Computational thinking in university students: The role of fluid intelligence and visuospatial ability

Abstract

Computational thinking (CT) is a set of problem-solving skills with high relevance in education and work contexts. The present paper explores the role of key cognitive factors underlying CT performance in non-programming university students. We collected data from 97 non-programming adults in higher education in a supervised setting. Fluid intelligence, crystallized intelligence, and visuospatial ability were assessed using computerized adaptive tests; CT was measured using the Computational Thinking test. The direct and indirect effects of gender and visuospatial ability through fluid intelligence on CT were tested in a serial multiple mediator model. Fluid intelligence predicted CT when controlling for the effects of gender, age, and visuospatial ability, while crystallized intelligence did not predict CT. Men had a small advantage in CT performance when holding the effects of cognitive abilities constant. Despite its large correlation with gender and CT, visuospatial ability did not directly influence CT performance. Overall, we found that programming-naive computational thinkers draw on their reasoning ability that does not rely on previously acquired knowledge to solve CT problems. Visuospatial ability and CT were spuriously associated. Drawing on the process overlap theory we propose that tests of fluid intelligence and CT sample an overlapping set of underlying visuospatial processes.