How broad is computational thinking? A longitudinal study of practices shaping learning in computer science

Computer science is becoming a mainstream school subject, yet we know relatively little about teaching, learning, and assessing computer science at the primary and secondary level. Few studies have followed the long-term trajectories of early computer science learners. We present a longitudinal study of a school cohort (N=48) across a three-year computer science curriculum in grades 6-8. We analyzed students' Scratch projects in terms of elaboration and computational thinking content, and modeled their association with performance on a summative open-ended assessment of computational thinking. Both metrics were associated with performance on the summative task, but engagement had a much more substantial effect. This supports the idea that early computer science experience should be designed to support students in working on personally-meaningful projects. Developing computational literacy practices may be more important for long-term growth in computational thinking than a primary emphasis on content knowledge.