The abstraction transition taxonomy

Abstract

The abstraction transition taxonomy: developing desired learning outcomes through the lens of situated cognition. In: Clear, A., Sanders, K. {sesper, srfoster, mfecho, bsimon}@cs.ucsd.edu ABSTRACTWe report on a post-hoc analysis of introductory programming lecture materials. The purpose of this analysis is to identify what knowledge and skills we are asking students to acquire, as situated in the activity, tools, and culture of what programmers do and how they think. The specific materials analyzed are the 133 Peer Instruction questions used in lecture to support cognitive apprenticeship -honoring the situated nature of knowledge. We propose an Abstraction Transition Taxonomy for classifying the kinds of knowing and practices we engage students in as we seek to apprentice them into the programming world. We find students are asked to answer questions expressed using three levels of abstraction: English, CS Speak, and Code. Moreover, many questions involve asking students to transition between levels of abstraction within the context of a computational problem. Finally, by applying our taxonomy in classifying a range of introductory programming exams, we find that summative assessments (including our own) tend to emphasize a small range of the skills fostered in students during the formative/apprenticeship phase.

Experience Report

“…The course design incorporates a strong code comprehension element, in line with a number of researchers' recommendations: preclass activities encourage students to practice their code-reading ability [18]; PI questions foster understanding and analysis of programs at a range of levels of detail [4,24]; the concept of the notional machine is introduced to learners during class, and they are shown how the activities they have been undertaking should help to develop their own notional machine understanding [6,7,29]; labs encourage the development of a number of smaller programs rather than just one large one [10]; some lab sessions encourage students to explore complete programs and draw out re-usable patterns [19]; the staff-student ratio in the labs is high to provide sufficient time for students to talk to staff about their programs, displaying their understanding. These discussions are intended to emphasise the importance of understanding over simply getting programs working.…”

The Stained Glass of Knowledge: On Understanding Novice Mental Models of Computing

“…• Sociocultural Factors [23,123,132,160] • Communication with Classroom Community [21,61,81,97,132,152,161] • Active or Group Learning Dynamics [21,25,90,150,151] • Learner's Intent in Taking a Course [15] • Beliefs Surrounding Pedagogical Tools [4,146] • Interpretation of a Pedagogical Tool's Use [4,20,146,165] • Ordering of Topics in the Course [168] • Classroom's Spatial Design [8,89]…”

“…The spectrum of representational methods and generalizations one may be working through, especially during the problem solving process. In Computer Science Education, Cutts et al [25] characterized three main levels of abstraction that one moves between at various intervals of the problem solving process: "normal language", "CS speak" (utilization of jargon and pseudocode-esque natural language), and "executable code".…”