Assignment Verb Decoder: what 'analyse' vs 'evaluate' actually asks

The prompt decoder

A prompt like "Describe the two models, then evaluate them against the evidence" contains two jobs of very different size. Describe is a level-2 errand — accurate coverage. Evaluate is a level-5 verdict — criteria, weighing, commitment. The top band lives almost entirely in the second clause, yet the describe clause is where most answers spend most of their words. The decoder finds every directive verb in your prompt, places each one on the ladder, marks the highest as the primary job, and turns the whole stack into a checklist: one line per verb, ordered so the top verb gets the biggest share of your plan.

Two reading habits make the decoder unnecessary for most prompts after a few runs. First, find every verb before deciding on any of them — the "then" and "and" clauses are where markers park the marks. Second, budget backwards from the top verb: on a 2,000-word answer to "describe, compare, evaluate", a sound split gives the evaluation more than half the words and the description barely a fifth, because description is worth level-2 marks and evaluation carries the band. When the top verb is argue or construct, the thesis has to carry a position from the first line — the thesis statement generator drafts one you can push against.

The verb ladder: Bloom's taxonomy in plain language

Bloom's taxonomy is a ranking of thinking skills, first published in 1956 and revised in 2001, that runs from remembering facts at the bottom to creating new work at the top. Markers borrow it because it converts "quality of thinking" — which is invisible — into command verbs, which are checkable. Each rung up the ladder asks the writer to do something harder with the same material: repeat it, restate it, use it, take it apart, weigh it up, build with it. Mark schemes translate the ladder into bands, so the verb a prompt uses is a fairly reliable preview of where the marks sit.

Two caveats keep the ladder honest. First, it is a map, not a law: some boards pull discuss up to level 5, and a few mark schemes treat explain as a level-2 word. When your module's own command-word list exists, it beats any general table. Second, the ladder says nothing about difficulty of content — a level-1 list question on a hard topic can be harder to answer than a level-5 judgement on an easy one. What the ladder tracks is what the marker is allowed to reward: a perfect answer to the wrong rung still collects the wrong rung's marks.

What each jump up the ladder looks like in a paragraph

Levels are easiest to see in pairs. Each pair below takes the same material — all invented and labeled illustrative — and shows the lower-rung answer, then the higher-rung answer. The content is identical; the verb's work is what changes.

Describe → Analyse

Describe: "The company's sales fell for three quarters in a row. Its costs rose over the same period, and two competitors entered the market with lower prices."

Analyse: "The sales fall tracks the competitor entries more closely than the cost rise: both entrants launched at lower price points in exactly the two regions where volume dropped first, which points to price competition — not the cost base — as the operative mechanism."

What changed: the describe version inventories three facts and stops. The analyse version puts the facts in contact with each other and names the mechanism they jointly suggest. Same case, one rung apart. The analytical essay guide turns this move into a repeatable paragraph plan.

Explain → Evaluate

Explain: "The clinic cut its waiting list because it opened weekend sessions and gave senior nurses the authority to triage patients without a doctor first."

Evaluate: "Judged against the brief's own criteria — speed and fairness — the clinic's reform is only partly a success. Weekend sessions shortened waits, and triage authority cut them further for urban patients; but rural sites could not staff weekends, so the reform widened the gap it was meant to close. On balance it succeeds on speed and fails on fairness."

What changed: the explain version answers "why did it work?" The evaluate version asks "how well, against what standard, and where does it stop working?" — then commits to a verdict. Weighing evidence and landing the judgement is the core craft of the discussion section in any report.

Discuss → Argue

Discuss: "Some researchers argue that exam coursework should be replaced entirely; others say exams are the fairest measure. There are good points on both sides."

Argue: "Coursework should weight no more than a third of the final grade: it measures stamina rather than attainment, its fairness collapses the moment help moves from the classroom to the chat window, and a capped share keeps its genuine benefit — testing sustained work — without letting it replace the only format every candidate faces equally."

