For example, Horn, Crouser and Marina (2012 suggested that children might have difficulty in operating the mouse on a graphical user interface. The emergence of touch surface technology has removed such barriers. Prior studies of hybrid coding interfaces advocated for providing users with the flexibility to choose a preferred modality (Horn, Crouser & Marina, 2012), but the impact of different systems on learning dynamics (e.g., peer interaction and interaction between teachers and students) is largely overlooked. Oviatt el. al’s (2012) research revealed the impacts of different interfaces on ideation and problem solving for high school and university students, but how different systems would affect young children in learning computational thinking and problem solving is still unclear. A recent study (Sapounidis, Demetriadis & Stamelos, 2015) compared children’s performance with graphical input and tangible input for robotics, and it suggested fewer programming errors occurred and better debugging was achieved with tangible input. This research examined different input methods for one unified output presentation (a physical robot). However, there was no detailed discussion on the potential affordances of different embodiment (graphical/tangible) of the problem itself, which could affect the problem-solving process (Czarnocka, 1995). In other words, there still lacks investigation on the affordance of different combinations of input and output modalities on passing problem-solving-related knowledge to children.