Refurbishing the Educational Escape Room for Programming: Lowering the Threshold and Raising the Ceiling

This paper presents the design, development, and evaluation of a web-based educational escape room game teaching basic Python programming concepts to K-12 students, with the aim of making programming more accessible ("lowering the threshold") while still supporting meaningful learning and engagement ("raising the ceiling"). Motivated by reported difficulties in university programming courses and by challenges in integrating programming into Swedish and international K-12 curricula, the authors followed a five-step design science process. The resulting game, titled Escape with Python, contains ten rooms covering five programming concepts (variables, sequences, conditions, iterations, and functions), with each concept revisited at increasing difficulty and delivered through short character dialogues and puzzle-based challenges.

The game was tested by 32 students, mostly around grade 9, in mid and northern Sweden during the 2021 and 2022 academic years, using a questionnaire mixing closed and open-ended items. Closed-ended ratings of programming knowledge and interest on a 1 to 10 scale showed averages of roughly 4.6 both before and after play, with no meaningful overall shift. Counts of individual change, however, were more positive: twelve students reported increased knowledge against five reporting a decrease, and seven reported increased interest against six reporting a decrease. Thematic analysis of the open-ended responses grouped feedback into graphics, mechanics, content, and learning and engagement. Students valued the simple retro-style visuals and the early accessibility of the game, but requested greater consistency in mechanics, more freedom such as writing real code, faster dialogues, clearer hints, a life system to discourage guessing, and more varied narrative themes.

The authors acknowledge limitations, including the small and self-selected sample, a relatively simple development tool, and questionnaire-based measurement without pre- and post-testing of learning outcomes. They note that the closed-ended results contain internal inconsistencies, and they plan future work combining student data with teacher evaluations, rebuilding the prototype with more professional tools, and conducting larger-scale testing.