Exploring the Relationship Between Coding, Computational Thinking and Problem Solving in Early Elementary School Students

• In Event: Fostering Young Children's Computational Thinking

Abstract

Objective: This paper explores the relationship between Computational Thinking (CT) and improvements in problem-solving skills in early elementary school children who participate in a coding curriculum.

Theoretic Framework: In 2006, Jeannette Wing proposed that acquiring Computational Thinking skills can have collateral benefits for the development of other domains of thinking. There is evidence that learning to code can improve Computational Thinking skills (Grover & Pea, 2013, Lye and Koh 2014; Fraillon et al., 2018), but few studies have explored the impact of coding-acquired CT skills on other areas of cognitive development in young children. In a pilot study, we observed that learning to code by exposure to an age-appropriate coding curriculum was associated with improvements in young children's performance in solving non-coding (e.g., unplugged) problems. As a test of Wing’s hypothesis, we explored whether this improvement in problem solving correlated with acquisition of Computational Thinking skills.

Methods: Nearly 600 2nd grade students from the Norfolk, Virginia school system completed a six-week Coding as Literacy (CAL) coding curriculum that included 12 hours of instruction and play with the KIBO educational robotics platform. KIBO is a programmable screen-free robot designed to engage young children in playful and creative coding.

Data Sources: Participating students were administered a battery of age-appropriate formative and summative assessments that were designed to evaluate three domains: coding skills (Solve-Its), Computational Thinking (TACTIC-KIBO) and unplugged problem-solving abilities (Tech-Check). For comparison purposes, some of the same assessments were administered to a group of 2nd graders who were not exposed to the CAL curriculum.

Results: A total of N=669 children participated in the Norfolk CAL study from a total of 10 public schools. N=8 schools N=580 (51.7% M, 49.3% F) children received the CAL curriculum and N=2 schools with n=89 children (44.9% M, 55.5% F) did not participate in coding and received the Tech Check problem solving assessment. This paper describes the results obtained in the Norfolk CAL study in terms of Tech Check, TACTIC, and Solve-it measures and the correlations among them.

Significance: Learning to code is a valuable skill in itself but it is important to know if it can reliably translate into improvements in other domains of thinking. This study is among the first to explore how teaching young children to code may affect the development of other problem-solving skills. Computational Thinking skills can be considered a form of generalization of the knowledge gained from learning to code. Acquisition of CT skills may permit young children to address a broader range of problems. We plan to further examine the elements of the coding curriculum that encourage this type of generalization of learning in the hope of improving future teaching strategies. We discuss the implications of our current findings for future studies as well as for the design of coding and other educational initiatives for young children.

Authors

• Emily Relkin, Tufts University

• Marina U. Bers, Tufts University