Assessing Algorithmic Thinking Skills in Early Childhood Education

Abstract

This chapter presents part of a wider project aimed at developing computational thinking assessment instruments for first and second grade primary school students. The applicability of the specific proposed tool, which concerns merely the algorithmic thinking (AT), was tested within the Environmental Study course (ESc). The main pillar of the work is the computational environment PhysGramming. The assessment of AT was based on mental tasks involving puzzles which require AT abilities. The AT test comprised of four puzzles with 4, 6, 9, and 12 pieces respectively, and the puzzle-solving performance was measured at the nominal level (success/failure). Latent class analysis (LCA), a robust multivariate method for categorical data, was implemented, which distinguished two clusters/latent classes corresponding to two distinct levels of AT. Moreover, LCA with covariates, such as gender, grade, achievement in ESc, and the use of plan revealed the association of the above variables with the AT skill-levels. Finally, the results and their implications for theory and practice are discussed.