Business Paper 4 — Global Business
Examiner Mindset Guide
Compiled 2026-09-29, from research/veridian/WBS14-verified-facts.md (the primary, provenance-checked
source) and the real Pearson mark schemes/Examiners' Reports it cites, cross-checked against
content/veridian/curriculum/wbs14/ and every other file already in
content/veridian/technique-hub/wbs14/ before writing, specifically forge-exam-technique-guide.md,
forge-level-threshold-guide.md and signal-examiner-intelligence.md. Full citation list at the end.
WBS14 Examiner Mindset Guide — Unit 4: Global Business
What this document is (and isn't)
Three sibling documents already cover adjacent ground, and none of them asks this document's
question. forge-exam-technique-guide.md gives the tariff-by-tariff bands and what each command word
rewards. forge-level-threshold-guide.md works through the exact boundary lines inside those bands,
mark by mark. signal-examiner-intelligence.md studies how a report's language behaves — the
repeated phrases, the paired verdict-openers. All three treat "how this paper is marked" as settled
background and build technique on top of it.
This document asks the background question directly: what is an examiner actually doing, mechanically, when they mark a WBS14 answer — and does that mechanism change across the paper? The honest answer, checked against the facts bank rather than assumed, is that it changes exactly once, at a real tariff boundary, and three quotes from the same real series show the change happening in the examiners' own words.
Every quote below is a short, attributed fragment under Pearson's own copyrighted wording (per this course's standing copyright policy); the connecting sentences are paraphrase. No likelihood, grade boundary, or "confirmed in every series" claim appears anywhere below unless the series count is stated explicitly next to it.
1. Two real marking mechanisms on one paper — checked, not assumed
The task this document was set to answer was: does WBS14 mark by discrete point-tallying (a named mark for each of several components) or by holistic best-fit against a level descriptor (the whole response matched against a band, then placed within it)? Checking rather than assuming turns up both, on the same paper, split cleanly along the mark-tariff line already confirmed elsewhere in this course.
Below 8 marks: discrete component-tallying. The two Section A openers, Q1(a) (4-mark
Calculate/Construct) and Q1(b) (4-mark Explain), are marked as a named split of Knowledge 1 /
Application 2 / Analysis 1 — not a level descriptor at all. The real mark scheme for the diagram
variant of Q1(a) (October 2024, wheat-tariff question) allocates the marks by component: one mark for
correctly labelled axes, up to two for the equilibrium shift, one for the new equilibrium. The
calculation variant (June 2022, GDP per capita) allocates them the same way: one mark for the formula,
one each for the two data points, one for the final figure — with an explicit rule that the right
number without the $ sign loses exactly one of those four marks, nothing more. This is genuine
discrete marking: a candidate can be right about three of the four components and wrong about the
fourth, and the mark scheme prices that outcome exactly.
8 marks and above: holistic best-fit against a level descriptor. Q1(c) (8-mark Discuss), Q1(d) and
Q1(e) (12-mark Assess), and Q2/Q3 (20-mark Evaluate) are marked against level-descriptor tables — three
bands for Discuss (L1 1–2 / L2 3–5 / L3 6–8), four bands for Assess and Evaluate — each one a single
integrated descriptor blending knowledge, application, analysis and evaluation into one holistic
judgement of the whole response, not a separate tally of points earned. forge-level-threshold-guide.md
§0 confirms this directly from the real mark-scheme table structure: there is no internal 6/7 boundary
inside the Assess-12 Level 3 band, for instance, because the whole 5–8 range is one descriptor, not four
one-mark rungs.
The switch happens exactly at the tariff line already confirmed elsewhere in this course, not
gradually. Nothing on this paper sits between the two systems — no 6-mark question exists on WBS14
at all (the spec's own Analyse command, and its 2-mark Define, are never used on this specific paper in
any series this course has read; forge-exam-technique-guide.md §1 already confirms this). A candidate
sitting Q1(a)/(b) is being marked component by component; turning the page to Q1(c) puts them under a
completely different kind of judgement for the rest of the paper.
2. What gets checked first — in the examiner's own stated order, not inferred
Knowing which system applies doesn't say what an examiner looks for first inside it. Two real, independent pieces of evidence answer that for each system.
Under point-tallying: the mechanism/formula is checked before anything built on it. The mark
scheme's own bullet order for both Q1(a) and Q1(b) always lists the Knowledge point — the formula, or
the named benefit — first, before the Application and Analysis marks that follow it. The June 2022
Examiners' Report on the GDP-per-capita question confirms this is exactly how it plays out in
practice: candidates "that knew the right formula inevitably did well," while marks lost by others
were losses of individual components (a missing $, a decimal-place slip) layered on top of a
correct method — never described as a candidate losing the whole question over the formula, but always
as the formula being the first and most decisive thing checked.
