 THE HEARING / LEARNING AND USABILITY / THIRD REVIEW



 What does the learner take away?
 Nine fresh AI perspectives, concrete revisions and an observed-classroom pilot plan.

 Instructional judgment: the activity is strongest at teaching a bounded research move: identify how an
 action, rule or resource affects a political opportunity; distinguish the record from an inferred effect; test a
 competing account and a comparison. The evidence is less suited to settling the total cause of national
 election outcomes. The revised instructions make that scope explicit.

 The strongest recurring problem was placement, rather than lack of information. Essential actor
 relationships were easy to miss in disclosures; source cautions were easier to repeat than apply; and two
 source-answer/decision cycles made a long hearing page. The new structure brings the key context forward
 and separates each cycle without removing the second question or any source.

   ID       Fresh perspective                      Actual scope


   T1       Low-preparation first-year             Browser through evidence selection, then a clearly labelled cold read
                                                   of remaining screens.


   T2       High-school student, age 16            All six stages; 72-word main and 39-word comparison drafts; two
                                                   source decisions; no final filing.


   T3       Advanced undergraduate                 Cold read and a substantive sample account, rival and comparison.


   T4       Curious nonstudent adult               All stages inspected; 72-word draft and one challenge/decision; no
                                                   completed comparison or filing.


   T5       Educational Instagram creator          Student-text and entry-image review; proposed an accurate short
            lens                                   demonstration.


   T6       Average working first-year             Cold comprehension review: Serbia and Malawi comparison.


   T7       Instructional designer                 Cold review of goals, scaffolding, transfer and feasibility.


   T8       Political-science faculty              Cold review of institutional context, evidence and comparison.


   T9       Game designer / UX / art direction     Cold content and entry-image review of agency, structure and
                                                   cognitive burden.

 All are fresh independent AI simulations, without previous feedback or code. Three used the interface to varying extents;
 six were content-only. No one in this round verified a newly filed complete record. These are not real students, faculty,
 creators or a completion-rate study.




THE HEARING / LEARNING AND USABILITY REVIEW / 30 OCTOBER 2026                                                                1
 IMPLEMENTED CHANGES



 More context where it matters.
 The change is to guidance and structure, with the complete analytical work retained.


   Finding                                             Revision implemented


   The basic political relationships became clear      A visible three-part power map identifies office/holder,
   only after extra reading.                           challengers/checks and opportunity at stake. Full background,
                                                       names, dates and added sequence help remain available.


   The target could drift from an opportunity to       A required short research-target label names one event or
   the overall winner.                                 opportunity, actor and period. It carries into drafting, comparison
                                                       and the readable report, outside the paragraph word count.


   A directly documented step could make the           Keep the recorded fact. Name the unresolved motivation,
   rival task feel artificial.                         implementation or broader-effect link that both accounts address.
                                                       An alternative is a hypothesis to test, not evidence of equal
                                                       plausibility.


   Correct practice recognition could leave the        Each existing decision reason points to a phrase or link in the
   original overclaim untouched.                       learner response and a case ID. A connection inspector stays
                                                       available while drafting. A person still judges the warrant.


   Source limits could sound as if nothing can be      Each card pairs a recorded, source-supported point with where the
   concluded.                                          claim stops. Full source text, provenance and limitations remain.
                                                       Detailed MEC numerical audit is layered under an explicit notice.


   The hearing was a long page with two                A hearing agenda shows one question at a time or both, with
   answer/decision cycles.                             separate attempt and decision status. Both questions remain
                                                       required and every attempt, source and decision remains available.


   An unfamiliar second case could become              The named target and current connection travel into comparison. A
   another summary assignment.                         topic index helps locate relevant evidence; all six cards remain
                                                       accessible and a failed analogy is legitimate.


   Casual visitors felt obliged to complete a class    They can save an explicitly labelled readable draft at any point. A
   assignment.                                         complete hearing still meets the same source, challenge, writing
                                                       and comparison requirements as class mode.

 Entry now names the transferable skill and gives case-specific research hooks. The optional closing reflection asks for a
 supported statement, a further inference and a reusable check. The teacher guide adds a practical observation sheet
 rather than treating a correct multiple-choice response as independent transfer.




THE HEARING / LEARNING AND USABILITY REVIEW / 30 OCTOBER 2026                                                                2
 RIGOR / VERIFICATION / RETEST



 Keep the intellectual demand.
 The program checks a record structure. The teacher or reader checks its meaning.


