Statistics 1

Diagram Masterclass

#FORGE Diagram/Graph Masterclass — WST01 (Statistics 1)

Provenance (2026-09-29): Read for this document — research/veridian/WST01-verified-facts.md in full, grepped specifically for "diagram", "graph", "histogram", "scatter", "whisker", "B1", "draw" and "label"; the two existing per-topic deep-dives that already own diagram content in depth, codex-deep-dive-outliers-boxplots-comparing-distributions.md and codex-deep-dive-sampling-tree-diagrams-venn-diagrams.md; the regression deep-dive's own §5 negative finding, codex-deep-dive-regression-gradient-interpretation-extrapolation.md; and forge-exam-technique-guide.md (its §2.1 mark-type definitions and §4 checklist). No new PDF fetches were made. This document adds one thing none of the above already does: a single cross-cutting answer to "does drawing/labelling THIS diagram earn its own mark on WST01, and if so exactly how" — organised by diagram type rather than by topic — while sending every reader who wants the full worked mechanism for one diagram type to the deep-dive that already owns it, so nothing below is a second copy of material that already exists elsewhere in this repo.

#Verdict: yes, but only for some diagram types — not a blanket rule

WST01's spec (2.1) says outright that "drawing of histograms/stem-and-leaf/box plots will not be the direct focus of exam questions — reading/using them will be." Taken alone, that sentence would suggest diagrams never carry their own marks on this paper. The real mark schemes contradict that for two diagram types specifically — box plots and tree diagrams — where drawing/labelling the diagram itself is exactly what a real, verified mark rewards. For two other diagram types the spec explicitly permits — histograms and scatter diagrams with a fitted regression line — the reviewed record (14 mark-scheme series, 6 with a read examiner report) contains zero confirmed instances of the drawing itself being separately marked. Venn diagrams sit in between: filling in a given diagram correctly is real, evidenced mark-scheme territory, but constructing one from a blank grid is not what this paper actually tests.

#1. The general mechanism: how a diagram earns a mark at all

The mark schemes' own "General Marking Guidance" (already quoted in forge-exam-technique-guide.md §2.1) defines a B mark as "independent accuracy marks where there is no method (e.g. often given for a comment or for a graph)." That parenthetical is the whole basis for every diagram-specific mark below: a correctly-drawn or correctly-labelled diagram feature can earn a mark standing entirely on its own, with no method line required and no dependency on any other part of the question. Several of these B marks are also ft (follow-through) — credited using the candidate's own earlier (possibly wrong) values, provided the diagram is built from them correctly.

#2. Box plots — the paper's most heavily diagram-marked skill

This is the strongest real evidence of separately-creditable diagram marks anywhere on WST01. Two independently verified anchors:

Jan 2021 Q2(d) (box-plot construction from a stem-and-leaf-derived five-number summary): a first A1 for the box and both whiskers drawn from the candidate's own values, and a second A1, verbatim, "for upper whisker ending at 97 and lower whisker ending at 54 or 50 and only one outlier, shown at 43 / Allow ± 0.5 of a square for accuracy" — a real, quoted drawing-tolerance rule, and a genuine asymmetry (the lower whisker is credited at either the true value or the fence itself; the upper whisker only ever at the true value).
June 2019 Q2(c) (redraw a box plot after two data values change): a full 7-mark sequence — B1/M1/A1/M1/A1ft/A1/A1 — where whisker positions, the shifted quartile, and the outlier count are each their own credited line.

The full mechanism (the fence-direction rule, the whisker-termination asymmetry, the recompute-cascade proof, and both anchors worked in depth) is already covered in codex-deep-dive-outliers-boxplots-comparing-distributions.md §§1–2.4 — not repeated here.

#3. Tree diagrams — the branch is the mark

Jan 2023 Q2(a) (two-bag transfer, without replacement) mark scheme, verbatim: "B1 for 8/13 & 5/13 in the correct place on the 3rd branches" and "B1 for 7/13 & 6/13 in BOTH correct places on the 3rd branches." Two real, separate B marks for correctly placing values on specific branches — and the second bullet credits the same pair of fractions regardless of which of the two physical transfer orders produced them, because both orders rebuild the destination bag identically. This is a genuine diagram-labelling skill, distinct from the arithmetic that follows it: a student can lose these marks even while going on to compute a correct final probability, and can earn them even on a part they don't finish. Full derivation (why either order is credited, and the general counting reason) is in codex-deep-dive-sampling-tree-diagrams-venn-diagrams.md §2.3 — not repeated here.

