STATE - Assessment Form Catalog

Current truth

DOL English V2 uses a canonical assessment-form layer to keep Goal, Recommendation, Metrics, and Exercise Bank routing consistent while allowing leaf taxonomy to evolve independently.

Rules

• Every attempt/result carries one assessment_form_id.
• Core form metadata:
  • program_id,
  • skill_id,
  • scoring_family (objective | subjective_ai | hybrid),
  • score_profile_id,
  • timing_family (untimed | timed | adaptive).
• Onboarding goal form/skill options are sourced from this catalog (not hardcoded per screen).
• Goal comparison and metrics interpretability depend on assessment_form_id + score_profile_id.
• Recommendation fallback for ineligible goal-gap prioritizes same assessment_form_id.
• Week-1 activation starter selection should consume catalog form metadata:
  • resolve starter forms within selected program first,
  • preserve assessment-form continuity between activation step-1 and step-2 when possible.
• Leaf filter/taxonomy changes do not alter core contracts as long as form mapping stays valid.

Program baseline forms (V1)

• SAT:
  • sat_verbal_mixed_obj,
  • sat_verbal_one_type_obj,
  • sat_math_mixed_obj,
  • sat_math_one_type_obj.
• TOEIC:
  • toeic_reading_parted_obj,
  • toeic_listening_parted_obj,
  • toeic_writing_task_ai,
  • toeic_speaking_part_ai.
• Communication:
  • com_speaking_scenario_ai,
  • com_writing_scenario_ai,
  • com_microdrill_obj,
  • com_integrated_mixed.

Decision trace

• DEC-0071
• DEC-0072
• DEC-0073
• DEC-0080