Technique Masterclasses

Business Paper 3 — Business Decisions and Strategy

Examiner Intelligence

WBS13 Examiner Intelligence

Provenance — written 2026-09-28. Every claim below traces to research/veridian/WBS13-verified-facts.md (2,057 lines, read in full this pass), which is itself built from primary Pearson mark schemes and examiners' reports (ER), fetched either directly from qualifications.pearson.com or from a genuine local Pearson archive, with one flagged exception (October 2025, mirror-sourced — see that file's own provenance note). This document was also checked against every other file already in this folder — forge-exam-technique-guide.md, forge-error-pattern-tracker.md, signal-frequency-and-boundaries.md, forge-vocabulary-precision-list.md, forge-battle-card.md, the seven codex-deep-dive-*.md files, codex-definitions-and-application-bank.md, forge-exam-day-protocol-extended.md, forge-forty-eight-hour-checklist.md and forge-mock-paper-protocol.md — and against content/veridian/question-bank/wbs13.ts and the seven curriculum lessons in content/veridian/curriculum/wbs13/, to avoid restating what already exists there.

What this document is, and isn't. forge-error-pattern-tracker.md already catalogues what candidates get wrong, topic by topic, as a fillable log. forge-exam-technique-guide.md already tells you what to do about tariffs and command words. Neither asks the question this document asks: read across every examiner's report this course has mined, as a body of writing, and find the phrases and rhetorical moves the examiners themselves keep reaching for — not the individual mistakes, but the shape of how these reports talk, series after series. That shape is itself evidence about how a WBS13 script actually gets read, independent of which topic it happens to be about.

Scope, stated honestly. Twelve WBS13 series have been located for this course's research (Oct 2021 → Oct 2025). Examiners' reports carrying quotable commentary were read, to varying depth, for ten of them: Oct 2021, Jan 2022 (Q1a/b only), June 2022, Oct 2022, Jan 2023, June 2023, Oct 2023, Jan 2024, Summer/June 2024, Jan 2025. No examiner's report has been located or mined for October 2024 or October 2025 — the Oct 2025 material in this course is mark-scheme-only, and no report commentary from that series appears anywhere below. A pattern below is only called a pattern where it recurs in the actual, cited wording of two or more of those ten series; where a finding is real but sits in only one report, that is said plainly. Quotations are kept under ~15 words each and attributed to their series; nothing longer is reproduced, in line with this course's standing citation discipline. No likelihood percentage or grade boundary appears anywhere below.

1. The disclaimer every report leads with: marking is holistic, not a checklist

The single most consistently repeated sentence in the whole mined set. Confirmed, in near-identical wording, in all six of the series whose reports state it directly — Oct 2021, June 2022, Oct 2022, Jan 2023, June 2023, Jan 2025:

"...applied in a holistic way rather than looking for individual Assessment Objectives" (near-verbatim across all six)

with the same warning that a candidate "may only achieve Level 2 if the evaluation is weak" even after attempting every element examiners look for. Nothing else in the ten reports mined is repeated this literally, this often, or this early in the document — it reads as a standing instruction to whoever is about to read the rest of the report, not a comment about one series' candidates.

What it implies: examiners are pre-empting the single most natural student misreading of a levels-based mark scheme — that hitting every bullet point mechanically earns the marks. The phrasing is defensive, almost corrective in tone, as if it is answering a complaint examiners anticipate hearing rather than describing what candidates actually did that series. A script that "does" evaluation without doing it well is being told, in advance, that it will not reach the top band on the strength of attempting it alone.

2. "A conclusion is required... but was not often seen" — a fixed formula for 12-mark Assess

Confirmed across four series, each on a different named question: Oct 2021 (British Airways), June 2023 (Lush, employee share ownership), Oct 2023 (Five Guys, financial rewards), Jan 2024 (Patagonia, transformative leadership). The reports state the rule and its violation in the same structural shape every time:

"required for 12 mark questions... but was not often seen" (paraphrase of the recurring shape; Oct 2023, near-verbatim)

What it implies: this is not a one-off note about one weak cohort — it is stated as a standing expectation Pearson's examiners assume will be missed, four separate times, across three academic years. The tone is matter-of-fact rather than alarmed, which itself is informative: examiners are not surprised when this happens, and appear to have built their own reporting habit — restating the rule, then noting its absence — around the fact that it recurs. A script that reaches a supported judgement in its final two or three sentences is doing something the examiner's own writing suggests most scripts, series after series, do not.

3. The dominant complaint, regardless of topic: answering the framework instead of the specific ask

This is the pattern with the widest reach in the mined set — not confined to one spec sub-section, but recurring across topics that share nothing except being the named, specific element of a question. Confirmed across four series, six separate questions:

What it implies: across four different series and six different questions spanning five different spec sub-sections, the single most frequently named failure mode is not "didn't know the theory" — it is "answered a nearby, more familiar question instead of the one actually asked." Several of these instances involve a taught framework (Ansoff's Matrix especially, named twice as the thing candidates default to) displacing engagement with the extract's specific detail. This reads as examiners distinguishing between candidates who have memorised a model and candidates who can apply it to the one business, and the one decision, actually named in the stem — and treating the second as categorically different from, not just a better version of, the first.

