A Placement Diagnostic for a Multi-Entry Zero-to-Competitive Python Curriculum

We describe the design of a placement diagnostic instrument for Python Pathway, a five-volume zero-to-competitive Python curriculum targeting Vietnamese K-12 learners on a chuyên Tin / IOI feeder trajectory. The curriculum is multi-entry by design: Volume 0 (grades 3–5, unplugged computationalthinking literacy) and Volume 1 (grades 5–12, Python syntax and structure) both function as legitimate starting points, and older self-directed readers with prior programming exposure may legitimately skip into mid-Vol 1 or Vol 2. A self-directed reader who mis-places either stalls or disengages without recourse. The instrument must therefore produce five routing decisions — Vol 0 entry, Vol 1 entry, mid-Vol 1 skip, Vol 2 direct, or pre-Vol-0 remediation — with enough precision that a first-time reader at home can take the diagnostic and start the right volume the same afternoon. We specify a 15–20 item pool across four fixed constructs (reading-fluency floor, numeracy floor, prior-CS notionalmachine knowledge, computational-thinking disposition) and one adaptive block (Vol 2 candidacy, administered only to learners scoring perfect on prior-CS). We describe scoring (conjunctive across constructs, compensatory within construct), routing (explicit decision tree with item-level mid-Vol-1-skip chapter recommendations), and delivery trade-offs (online as primary form with Parsons puzzles and auto-grading, paper as secondary form with short-trace substitution). We lay out a revalidation loop that activates at N ≥ 100 administrations, independent of calendar. We discuss limitations: the program does not run empirical cohort validation, so v1 thresholds are expertconsensus; criterion validity is deferred to the revalidation cycle. The instrument is distributed as advisory routing, not certification, and is designed to survive the research program's no-cohort constraint.