Examiner Intelligence

Compiled 2026-09-29, from research/veridian/WBS14-verified-facts.md (the verified-facts bank; every citation below traces to a passage that document itself marks as a real, dated Pearson Examiner's Report or Principal Examiner Feedback document, not a mark scheme) and cross-checked against every other file in this folder — forge-exam-technique-guide.md, forge-vocabulary- precision-list.md, forge-error-pattern-tracker.md, forge-level-threshold-guide.md, the seven CODEX deep-dives and seven SIGNAL topic briefs — plus content/veridian/question-bank/wbs14.ts, so nothing here restates what those already cover. Not affiliated with or endorsed by Pearson Edexcel. Direct quotation from Pearson's own copyrighted wording is kept to short, attributed fragments (under ~15 words) throughout; everything else is paraphrase, and every "seen in N of M" claim below states the real count rather than a likelihood.

#WBS14 Examiner Intelligence — Unit 4: Global Business

#What this document is (and isn't)

Every other document in this folder mines the examiner reports for content — which mechanism was missed, which term was confused with which, which band a specific answer would land in. This one asks a different question: setting the content aside, how do the reports themselves actually talk? Do they reuse the same sentence shapes, the same small set of words, series after series — and if so, what does that reveal about what an examiner is actually looking for when they read an answer?

This only works with documents that contain real prose commentary, not just indicative content. Of the five series the facts bank read in full, only three carry a genuine Examiner's Report or Principal Examiner Feedback: June 2022 (Examiners' Report), October 2023 (Principal Examiner's Report), and January 2024 (Principal Examiner Feedback). October 2021 and January 2023 are Mark Scheme documents only — no narrative commentary exists for either, so they contribute no evidence here. One further series, October 2024, was added later by a single targeted pull that also fetched a genuine Principal Examiner Feedback (for the niche-markets Q2 addendum only, not that series' full paper). These four series — June 2022, October 2023, January 2024, October 2024 — are the entire evidence base for this document. Every pattern below is checked against them and only them; nothing here is inferred from a mark scheme's indicative content alone, and nothing is extrapolated from a series this course hasn't actually read in narrative form.

A pattern found in only one of those four series is labelled a single finding, not a pattern, per this course's own standing rule — and where a sibling document already treats one of those single findings in full, this document cross-references it rather than re-arguing it.

---

#0. A provenance conflict this document will not paper over

Before any pattern claim: the facts bank's own "Trap-taxonomy density pass" (dated 2026-09-25) checked directly and concluded that October 2022 (Q1(c), Q2 and Q3), January 2022 (Q3) and June 2023 (Q3) have no matching Examiner's Report anywhere in this course's archive — Mark Scheme documents only. forge-error-pattern-tracker.md, compiled in this same folder on 2026-09-29, independently repeats that exact conclusion for all five of those series-question combinations.

content/veridian/question-bank/wbs14.ts, also dated 2026-09-25 and itself claiming independent re-verification against primary-source PDFs, nonetheless carries examinerReport-typed fields — source: "examiner-report", verified: true — with full prose text for four of those five: Oct 2022 Q1(c), Oct 2022 Q2, Jan 2022 Q3, and June 2023 Q3. (wbs14.ts has no entry at all for Oct 2022 Q3, the Glovo five-forces question, so that fifth item carries no conflict — only a mark scheme, consistently, in both files.)

This is a real, unresolved conflict between two files this course treats as independently verified, not a gap this document can close from the facts bank's text alone. None of the "examiner report" wording attributed to October 2022, January 2022 or June 2023 in wbs14.ts is used as evidence anywhere below — if those series really do have no genuine Examiner's Report, that wording is either mislabelled Mark-Scheme text or sourced from a document nobody else in this course has independently confirmed exists. Whoever next touches wbs14.ts should re-check its four examinerReport fields against a primary-source PDF before relying on them as genuine report language, in either direction.

---

#1. The recurring hinge: how a report signals "this is what separated the better answers"

Across all four confirmed series, a report's paragraph on a specific question almost never simply says an answer was wrong. It states what the general response did, names the one thing that capped it, and then pivots — using a small, repeated set of near-identical phrases — to what the higher-scoring group did instead.

Most strikingly, the exact same two words open that pivot in two different series, independently re-verified word for word by the facts bank both times:

• June 2022, Q1(c): "Better responses based their answer on the move that would occur from the primary sector to the secondary sector."
• January 2024, Q1(d): "better responses considered whether a business was entering a market to sell or to produce."

The same slot is filled by a close variant elsewhere in the same two series — "those that did make this distinction answered the question well" (June 2022, Q1(d)); "Those candidates that knew the right formula inevitably did well" (June 2022, Q1(a)) — and, in the two series this pull confirms only by paraphrase rather than a marked verbatim quote, by "the strongest answers concluded..." (June 2022, Q2) and "the strongest answers considered..." (January 2024, Q1(e)).

Even where the exact wording differs, October 2023 and October 2024 fill the same functional slot: a report names what the general cohort did, then names the specific thing that moved a smaller group higher — "candidates who did understand the term often 'dismissed it without development'" (October 2023, Q1(e), naming the cap rather than the cure) and "candidates who scored well 'expand[ed] on the extract, often bringing in other examples'" (October 2024, Q2, naming the cure directly).

Seen in: this three-part shape appears in some form in every one of the four confirmed series; the literal words "better responses" are confirmed word-for-word in two of the four (June 2022, January 2024). That second, narrower claim is the one worth calling a pattern rather than a coincidence — a report is not improvising fresh language each series to describe the same kind of gap.