4. The same conflation, named almost identically, on two different questions a series apart

A narrower but sharper version of pattern 3, worth separating out because the wording is close to verbatim across two non-adjacent series on structurally identical questions (both 12-mark Assess, both "will strategy X reduce labour turnover for company Y"):

What it implies: this is not two examiners independently noticing the same generic problem — the underlying mechanism is identical (a reward scheme's motivational effect is not the same claim as its retention effect, and only the second answers a turnover question), and the report-writing follows the same template both times: name the wrong mechanism candidates reached for, then state which one the question actually required. That a marking team reproduces this exact diagnostic structure a full exam year apart suggests it is a known, anticipated confusion for this specific question type on this paper, not an incidental observation.

5. Naming a theorist is explicitly, repeatedly said not to matter

Confirmed across the same two series as pattern 4, in language close enough to be the same boilerplate reused:

What it implies: examiners appear to be correcting a specific, recurring student habit — reaching for a theorist's name as if the name itself earns credit — twice, in near-identical wording, within four months of each other. Combined with pattern 1's "holistic, not a checklist" warning, this reads as the same underlying marking philosophy applied to a smaller unit: naming Taylor is a checklist item, not an argument, and the reports say so almost formulaically when it happens.

6. Copying the extract draws sharper, more disapproving language than a genuine gap in knowledge

Confirmed across two series, on two unrelated questions:

What it implies: set these against pattern 1's dry, procedural tone ("applied in a holistic way...") and pattern 2's matter-of-fact "not often seen." Both of those describe an absence — a missing conclusion, a misapplied framework — in neutral, almost administrative language. "Simply copied" and "just paraphrasing" are different in register: they name an action the candidate visibly chose to do, not a gap in what they knew. The two instances of this specific tone are too few to generalise about examiner attitude broadly, but the contrast within the mined set itself is real: a script that recites the source material back is described, both times it is caught, in noticeably more pointed language than a script that simply reasons weakly.

7. Two questions are sometimes benchmarked directly against their own paper-neighbour

A narrow but genuine pattern, confirmed twice:

What it implies: examiners are not only judging each question in isolation against its own mark scheme — at least twice, they explicitly rank one question's performance against a specific neighbour on the same sitting. This is a small sample, but it shows the reports occasionally carry comparative information that a topic-by-topic reading of the mark scheme alone would not surface: some spec points are flagged as harder relative to what else was on that paper, not just difficult in the abstract.

8. Two visibly different registers inside the same report: rule-bound arithmetic, discretionary prose

This is a structural observation about every report mined, not a single quoted phrase, so it is reported as an interpretation of the body of text rather than a citation count. Every report that covers both a Calculate question and an Assess/Evaluate question shifts vocabulary between the two. For calculation marking, the language is close to algorithmic:

For extended-writing marking, by contrast, the same reports use consistently impressionistic vocabulary — "holistic," "chains of reasoning," "balanced," "an awareness of competing arguments" — language that explicitly resists being reduced to a checklist (see pattern 1).

What it implies: this is this document's own reading of the evidence, not a quoted claim, so it is flagged as interpretation. The two vocabularies suggest two different marking philosophies operating side by side on the same paper: Calculate questions are marked by a fixed, near-mechanical rule a candidate can memorise in advance and apply exactly; Assess and Evaluate questions are marked by a holistic judgement a candidate cannot fully anticipate by memorising a rule, only by practising the kind of reasoning the level descriptors describe. Revision time is not interchangeable between the two — arithmetic precision is worth drilling to exactness, while extended answers need practised judgement, not a longer memorised checklist.

9. Real findings that are NOT patterns — single-series only, stated as such

Per this document's own hard rule, the following are genuine, primary-source examiner findings, but each is currently backed by only one series in this course's mined material. They are recorded here because they are real and specific, not because they meet this document's own bar for a "pattern" — a second occurrence has not been found, and none is implied.

10. What this implies for how to write an answer on this paper

Strictly derived from patterns 1–8 above, not new claims:

  1. A levels-based answer is read for the quality of its reasoning and its balance, not for how many of the "right" elements it mentions — attempting evaluation is not the same as achieving it (pattern 1).
  2. A 12-mark Assess answer without an explicit final judgement is, on this paper's own recent history, the norm rather than the exception among weaker scripts — finishing with one is a distinguishing move, confirmed valuable across four separate series (pattern 2).
  3. The most reliably punished error on this paper is not a missing fact — it is answering the general topic instead of the specific, named element of the question (the specific initiative, the specific outcome asked about), confirmed across four series and six different questions (pattern 3), and twice in near-identical form when the specific outcome is "reduces labour turnover" against the easier, adjacent claim of "increases motivation/productivity" (pattern 4).
  4. Naming a theorist is not, by itself, credit-bearing — twice stated in near-identical terms (pattern 5).
  5. Extract material must be used as evidence for an argument, not reproduced — the language examiners use for this specific failure is unusually pointed compared to how they describe other gaps (pattern 6).

Sources

research/veridian/WBS13-verified-facts.md, read in full (2,057 lines). Ten examiners' reports cited by series, as that file records them: Oct 2021, Jan 2022 (partial), June 2022, Oct 2022, Jan 2023, June 2023, Oct 2023, Jan 2024, Summer/June 2024, Jan 2025. content/veridian/technique-hub/ wbs13/forge-error-pattern-tracker.md, forge-exam-technique-guide.md and signal-frequency-and-boundaries.md, checked for overlap before writing this document — none of the eight patterns above is restated from any of them. Not affiliated with or endorsed by Pearson Edexcel.