#4. Venn diagrams — filling in a given diagram is tested; drawing one from scratch is not

Two real, distinct skills, both about a diagram the question already supplies:

The blank-region convention (Jun 2022 Q4(c), verbatim): "some candidates left out the 0 in the outside region of N and should be reminded that blank spaces are not assumed to be 0s in Venn diagrams." This is a labelling convention with real marks riding on it — an empty region is read as "not attempted," not as zero.
Adding a new region to a given diagram (Oct 2021 Q1(d), verbatim): "On the Venn diagram show a possible position for the event D" — B1 for a suitable event D drawn with a valid overlap pattern. This is the one place on this paper where a student draws a genuinely new diagram element, not just reads or completes given values.

Both are cross-referenced (not re-derived) in codex-deep-dive-sampling-tree-diagrams-venn-diagrams.md §§3.2 and 3.4. No anchor in the reviewed record has a student constructing a full Venn diagram's region layout from a blank grid — every real question supplies the diagram (or most of it) and marks what the student does with or to it.

#5. Where the spec permits a drawn diagram but no evidence of a separate mark exists

Two genuine negative findings, already established elsewhere in this repo and not re-derived here:

Scatter diagrams and the fitted regression line. Spec 4.1's own guidance states "students may be required to draw it on a scatter diagram." codex-deep-dive-regression-gradient-interpretation- extrapolation.md §5/§6 already checked this specifically and found zero confirmed worked instances across all 14 mark-scheme series and all 6 read examiner reports for this unit. The skill is real and spec-named; it is simply unevidenced in the material this course has verified so far — stated there as a fact, not a likelihood, and repeated here only so a reader of this masterclass doesn't have to re-discover it.
Histograms. Spec 2.1 explicitly states histogram-drawing "will not be the direct focus of exam questions." The one confirmed histogram angle in the reviewed record (Jan 2023 Q1(a)) tests reading a given histogram correctly — recovering frequencies from area, not height — not constructing one.

Stem-and-leaf diagrams follow the same pattern as histograms: every confirmed mark-scheme angle (finding a missing leaf digit from a given statistic, reading $Q_1$ / $Q_3$ off a diagram) is a reading or inference skill applied to a diagram the question already supplies, never a construction mark — consistent with spec 2.1's guidance and covered in depth by codex-deep-dive-reading-and-comparing- distributions's own lesson content, not duplicated here.

#6. Don't confuse a self-help sketch with a marked diagram

forge-exam-technique-guide.md §5 already documents Pearson's own repeated advice to sketch a Normal curve on any Normal-distribution question, quoting the examiner report directly: "a simple diagram would have helped many to avoid this error" (Jan 2023 Q5(a)) and "an error that could probably have been avoided if they had drawn a suitable diagram" (Jan 2021 Q3(b)). This is genuinely useful advice — but it is advice about avoiding a different, already-marked error (subtracting from 1 wrongly, mis-signing a $z$ -value), not evidence of a diagram-drawing mark in its own right. No mark scheme anywhere in the reviewed record awards a B mark for the sketch itself on a Normal-distribution question. Treating that sketch as if it earned marks the way a box-plot whisker or a tree-diagram branch does would be the exact mistake this masterclass exists to prevent.

#Sources

research/veridian/WST01-verified-facts.md (spec text §1; the Jan 2021 Q2, June 2019 Q2(c), Jan 2023 Q2(a), Jun 2022 Q4(c) and Oct 2021 Q1(d) mark-scheme/examiner-report citations, each independently verified against the primary PDF in the facts bank's own per-lesson audit sections). content/veridian/technique-hub/wst01/codex-deep-dive-outliers-boxplots-comparing-distributions.md, codex-deep-dive-sampling-tree-diagrams-venn-diagrams.md, codex-deep-dive-regression-gradient-interpretation-extrapolation.md, and forge-exam-technique-guide.md (cross-referenced throughout, none of their content reproduced beyond short attributed quotes).