Assessment Design for a Star-Topology K-12 Programming Curriculum

Star-topology curricula — in which a linear spine of content-independent cognitive milestones is crossed with a set of algorithmic clusters — displace the single-axis completion model that conventional K-12 programming assessment presupposes. When the curriculum is two-dimensional, a one-dimensional test necessarily collapses one axis into the other and produces passfail signals that do not distinguish a learner who has advanced the spine with shallow cluster coverage from one who has mastered a few clusters but is spine-stalled. This paper proposes an assessment framework tailored to the two-axis structure: spinemilestone markers and per-cluster fluency checks, gated jointly at each volume boundary. The framework operationalises the seven-milestone spine of the Spine-Cluster architecture against K-12 behavioural observables, decomposes cluster-fluency into a solve-recognise-adapt triad sampled along the four-axis variation space, and specifies a two-part boundary gate that runs criterionreferenced against a single rubric. We commit to four structural decisions carried from the programming-education literature: multi-dimensional reporting per Blackwell-Burnett and Messick, a tracing / explaining / writing triad per Lister et al. (2004), a mastery-learning gate per Bloom and Guskey with formative / summative partition, and the Anderson-Krathwohl revised taxonomy as a within-cell task-type generator. The proposed framework is instrumentable without cohort data, is anchorable against existing problem banks (VOI / VMO) as external criterion reference, and addresses the novel structural-validity problem that star-topology curricula raise for traditional assessment designs. No empirical validation is reported; the contribution is a design artefact grounded in existing literature and delivered in a form directly consumable by the Python Pathway series (Vol 0 through Vol 4).