This article reports on a three and a half year design-led project investigating the use of open-ended learning to teach programming to students of interaction design. Our hypothesis is that a more open-ended approach to teaching programming, characterized by both creativity and self-reflection, would improve learning outcomes among our cohort of aspiring HCI practitioners. The objective of our design-led action research was to determine how to effectively embed open-endedness, student-led teaching, and self-reflection into an online programming class. Each of these notions has been studied separately before, but there is a dearth of published work into their actual design and implementation in practice. In service of that objective we present our contribution in two parts: a qualitatively-derived understanding of student attitudes toward open-ended blended learning, as well as a matching set of design principles for future open-ended HCI education. The project was motivated by a search for better educational outcomes, both in terms of student coding self-efficacy and quantitative metrics of cohort performance (e.g., failure rates). The first year programming course within our interaction design-focussed Bachelors program has had the highest failure rate of any core unit for over a decade. Unfortunately, the COVID-19 pandemic confounded any year-to-year quantitative comparison of the learning efficacy of our successive prototypes. There is simply no way to fairly compare the experiences of pre-pandemic and pandemic-affected student cohorts. However, the experience of teaching this material in face-to-face, fully online, and hybrid modalities throughout the pandemic has aided our qualitative exploration of why open-ended learning helps some students but seems to harm others. Through three sets of student interviews, platform data, and insights gained from both the instructional and platform design process, we show that open-ended learning can empower students, but can also exacerbate fears and anxieties around inadequacy and failure. Through seven semesters of iterating on our designs, interviewing students and reflecting on our interventions, we've developed a set of classroom-validated design principles for teaching programming to HCI students without strong computational backgrounds.

No abstract