Evaluations Guide

Provenance (checked 2026-09-29): written as an independent document in this course's WBS14 batch, using content/veridian/technique-hub/wbs12/forge-evaluations-guide.md as the reference for shape and rigor only — that document's own history is a caution, not a template for content: it once carried an entirely fabricated "trigger phrase" mechanism (a claimed Level-4 vocabulary requirement), confirmed absent from every real WBS12 examiner report checked. Before writing a single new claim below, this pass read research/veridian/WBS14-verified-facts.md in full (2,533 lines), all eight lessons in content/veridian/curriculum/wbs14/, and every file already in content/veridian/technique-hub/wbs14/ — in particular forge-exam-technique-guide.md, forge-level-threshold-guide.md, and signal-examiner-intelligence.md, which between them already carry the tariff-by-tariff descriptor quotes, a boundary-by-boundary drill with three fresh real anchors, and a study of how examiner-report language itself behaves. That existing coverage is the reason this document takes the shape it does. Unlike WBS12, where a fabricated draft needed a full rebuild, WBS14's AO4 ground has already had a genuine, well-sourced pass — this document's job is to map that ground honestly, run the one check that pass could not run on itself (a direct search for the WBS12-style fabrication pattern, across the whole WBS14 build), state the one throughline that ties the existing pieces together, and name plainly what is still actually missing. Every specific claim below traces to the facts bank, a real spec/mark-scheme quotation already verified there, or a named file in this course. Direct quotation from Pearson's own wording is kept to short, attributed fragments under 15 words throughout.

#WBS14 — Evaluations: A Map of What's Real, and What Is Still Missing

#0. What this document is (and isn't)

This is not a second worked-boundary drill. forge-level-threshold-guide.md already does that job, with three real anchors this pass deliberately does not re-derive (EU fishing subsidies for the 8-mark Discuss ladder, the WTO for the 12-mark Assess ladder, Porter's five forces/pet food for the 20-mark Evaluate ladder). It is not a second study of examiner-report language — signal-examiner-intelligence.md already covers that, including the one real, two-series pattern ("it depends" beats a flat side, when a report says so explicitly) that sits closest to what a WBS12-style "Evaluations Guide" would normally open with. And it is not a rebuild of the tariff ladders themselves — forge-exam-technique-guide.md §2 already states each command word's real, confirmed band and its real descriptor quote.

What is left, and what this document actually does:

1. States, in one place, exactly which command words on this paper carry zero AO4 — never assembled as its own point anywhere else in this folder (§1).
2. Names the one real, one-sentence difference between what Assess's ceiling asks for and what Evaluate's ceiling asks for — implied across two existing files but never stated as a single, direct contrast (§2).
3. Runs a direct fabrication check — the actual thing a WBS14 "Evaluations Guide" owes this course, given why the WBS12 version of this file had to be rebuilt (§3).
4. Spot-checks the WBS14 build's own "confirmed on every series" discipline, since that is the second failure mode the WBS12 rebuild found (§4).
5. Maps where the real, already-built AO4 material actually lives, so nobody re-derives it a third time (§5).
6. States plainly what is still genuinely unverified (§6) — and stops there, rather than filling the gap with something that sounds plausible.

---

#1. Where AO4 lives on this paper, and where it flatly doesn't

Every mark on this paper sits under one of eight command words, each attached to a fixed tariff and a fixed AO mix (spec Appendix 6, verified in research/veridian/WBS14-verified-facts.md's own paper-mechanics section). Read straight down that table, four of the eight carry no evaluation component at all, by the table's own wording:

Two things worth stating plainly, because they are easy to get backwards under time pressure:

• A 6-mark Analyse question is not a smaller Assess question. It shares the same "chain of reasoning" language as Discuss and Assess, but the table's own AO column rules out evaluation entirely — a "however" sentence on a 6-marker earns nothing it wasn't already going to earn from the chain itself. Nothing in research/veridian/WBS14-verified-facts.md records a WBS14 examiner report stating this in as many words for a specific series (the closest confirmed WBS14-specific evidence is the spec's own table wording, not a report quote — this document does not borrow WBS12's own "Analyse questions do not have any AO4" report quote and relabel it as WBS14's, since that quote belongs to a different paper's own examiner report).
• The spec sets no per-question AO breakdown anywhere on this paper. paper-anatomy.ts states this directly, after checking: WBS14's mark schemes are organised by command word and level-descriptor band, not by a per-question AO1–4 split, and the only AO percentages Pearson actually publishes are whole-unit ones (AO4 = 27.5% of IA2's marks; 6.9% of the full IAL award). The 12-mark Assess table's own four bands are the closest thing to a per-question AO4 rubric this paper has, and even that band structure is not labelled "AO4" inside the mark scheme itself — it is one integrated descriptor per level, blending all four AOs together, confirmed directly from the October 2025 mark scheme's own table structure (forge-level-threshold-guide.md §0).

