Critical analysis vs description: the difference that decides your grade

If you have ever received a mark in the high fifties with the comment 'well researched but largely descriptive', you have run into the single most common ceiling in UK higher education. It is frustrating because the work is not wrong — the reading is done, the sources are real, the writing is clear. It simply does not do the thing the marking criteria are measuring.

The gap between description and analysis is not a matter of vocabulary or confidence. It is a matter of what your sentences are doing. This article shows the difference at sentence level, gives you a test you can apply to any paragraph in ten seconds, and sets out the four moves that reliably convert one into the other.

What markers mean by 'too descriptive'

Description tells the reader what exists: what a theory says, what a study found, what a case decided, what happened. It is accurate, necessary and worth roughly a 2:2. Analysis tells the reader what to make of it: why the finding is more or less convincing, what it implies, how it sits against a rival account, where it stops working.

The reason description feels like enough is that it takes real effort to produce. You read four articles and summarised them faithfully — that is work. But from the marker's side, description demonstrates only that you understood your sources. Every criterion above about 60 per cent asks for something more: judgement, evaluation, the ability to weigh.

Look at any UK marking rubric and the verbs give it away. The lower bands say 'describes', 'outlines', 'demonstrates knowledge of'. The upper bands say 'evaluates', 'critiques', 'synthesises', 'justifies'. Those verbs are instructions. Your sentences should be able to answer which verb they are performing.

The ten-second test: could a well-read stranger disagree?

Take any sentence in your draft and ask whether an informed person could reasonably disagree with it. 'Herzberg distinguished between hygiene factors and motivators' — no one can disagree; that is description. 'Herzberg's two-factor model survives in practice mainly because it is easy to teach, not because the evidence supports the separation' — that is contestable, and therefore analytical.

Apply the test to a whole paragraph and the diagnosis becomes obvious. Highlight every sentence that fails the test. If more than half your paragraph is highlighted, the paragraph is a summary with a conclusion bolted on, and no amount of stronger phrasing will change that.

The point is not to eliminate description. It is to make sure description is doing a job — setting up a claim that follows — rather than filling the space where the claim should be.

Before and after: six rewritten sentences

The pairs below are drawn from real coursework across several disciplines, lightly anonymised. In every case the 'after' version uses the same source and the same knowledge — nothing extra was read. What changed is that the sentence now takes a position, and the position is defensible.

• Descriptive: 'Maslow's hierarchy of needs consists of five levels, beginning with physiological needs.' → Critical: 'Maslow's hierarchy remains a fixture of management teaching despite thin empirical support, which suggests its appeal is pedagogical rather than explanatory.'
• Descriptive: 'The study used a sample of 120 undergraduate students from one university.' → Critical: 'Drawing all 120 participants from a single undergraduate cohort limits the claim to a population with unusually homogeneous working patterns, which matters because the paper generalises to the wider labour market.'
• Descriptive: 'Donoghue v Stevenson established the modern law of negligence.' → Critical: 'Donoghue is routinely cited as the origin of the modern duty of care, but the neighbour principle did far less work in the following decade than that framing implies; the expansion came later, through Anns and its retreat.'
• Descriptive: 'Inflation rose sharply during 2022 and 2023.' → Critical: 'The 2022–23 inflation episode is difficult to attribute to a single cause, and the disagreement between energy-shock and demand-side accounts matters because the two imply opposite policy responses.'
• Descriptive: 'Both articles discuss the effects of social media on adolescent wellbeing.' → Critical: 'Both articles report a negative association, but only one measures usage objectively; the gap between self-reported and logged screen time is large enough to account for much of the difference in effect size.'
• Descriptive: 'The Nightingale wards were designed to improve ventilation.' → Critical: 'The ventilation rationale for Nightingale wards was constructed retrospectively as miasma theory lost credibility, which complicates the usual account of design following evidence.'

Four moves that create criticality

Criticality is not a personality trait, it is a set of repeatable moves. Four of them will cover almost everything you need in undergraduate and taught postgraduate work.

The first is comparison: put two sources in the same sentence and say how they differ. The second is evaluating method: ask how the knowledge was produced and what that method can and cannot show. The third is identifying assumptions: name what the author took for granted, and ask what happens if it does not hold. The fourth is drawing implications: state the consequence of the finding being true, especially where it is inconvenient.

Notice that none of these requires you to invent new knowledge or to be cleverer than published academics. They require you to be specific about what is on the page in front of you. That is why criticality improves quickly once you start practising it deliberately — it is a habit of attention, not a gift.

• Compare: 'X argues A, whereas Y, using the same dataset, concludes B because…'
• Evaluate method: 'Because the design is cross-sectional, the association cannot support the causal claim made in the discussion.'
• Surface assumptions: 'This depends on treating engagement as a proxy for learning, which the study does not defend.'
• Draw implications: 'If that holds, the policy recommendation in section 5 addresses the wrong mechanism.'

Where description still belongs

Over-correcting is a real risk. Students who have just been told to be more critical sometimes strip out the context, and the result is assertion — strong claims about material the reader has not been given. That scores no better than description, and sometimes worse, because it looks unevidenced.

The workable ratio in most humanities and social science writing is roughly one part description to two parts analysis, arranged so the description comes first and briefly. Give the reader precisely what they need to follow the point, then spend the rest of the paragraph doing something with it.

In some contexts description carries more weight legitimately: a methods section, a case summary in a law essay, a reflective account in a nursing portfolio. Even there, selection is analytical — what you choose to include and omit reveals judgement, and a well-chosen three-sentence case summary is a critical act.

A red-pen pass for your own draft

Do this in the last hour before you submit, on paper or with the highlighter tool, and treat it as a mechanical process rather than a creative one. It takes about twenty minutes for 2,000 words and it converts marks more reliably than any other single edit.

• Highlight every sentence a well-read stranger could not disagree with.
• For each paragraph, find the contestable claim — if there isn't one, write it now, at the end.
• Delete or compress any description that does not set up a claim in the same paragraph.
• Check every paragraph ends with your judgement rather than a citation or a summary.
• Search the draft for 'shows', 'states', 'discusses' and 'talks about' — these verbs almost always mark a descriptive sentence that could be doing more.
• Read the topic sentences alone; if they are a list of authors, rewrite them as claims.