IB Computer Science Command Terms: What Each Word Wants From Your Answer
Every IB Computer Science question begins with a command term, and the term — not the topic — decides your answer’s shape. “State” answered with an essay wastes time you need elsewhere; “evaluate” answered with one side caps your marks however accurate it is. The 2027-syllabus papers split marks 50/50 between knowledge/application and analysis objectives, so roughly half your marks depend on answering at the depth the term demands. Here is the working reference.
The quick rule
Read the command term first, decide your answer’s shape second, then fill in content. Depth ladder, bottom to top:
| Depth | Terms | Response shape |
|---|---|---|
| Recall | State, Define, Identify, Name | One precise line. No reasoning, no padding. |
| Understanding | Outline, Describe | Key points / characteristics. “What”, not “why”. |
| Reasoning | Explain, Distinguish, Compare, Contrast | What and why; differences with both sides shown. |
| Judgement | Discuss, Evaluate, Justify, To what extent | Balanced argument plus a conclusion. |
Recall terms: precision beats volume
State and Define want exactness — the precise term, value or meaning, nothing around it. A model answer to “state the purpose of a primary key” is one line long. If your answer to a state-question exceeds two lines, you have misread it or you are hedging; hedging suggests the recall isn’t secure yet. Our smart practice system drills exactly this: instant feedback on definition-level answers, the way flashcards wish they worked.
Understanding terms: characteristics, not causes
Outline = brief account, key points only. Describe = detailed account of features or characteristics. Neither asks why — save the reasoning. Describe-questions about networks or system components reward structured answers: one feature per bullet or sentence, each concrete.
Reasoning terms: where most marks leak
Explain always wants what and why — a mechanism, a because-clause. If your explain-answer has no “because” or “so that”, it is a description wearing the wrong label.
Distinguish (differences only) and Compare (similarities and differences) reward a table — two columns, term by term. Examiners mark against point-lists; a table makes your points impossible to miss and fast to write.
Judgement terms: balance, then commit
Discuss and Evaluate require both sides and then a judgement. “To what extent” wants the same with an explicit position on the spectrum. Justify supports a given conclusion with reasons. The marking bands for extended response (8–10 mark questions, typically Paper 1 Section A) read:
| Band | Marks | Descriptor |
|---|---|---|
| Excellent | 9–10 | Clear, detailed, accurate; addresses all aspects; well-structured with examples |
| Good | 7–8 | Most aspects covered; generally accurate; some detail |
| Adequate | 5–6 | Some aspects addressed; basic understanding; limited detail |
| Limited | 3–4 | Superficial; few aspects; inaccuracies |
| Poor | 1–2 | Very limited; significant gaps |
Note what “excellent” demands: all aspects, structure, examples. A one-sided rant, however confident, lives in “adequate” at best.
How to train this
- Term-tag your practice. Take any past question and, before answering, write the command term’s required shape in the margin (“what+why”, “table, both sides”).
- Self-mark against the shape. Did the answer physically contain a because-clause, a table, a conclusion? If not, redo the shape before improving content.
- Drill recall separately. Definitions are the cheap marks — bank them with spaced drilling so exam time goes to reasoning terms.
- One judgement per essay. Practise writing a one-sentence position, then two paragraphs of support, then the counter, then the verdict. That skeleton fits every evaluate question.
For the full course map — papers, themes, case study timing — see the IB CS complete guide. We teach command-term discipline in every IB CS class, with instant-feedback drills in our smart practice system; the notes page has sample packs, and message us for a trial.