   Retained or strengthened         Verification / boundary


   Historical evidence              All 24 historical card bodies and all 12 active/12 legacy practice answer keys are
                                    preserved. Offices, dates, denominators, original links and unresolved source
                                    discrepancies remain.


   Analytical work                  Two distinct supports, a separate complication, actor/action/option connection, two
                                    practice attempts and reasoned decisions, a rival and discriminating evidence, and a
                                    second case remain.


   Authored writing                 Main explanation remains 60-85 words; comparison remains 30-45. The new target is
                                    a short label. No historical paragraph is written or repaired by the app.


   Current and earlier work         Target/view/reflection data persist and import. Old completion timestamps remain as
                                    history if current requirements are missing. Changing a view or optional reflection
                                    does not undo valid completion.


   No substitute score              No score for a political conclusion, speed, caution words or changing views. A
                                    complete record can still contain poor reasoning. No semantic auto-grading or
                                    learning-gain claim.


   Engineering checks               594 assertions pass: 268 new and 326 prior. Prior complete fixtures include the new
                                    target; every earlier analytical gate is retained. These are event/state and
                                    HTML-output tests, not browser or learning evidence.


 A reporting inconsistency found and corrected
 Independent engineering review found old comparison-purpose labels still in the export, even though the
 screen used neutral labels. That could revive a false outcome contrast. The report now matches the
 interface; new consistency regressions pass.

 Informed learner retest
 Verified visible power map, required target, target carried into comparison, current-connection inspection,
 Question 2 and Show both navigation, separate attempt/decision status, and both decisions still required.
 Source-supported points and card limits were visible. Audit expansion readback, report preview and casual
 draft delivery were interrupted or outside that sampled retest; they are not claimed as browser-verified.

 The first two browser attempts reached a different prototype because of preview routing. They were excluded entirely.
 Correct title and country choices were verified before accepted observations. Those setup/protocol errors are not evidence
 of product confusion.




THE HEARING / LEARNING AND USABILITY REVIEW / 30 OCTOBER 2026                                                                 3
 REVIEW DISPOSITIONS / FIRST REAL CLASS



 How to test the teaching claim.
 All nine full reports and the detailed response register are included in the kit.


   Perspective                What changed / what remains to test


   T1 low-preparation         Earlier neutral example and plainer terminology; positive source points, short target and
   learner                    layered audit. Full independent completion remains untested in that profile.


   T2 high-school learner     Essential power context is visible, the task names all required work, and the hearing
                              separates answer/check/decision. The later informed retest confirms the core interaction
                              changes.


   T3 advanced learner        Unresolved-link rival framing and persistent target support a higher analytical ceiling. The
                              app cannot decide whether an account is well warranted.


   T4 nonstudent              Voluntary readable drafts and mode-appropriate sharing language; complete analysis
                              retains all class requirements. No shorter route is represented as a complete hearing.


   T5 creator                 Concrete entry action, dated case context and supplied-claim labels. The proposed short
                              clip remains one move within the longer activity; no posting or engagement forecast.


   T6 average reader          Office relationships, simpler source/rival language, an invented denominator example and
                              topic-based comparison browsing.


   T7 instructional design    Own-writing decision prompts, documented-fact versus uncertain-link guidance, optional
                              reflection and a human-pilot sheet. Automatic relaxation of word limits was not adopted.


   T8 political science       Explicit institutional target, viable hypothesis-limit guidance and the Malawi threshold
                              qualification in the opening. No common regime classification is imposed.


   T9 game/UX                 Focused agenda, visible connection and selected evidence relevance. New timers, political
                              scores, simulated electoral effects and another required essay were not added.


 Is it doable in one class?
 The reconciled 55-minute timetable remains a planning estimate. The high-school simulation produced
 substantive drafts and two decisions, but browser operation is not student time. Reading confidence,
 discussion and revising a novel comparison may require more time. The guide provides a carryover point
 and a two-session option that preserve all reasoning.

 Collect evidence of understanding
 Use the new guide sheet: phase times and help needed; distinct observations; an actor-to-option warrant; a
 claim-relevant limit; rival accounts of a common uncertain link; differing expected patterns; and a valid
 comparison. Ask the learner to explain one reusable research move without supplied options. Use
 unfamiliar later material to probe retention. A causal learning claim needs an appropriate comparison
 design.

 Physical-phone use, a full screen-reader audit, cloud-browser download completion, durable transfer and actual human
 learning gains remain unverified. The review supports a stronger design and a concrete pilot, not a claim of measured
 educational effectiveness.




THE HEARING / LEARNING AND USABILITY REVIEW / 30 OCTOBER 2026                                                                4
