The Standalone Test: Designing Coursework Figures That Survive a Marker's Glance

A marker working through a stack of submissions rarely reads in the order the author wrote. Eyes tend to land on the bold heading, drop down the page, and settle on whatever shape breaks the wall of text. By the time the discussion is read properly, an impression of the visuals has already formed. That is not cynicism on the marker's part; it is simply how documents are scanned. The practical consequence for UK coursework is straightforward: every figure should be built on the assumption that it will be judged before the paragraph that explains it.

Begin by Asking What the Argument Would Lose

The first test of any visual is not whether it looks polished but whether removing it would cost the argument something. Take the figure out and read the surrounding prose. If the discussion still works, the visual was ornamental, and ornament in assessed work invites the question of why the words were spent elsewhere.

A useful distinction is between figures that carry a claim and figures that merely restate one. A bar chart showing that three sources disagree about the scale of a policy's effect is doing real work: it makes a comparison visible at a glance and lets the reader see the spread before the prose unpacks it. A pie chart showing that 'most respondents agreed' when the sentence already says so is a second telling, not evidence.

Ask, too, whether a table would serve better than a chart. Where precision matters — exact coefficients, dated figures, small numbers of cases — a well-set table is usually more honest than a graphic that smooths the detail into a shape. Charts are for pattern; tables are for values.

Captions That Do Not Need the Paragraph

A self-contained caption answers the reader's questions in two or three lines, without requiring them to hunt through the body text. It should identify what is being shown, the population or dataset involved, the place, the period covered, the unit of measurement, and the source.

Convention in most UK departments places table captions above the table and figure captions below the figure. Numbering runs in separate sequences — Table 1, Table 2, Figure 1, Figure 2 — in the order the items first appear in the text, and every numbered item should be referred to at least once in the discussion. A figure that is never mentioned reads as an afterthought, however carefully it was made.

Keep captions descriptive rather than interpretive. The caption says what the reader is looking at; the argument about what it means belongs in the prose. Mixing the two produces captions that strain to make the point and prose that has nothing left to add.

Source Lines and the Referencing Markers Expect

Most UK institutions work from a version of Harvard referencing, commonly Cite Them Right, though departments vary and the programme handbook is the authority. Whatever the exact convention, the principle is constant: a reader must be able to trace the figure back to its origin without guesswork.

Where the figure is original, say so plainly — 'Source: author's own compilation, based on Office for National Statistics (2026) data' — and give the full reference in the reference list. Where a figure has been redrawn or reworked from a published source, 'Adapted from' is more accurate than 'Source', because it signals that the presentation has changed even if the underlying data has not. Where an image is reproduced unchanged, the original creator, title and date should appear, and where copyright or licence conditions apply, the permission statement belongs in the caption rather than buried in an appendix.

If the data reached you through a secondary source — a textbook citing an original survey, for instance — cite what you actually read, and make clear that you have not consulted the primary material. Markers notice the difference, and a tidy reference list will not rescue a figure whose provenance is vague.

The Three Sentences That Make a Figure Legible

Good practice around a figure follows a short sequence. The first sentence directs the eye and names the item, so the reader knows where to look. The second states the reading — the pattern, contrast or anomaly that matters. The third connects that reading to the claim being built in the paragraph.

The common failure is a sentence that simply announces the figure. 'Figure 1 shows the results' tells the reader nothing they cannot see. A more useful construction identifies the salient feature: 'Figure 1 shows that the gap between the two regions widens sharply after 2021, a divergence that the funding formula does not account for.'

Equally, resist the urge to narrate every number. A reader capable of reading a chart does not need to be told each value in turn. Reserve the prose for interpretation, and let the figure look after the arithmetic.

Accessibility Is Part of the Job, Not an Extra

Coursework is increasingly read on screen, and often by people using assistive technology. A few basics protect both readability and marks.

Provide alternative text for every image: a short description of what the figure shows and, where relevant, the key finding. Keep text in figures at a legible size — roughly the same as the body text once the figure is placed at its final dimensions. Check contrast between text and background, and do not encode meaning in colour alone; a chart that distinguishes categories only by red and green becomes unreadable to a colour-blind reader and to anyone printing in greyscale. Different line styles, direct labelling or pattern fills do the same job more robustly.

Avoid three-dimensional charts, unnecessary gridlines and decorative shading. They consume space, reduce precision and have no argumentative purpose.

The Small Mechanics That Undermine Otherwise Good Work

Bar charts should normally start at zero; a truncated axis exaggerates differences and is easy to challenge. Axes need labels and units. Scales should be consistent across panels being compared. State the sample size where it matters, and round consistently. If a figure has been assembled from several datasets, say which parts came from where.

A Last Pass Before Submission

Before the file is uploaded, run through the following. Is every figure referred to in the text? Does each caption identify subject, place, period, unit and source? Is numbering sequential and consistent with the text? Would the figure survive being read alone, on a phone screen, in greyscale, by someone who has not read the surrounding paragraph?

If the answer to the last question is yes, the visuals are doing what they were included to do: arguing, rather than merely occupying space.