The practical consequence: AO4 is concentrated in three command words that together account for 64 of this paper's 80 marks (Q1c's 8, Q1d and Q1e's 12 each, and the two 20-mark essays) — and it is the Section A structure, not any individual question, that decides how much of the paper's 27.5% AO4 weighting a candidate actually gets a fair shot at. A rushed Section A opener does not just cost marks on the opener; per forge-exam-technique-guide.md's own timing analysis, it borrows directly from the two questions in the same section that carry AO4.

---

#2. The one-sentence difference between Assess's ceiling and Evaluate's ceiling

paper-anatomy.ts already states that Assess and Evaluate "share the same top-band move," and forge-exam-technique-guide.md already quotes both closing clauses side by side. Read together, though, neither file states the practical difference as a single, direct sentence a student can hold onto under exam pressure — so here it is, built only from the two real descriptor fragments both files already verify:

Assess's ceiling needs a stated, supported judgement. Evaluate's ceiling needs that judgement to go one step further and actually propose something. The Assess Level 4 descriptor closes on "leading to a supported judgement" — a verdict, reasoned and named. The Evaluate descriptor closes on "a perceptive conclusion proposing a solution/recommendation" — a verdict that then does something with itself. A 20-mark answer that reaches a genuinely supported judgement and stops there has done what an Assess Level 4 answer needs to do, not what an Evaluate one does; the extra half-step — turning the judgement into a recommendation — is the entire distance between the two ceilings, not a difference in depth.

Two real, already-verified anchors make the contrast concrete without repeating either sibling file's own worked material:

• Assess, stopping at the judgement itself. The real January 2026 WTO mark scheme (Q1d) credits the WTO's genuine liberalisation record — 164 member countries, the Doha round, a real dispute resolved between Australia and Canada over wine — directly against its own real limits: the Doha round "collapsed after 14 years," the WTO "cannot force countries to reduce trade barriers," and, in the mark scheme's own words, it is "just one of several factors responsible for trade liberalisation." A Level 4 answer weighs those and lands on a stated, supported verdict about how much the WTO's role actually matters. It does not need to then say what a business or a government should do about that — the command word doesn't ask it to.
• Evaluate, taking the extra step. The real October 2024 niche-markets mark scheme (Q2) credits the same shape of balance — a global niche's price-inelastic premium and lower competition, set against real, named risks including a successful niche being bought out (the mark scheme's own example: Green & Black's, acquired by Cadbury) — but the Principal Examiner Feedback for that series records what separated the strongest answers: candidates who scored well "bring[ing] in other examples" beyond the extract, with genuine balance seen specifically in the "possibility of takeover as the niche grows." A 20-mark answer that reaches the same kind of supported verdict the WTO example reaches, but never turns it into a stated recommendation — should this business protect the niche, diversify out of it, prepare for acquisition, or accept the risk as the cost of the premium it's already earning — has done Assess's job at Evaluate's length, not Evaluate's own job.

That is the one line worth carrying into the exam: a supported judgement is necessary at both tariffs, but only Evaluate is actually asking to be told what to do about it.

---

#3. Fabrication check: does WBS14's own build carry a WBS12-style "trigger phrase" claim anywhere?

The WBS12 evaluations guide exists in its current form because a prior draft invented a rule — that Pearson requires specific vocabulary ("only if," "provided that," "conditional on") to reach Level 4 — and presented it as "confirmed from examiner reports across every series." That claim did not check out anywhere in the WBS12 archive. WBS14's own build never went through an equivalent unverified-draft pass (no WBS14 "definitive guide" draft exists anywhere in the ~/Downloads/FINSTO/ source material this course pulled from — checked directly by filename search before writing this section), so there was no single draft to re-verify. But the risk that the same fabrication pattern crept in independently was worth checking directly rather than assuming absent.

Method: every file in content/veridian/curriculum/wbs14/ and content/veridian/technique-hub/wbs14/ was searched directly for "trigger phrase," "only if," "provided that," and "conditional on."

Result: "only if" appears repeatedly — but in every single instance, it is this course's own conditional-judgement-drill device (a fill-in-the-blank teaching prompt of the form "X is likely to Y only if ___"), never attributed to Pearson as a required phrase, and never claimed as something an examiner report confirms candidates must write verbatim. Not one instance anywhere in the WBS14 build claims that specific wording is required, rewarded, or checked for by a mark scheme. This is a genuinely different, legitimate use of the same three words — a study tool that trains the reasoning move (naming the actual condition a judgement depends on), not a claim about Pearson's own marking vocabulary — and it is the same real mechanism signal-examiner-intelligence.md already independently confirms from actual report language: best-fit, whole-response marking against descriptor wording, never phrase-matching. No trigger-phrase fabrication exists anywhere in the current WBS14 build.