Under levels-based marking: whether the response is actually about the right thing is checked before the quality of its assessment is. Two independent series show the same shape. June 2022's Q1(c) report states plainly that the term "employment patterns" was "widely misunderstood," and as a direct result "far too many candidates failed to score any marks at all" — not a lower band, zero credit, regardless of how well-developed or balanced the rest of that answer might otherwise have been. October 2023's Q1(e) shows the same shape on a different sub-point: candidates who misread "likely return on investment" as a quantitative investment-appraisal calculation rather than a qualitative location factor are reported as having "dismissed it without development" before moving on to other, better-understood factors. In both cases the examiner's report is describing the same first gate — is this response actually engaging with the term or concept the question names — being checked before any credit for balance, development or judgement becomes possible at all. Two independent series showing the same shape is this course's own bar for calling something a genuine pattern, not a single finding — met here, but not by more than that.
3. Three quotes from one series, read side by side
The clearest evidence for both of the points above comes from a single series — June 2022 — which happens to carry a full Examiners' Report across all three question types on the same paper. Reading its Q1(a), Q1(b) and Q1(c) commentary in sequence shows the mechanism change directly, in the examiner's own words, rather than inferred from a mark scheme's structure alone.
- Q1(a), 4-mark Calculate (point-tallied): "those candidates that knew the right formula inevitably did well" — a verdict about one named component (K1), reported as decisive on its own.
- Q1(b), 4-mark Explain (point-tallied): context and identification were "generally good," but analysis "often lacked enough development" — a verdict about which specific component (An1) fell short, not a verdict on the answer as a whole.
- Q1(c), 8-mark Discuss (levels-based): "balance (assessment) is definitely needed" even though "there is no requirement to provide a conclusion" — a verdict about a quality that only exists at the level of the whole response, not attachable to any single sentence or fact the way K1 or App2 are.
Put together, these three sentences — from the same examiner, the same paper, the same sitting — show the mindset shift stated plainly rather than left implicit: on Q1(a)/(b) the report talks in terms of which component a candidate got right or wrong; on Q1(c) it talks in terms of a property — balance — that the response either has or doesn't have as a whole. That is what "discrete point-tallying" versus "holistic best-fit" actually looks like from inside a real report, not as two abstract descriptions of a mark scheme.
What this implies, stated directly
- On Q1(a)/(b): defend the named mechanism first. The formula or the named benefit is reported as the decisive, checked-first component; everything documented above is about how much is lost when a later component slips (a sign, a decimal place, under-development), never about the earlier component being irrelevant once later ones are wrong.
- On Q1(c)/(d)/(e) and Q2/Q3: get the question's own key term or scope right before worrying about balance. Two independent series show the same real cost for getting this gate wrong — not a lower band, but "any marks at all" (June 2022) or a point abandoned mid-answer (October 2023) — which is a larger and earlier-acting risk than the balance/development points that dominate technique advice elsewhere in this folder.
Flagged limitations
- This paper's own General Marking Guidance text was never independently quoted verbatim in this course's evidence base. The facts bank confirms this boilerplate section exists and is structurally consistent across every series checked (same copyright notice, same document shape), but nothing in this course's sources reproduces its actual wording — so this document describes the mechanism visible in the level tables and the report commentary, not Pearson's own internal training language for examiners, moderation, or standardisation. That would need a direct quote from that section of a real mark scheme, which this course does not currently hold.
- The "checked first" claim for levels-based marking rests on two series (June 2022, October 2023), not a wider sample. That meets this course's own two-series bar for a pattern rather than a single finding, but it is two of the roughly 5–9 series (of ~15 held since first assessment, June 2020) this course has actually read — stated here rather than generalised further.
- No grade boundary or likelihood figure appears anywhere above. This paper's facts bank carries no
such data (already flagged in
signal-frequency-and-boundaries.md), and nothing here estimates how often a candidate actually loses the "any marks at all" gate in a real cohort — only that the two cited instances are real and independently verified.
Sources
research/veridian/WBS14-verified-facts.md — the Q1(a)/Q1(b) K1/App2/An1 mark-scheme splits and their
June 2022 Examiners' Report quotations (GDP-per-capita and Thailand-rubber questions); the June 2022
Q1(c) Examiners' Report quotation on "employment patterns" and the "balance (assessment) is definitely
needed" line; the October 2023 Q1(e) "return on investment"/"dismissed it without development" citation;
the October 2024 Q1(a) diagram mark-scheme component breakdown (axes/equilibrium/new equilibrium); the
confirmed Section A 4+4+8+12+12 structure and the confirmed absence of Define/Analyse anywhere on this
paper. content/veridian/technique-hub/wbs14/forge-exam-technique-guide.md §1–2 (the paper-wide
tariff/band map this document extends rather than restates). forge-level-threshold-guide.md §0 (the
confirmed single-integrated-descriptor structure and the absence of a real internal boundary inside the
Assess-12 Level 3 band). signal-frequency-and-boundaries.md (confirming this paper's facts bank holds
no grade-boundary or likelihood data).
Not affiliated with or endorsed by Pearson Edexcel. Every quoted fragment above is reproduced only as a short, attributed phrase well under the sentence it comes from; everything else is paraphrase.