By Anna Fergusson (University of Auckland)
Information
Data science education faces the challenge of teaching computer programming alongside statistical concepts without overwhelming novice learners. This poster presents "code blanks", a pedagogical approach that provides students with code that contains only some sections that are editable, in order to focus attention on key programming decisions. Code blanks use a combination of visual scaffolding and constrained interaction mechanisms (text boxes and dropdown menus) to support purposeful reasoning and problem-solving within authentic data science workflows. Unlike traditional use of "fill in the blanks" type exercises, our use of code blanks is not just to assess "the correct answer" but to provide a scaffold for building familiarity with code, that also supports student creativity by promoting a tinkering mindset. We will demonstrate what code blanks are, describe the four key use cases for code blanks when designing interactive online coding tasks, and present some initial results from our implementation with introductory level data science students focusing on their interactions with the online coding tasks and evidence of "tinkering".