What changed: the discuss version serves two positions and stops. The argue version owns one, states it in the first sentence, and spends the rest of the answer earning it — the whole craft of the argumentative essay.

'Compare and contrast' done properly

Compare and contrast is one instruction, not two, and the marks sit in the structure it implies. The failing version reads as two mini-essays: everything about subject A, then everything about subject B, with a sentence of comparison stapled at the end. The passing version runs the comparison through criteria: pick two or three dimensions both subjects share, put both subjects under each dimension, and close by judging which differences matter most for the question you were actually asked.

1. Set the criteria. Two or three is plenty, and they should come from the question or the case, not from a generic list. A question about growth wants growth-relevant criteria; a question about risk wants risk-relevant ones.
2. Run both sides per criterion. Each criterion paragraph names the criterion, shows subject A on it, shows subject B on it, and notes where they genuinely differ rather than where they merely differ in detail.
3. Judge the differences. End by ranking the differences by significance: which one would actually change a decision, and why. That ranking is the "contrast" half earning its keep.

The table is a worked shape, not content to transplant: your criteria must come from your prompt. For the full essay version of this structure — thesis options, paragraph orders, and the transitions that hold a comparison together — see the compare and contrast essay examples.

The 'to what extent' formula

"To what extent" is evaluate wearing a costume, and it has a four-sentence formula that fits any subject. Sentence one states your dial position — "largely", "partly, and only after 1945", "less than the standard account allows". Sentence two gives the strongest case for the high reading. Sentence three gives the strongest case against it, honestly. Sentence four re-states the dial with the weighing shown: "largely, but only where the roads were already built." An illustrative opener for a history-style question:

"The printing press mattered greatly — but as an amplifier, not a cause. (Dial, stated up front.) Its strongest claim is speed and reach: an argument that once travelled at the pace of a rider now travelled at the pace of a print run. (Case for.) Yet the same press served regimes that opposed reform, so distribution explains reach, not adoption. (Case against, kept honest.) The extent is therefore large wherever urban readership existed, and thin at the margins — an answer the essay then defends paragraph by paragraph. (Dial, re-read.)"

The formula's discipline is the commitment. Fence-sitting — "there are arguments on both sides" as a conclusion — is the single most common way extent questions drop out of the top band, because refusing to read the dial is refusing the exact skill the verb names. If your extent answer ends in "it depends", you have written the setup for an evaluation and then skipped the evaluation.

How markers weight rubrics

Rubrics rarely say "the verb is worth 40%". They say it in bands: a top band that requires judgement, a middle band that requires analysis, a lower band that rewards accurate coverage. The weighting below is an illustrative shape for a 2,000-word essay prompt that stacks describe, compare and evaluate — your actual rubric overrides it, and where the brief prints mark allocations, those numbers are the weighting.

Three practical consequences follow. First, mark-allocation arithmetic beats intuition: on a 30-mark question, roughly three marks per paragraph, and if the last clause of the prompt carries the top band, it deserves the most paragraphs. Second, a rubric's top band is written in the language of the highest verb — find the verb, and the top band tells you exactly what it wants to read. Third, stacked prompts punish uniform paragraphing: sections that serve the top verb get more room, and the setup sections get tightened, not deleted. The same weighting logic applies to reports, where the discussion section typically carries the evaluation marks the introduction and method cannot.

Higher-order drill

Eight pairs, one question each: which verb sits higher on the ladder? Answer fast, read the one-line reason, and beat your best score — the best score survives a reload, the answers do not. All pairs are illustrative teaching pairs, not from any real paper.

Quiz: five questions on command verbs

Five questions on reading a prompt like a marker — the ladder, the stacked verbs, the two famous costume verbs. Every wrong answer explains itself, and four options keep the guessing honest.

Frequently asked questions

Did this page change how you read your last prompt? 👍 Yes 👎 Not really