Observing abstraction in young children solving algorithmic tasks

Abstraction is considered an essential aspect of computational thinking. Primary schools are starting to include computational thinking into the curriculum. However, in order to guide their support, teachers need to know how to recognize abstraction. In this paper, we present how we can observe abstraction in young children tasked with solving an algorithmic assignment. In order to operationalize abstraction, we have used the layers of abstraction (LOA) model by Perrenet, Groote and Kaasenbrood. This model was originally used in the field of computer science and describes programming behavior at the level of software development, but has since been extended for use in primary education. We have operationalized this model for use with 5 to 6 year old students tasked with programming an educational robot. Their behavior has been related to each of the four layers of abstraction.
Students were individually instructed with programming Cubetto, an educational robot, to reach a number of destinations, increasing in the level of algorithmic complexity. We analyzed audio and video recordings of the students interacting with Cubetto and a teacher. Verbal and non- verbal behavior were categorized by two researchers and resulted in an observation schema.
We conclude that our operationalization of the LOA model is promising for characterizing young students’ abstraction. Future research is needed to operationalize abstraction for older primary school students.