Examination Technique — Paper 2

Examination technique — Paper 2 (theory, extended response)

Retrieval: the previous lesson deconstructed Paper 1's practical questions. This lesson applies the same discipline to Paper 2's theory-based extended-response questions, where the challenge shifts from "does the code work" to "does the written answer genuinely reach the command word's demand, in the space and time available."

Key vocabulary

• Extended response — a longer written answer (typically worth 4+ marks) requiring sustained reasoning, not a single fact or short calculation.
• Point–evidence–explain (or similar structuring) — a common way to build a strong extended response: make a point, support it with a specific detail/example, then explain its significance to the question actually asked.

Understand — what separates a strong extended response from a weak one

A weak extended response often lists correct facts without connecting them to the specific question asked. A strong one is selective (uses only what's relevant to the command word and scenario given, not everything the student happens to know about the topic) and applied (every point is tied back to the specific scenario in the question, not stated in the abstract).

Compare it — two responses, one question

Question (6 marks): "Evaluate whether a relational database or an in-memory Python dictionary is more appropriate for a small charity's one-off, single-event donation tracker (a few hundred donations, no requirement to keep records after the event)."

Response A: "A relational database stores data in tables with rows and columns, and uses SQL to query it. It supports ACID transactions, relationships between tables via foreign keys, and can enforce constraints. A Python dictionary stores key-value pairs in memory and is simpler to use for small programs."

Response B: "A dictionary is more appropriate here. The stated scale (a few hundred donations) is small enough that a dictionary's in-memory operations are more than fast enough, with none of a relational database's setup overhead. Persistence isn't required either — the brief explicitly says records don't need to survive after the event — which removes the main reason a database's durability would matter. A database becomes clearly justified once persistence, larger scale, or concurrent multi-user access genuinely apply, none of which this brief establishes."

Identify which response is stronger against the evaluate command word specifically, and explain why.

(Response B is stronger. Response A is accurate but entirely descriptive - it states facts about both options without ever weighing them against the SPECIFIC scenario given (small scale, no persistence requirement) or reaching a judgement; it would score well against a "describe" or "define" command word, but this question asked for "evaluate." Response B is genuinely evaluative: it directly ties two named features of the scenario (small scale, no persistence requirement) to a reasoned recommendation, and states the general condition under which the opposite conclusion would hold - exactly what "evaluate" demands: weighing evidence against stated criteria to reach a supported judgement.)

Structure it — building your own extended response

For the same question above, a strong response structure might be: state your overall judgement first (which option, briefly) → give your strongest reason, tied to a specific detail from the scenario → give a second reason or a genuine counter-consideration → state the general condition under which your judgement would change. This isn't the only valid structure, but writing your conclusion first (rather than building up to it at the very end) makes it far easier for a strong answer to be recognised as evaluative from the first sentence, rather than only becoming clear in the final line.

Common mistake

Treating "extended response" as "write everything you know about the topic." A 6-mark evaluate question rewards depth and application to the SPECIFIC scenario given far more than breadth — a shorter response that directly answers the actual question asked will typically outscore a longer one that recites general facts about the topic without applying them.

Check your understanding

A question asks students to "Analyse" why a particular sorting algorithm's performance degrades on a specific type of input. State one thing an analytical answer must do that a purely descriptive answer of the same algorithm would not. (An analytical answer must break down HOW or WHY the specific input causes the degradation - e.g. tracing through the mechanism that causes extra comparisons/swaps - rather than simply describing what the algorithm generally does step by step, which a descriptive answer would do without explaining the specific cause of the performance change.)

Challenge

Write your own 6-mark "evaluate" question on any theory topic from this course (data representation, computer systems, networking, databases, big data, functional programming, or ethical/legal/cultural issues), then write both a deliberately weak (descriptive) response and a strong (genuinely evaluative) response to it, in the style of the worked comparison above.

Looking ahead: the final lesson isn't about answering more questions — it's about how to read your own results back and turn them into a genuine improvement plan.