---

#2. Paired verdict-openers: examiners rank a question before they explain it

In every one of the four confirmed series, at least one question's commentary opens with a short, quoted evaluative label for the whole question's performance, before any detail follows — and the labels come in opposite pairs, reused across series rather than invented fresh each time:

"Widely misunderstood" and "widely understood" are the same phrase with the polarity flipped, one series apart from each other. "Tricky" and "straightforward" are opposites applied to the same series' two 12-mark Assess questions, Q1(d) and Q1(e) — the January 2024 report is explicitly ranking its own paper's two same-tariff questions against each other, not scoring each in isolation. October 2023's parallel comparison — that the return-on-investment question was the worse-answered of that series' own two 12-mark questions — is this document's own paraphrase of the facts bank's framing, not a marked verbatim quote, so it is presented here as an echo of the same comparative habit rather than a second confirmed instance of it.

Seen in: the paired-opposite-adjective opener, as a quoted verdict rather than a paraphrase, in all four confirmed series. The specific device of ranking a series' own two same-tariff questions against each other is confirmed by a direct quote in one series (January 2024) and echoed by paraphrase in a second (October 2023) — real, but the two are not equally strong evidence, and this document treats them accordingly rather than flattening them into one claim.

---

#3. What gets rewarded is a conditional verdict, not a side picked

Two of the four series show the same specific reward, in the same specific shape: the report singles out, as what separated the strongest answers, a judgement that depends on a named variable rather than a flat yes/no or either/or.

• June 2022, Q2 (adapting to local tastes: Nike vs. Aldi): the strongest answers concluded the right approach "depended on the product or service in question and its intended market" — rather than arguing that global businesses should always standardise, or should always localise.
• January 2024, Q1(e) (cultural/social factors): the strongest answers, per the facts bank's own paraphrase of the report, considered which products need more or less adaptation in the first place — the same move, applied to a different topic: not "adaptation is worth it" or "adaptation isn't," but a verdict that depends on which product is actually in question.

The June 2022 instance is a marked verbatim quote; the January 2024 instance is this course's paraphrase of the report's content, not a quoted sentence — again, real evidence, held to the weaker of the two standards it actually meets.

Seen in: 2 of the 4 confirmed series. Genuinely a pattern by this course's own rule (at least two series), but the smallest and least certain of the four in this document — worth teaching, not worth overstating.

---

#4. One single finding, for contrast — and where it's already covered

Not everything that looks like a pattern is one. June 2022's Q1(c) report states, of an 8-mark Discuss question with no formal conclusion requirement, that "balance (assessment) is definitely needed" — a real, specific, quotable instruction about how a lower-tariff command word is still marked. It appears in exactly one of the four confirmed series. Nothing in the other three narrows or contradicts it, but nothing in the other three repeats it either, so — per this course's own rule — it stays a single finding, not a generalised claim about how Discuss questions on this paper are marked. forge-vocabulary-precision-list.md §2 and forge-level-threshold-guide.md §1 already treat this quote in full; it is named here only to show what a correctly-labelled single finding looks like next to the three genuine patterns above, not to re-argue it.

---

#What this implies about how examiners are reading an answer

Put the three patterns together and a consistent picture emerges, distinct from "know the content" or "don't forget balance" — both already well covered elsewhere in this folder:

• The report is a diagnosis of a ceiling, not a verdict of right or wrong. Every instance in Pattern 1 names one specific, addable thing — a sectoral distinction, a sell-vs-produce distinction, a specific example beyond the extract — that would have moved an already-adequate answer up a band. That maps directly onto a level-descriptor mark scheme, which is built the same way: each band adds one more specific thing on top of the one below it, not a different kind of answer altogether.
• A report's language about a question's difficulty is about that cohort, that series, that extract — not a fixed property of the topic. Pattern 2's paired opposites (widely misunderstood / widely understood; tricky / straightforward) are reused across series precisely because each series gets to reapply either label to whichever question earns it that time. A topic called "tricky" once is not permanently tricky; a different series can — and, in this small sample, does — call a differently-cased version of a similar demand "straightforward."
• "It depends" beats a side, when the report says so explicitly. Pattern 3 is the smallest claim here, but it points the same direction as the stock closing-paragraph advice already documented in forge-exam-technique-guide.md (add a real conclusion) — except it specifies what kind of conclusion earned credit in the two cases this document can check: one that names the variable the verdict actually turns on, not one that just picks a side.

---

#Sources

research/veridian/WBS14-verified-facts.md — specifically its provenance table (June 2022 = full Examiner's Report; October 2023 = Principal Examiner's Report; January 2024 = Principal Examiner Feedback; October 2021 and January 2023 = Mark Scheme only, no narrative commentary), its per-question citations to those three series' report text actually used above (Q1(a), Q1(c), Q1(d), Q2 of June 2022; Q1(e), Q2 of October 2023; Q1(d), Q1(e) of January 2024), its October 2024 niche-markets addendum (Principal Examiner Feedback, Q2 only), and its "Trap-taxonomy density pass — REVISED 2026-09-25" finding that October 2022, January 2022 and June 2023 have no matching Examiner's Report in this course's archive. Cross- checked for non-duplication against every other file in content/veridian/technique-hub/wbs14/ and against content/veridian/question-bank/wbs14.ts (the source of the §0 provenance conflict). No new primary-source document was fetched for this pass — it is a re-read of report language already verified in the facts bank, organised by how the language itself behaves rather than by topic.