---

#4. Spot-check: does WBS14's "confirmed on every series" discipline hold?

The second thing the WBS12 rebuild found was a "confirmed every series, no exceptions" claim that had never actually been checked series-by-series. A direct search across the WBS14 build for "confirmed... every series," "every series... confirmed," and "without exception" returns nine matches. Eight of the nine are attached, explicitly and in the same sentence or the same document's own limitations section, to the real sample size being described — 5 to 8 of the roughly 15 series held since first assessment (June 2020), never the full population. paper-anatomy.ts, forge-exam-technique-guide.md, forge-question-selector-guide.md, forge-battle-card.md and forge-exam-day-protocol-extended.md each carry that hedge on their own respective claims. This is the correct discipline, consistently applied.

The ninth match is a different kind of claim and worth distinguishing rather than folding into the same count: assessing-global-markets-and-locations.ts's push/pull-factor teach block states that "every push factor, without exception, describes something true of the firm's current market." That is not an empirical frequency claim about how many exam series were checked — it is a definitional one (a push factor is, by the spec's own definition, a factor originating in the firm's existing market; a pull factor originates in a market it doesn't yet serve). Read as intended, it holds by definition and needs no series-by-series check to support it. It is named here only so a future pass checking this exact phrase doesn't mistake it for the empirical kind of overclaim this section was built to catch.

---

#5. Where the real AO4 material already lives

For anyone using this document to find the actual worked evaluation content on this paper, rather than re-deriving it:

---

#6. What's still genuinely open

forge-level-threshold-guide.md already flags this honestly, and it has not changed since: the 20-mark Evaluate tariff's own per-band descriptor text has never been independently re-quoted verbatim from a real WBS14 Q2/Q3 mark scheme, unlike the Discuss-8 and Assess-12 tables, where the actual band wording has been checked directly. Only the mark-range boundaries themselves (L1 1–4 / L2 5–8 / L3 9–14 / L4 15–20) are independently confirmed, across three real series. Everything this document says above about Evaluate's own closing requirement rests on the spec's command-word definition (Appendix 6, genuinely verified) and the confirmed mark ranges, not on a verbatim Q2/Q3 level-descriptor table — because no WBS14 source this course holds has actually fetched and quoted that specific table. This document does not close that gap, because doing so would mean either fetching a new primary-source mark scheme (out of scope for this pass) or writing something that sounds like a verified quote without being one. It stays open, named plainly, for whichever future pass fetches that specific document.

No grade boundary and no likelihood figure appears anywhere above, for the same reason forge-level- threshold-guide.md and signal-frequency-and-boundaries.md carry none: this paper's facts bank holds no such data.

---

#Sources

research/veridian/WBS14-verified-facts.md (paper-mechanics command-word table; the January 2026 WTO Q1(d) citation, verbatim indicative content, lines ~1088–1174; the October 2024 niche-markets Q2 citation and its Principal Examiner Feedback quotations, lines ~600–641). content/veridian/curriculum/wbs14/ paper-anatomy.ts (the AO4-weighting and no-per-question-AO-breakdown findings; the Assess/Evaluate shared-top-band-move framing). content/veridian/technique-hub/wbs14/forge-exam-technique-guide.md §2 (tariff ladders and descriptor quotes, cross-checked against directly, not restated in full). content/veridian/technique-hub/wbs14/forge-level-threshold-guide.md (the three existing boundary anchors, deliberately not re-derived here, and its own honest flag on the Evaluate-20 per-band wording gap, restated in §6 above rather than resolved). content/veridian/technique-hub/wbs14/signal-examiner- intelligence.md §3 (the existing "conditional verdict" pattern this document builds one further step on top of in §2, rather than re-arguing). International A Level Business Specification, Pearson Edexcel, Issue 1, September 2017, Appendix 6 (command-word table, quoted only in short attributed fragments). Direct filename search of ~/Downloads/FINSTO/ (confirming no WBS14-specific unverified draft exists in that source material, unlike WBS12). Direct grep of every file in content/veridian/curriculum/wbs14/ and content/veridian/technique-hub/wbs14/ for the fabrication and sample-size checks in §3 and §4.

Not affiliated with or endorsed by Pearson Edexcel. Every quoted fragment above is reproduced only as a short, attributed phrase, well under the whole sentence it comes from.