IGCSE/GCSE Computer Science: Theory That Gets Marked

Most students lose marks on IGCSE and GCSE Computer Science not because they don't understand the topic, but because they answer the question they think was asked instead of the one on the page. This subject is graded against a mark scheme with specific keywords and command words, and the exam boards (Cambridge, Edexcel, OCR, AQA) train examiners to award points only when those keywords show up in context.

Command words decide your structure

'State' wants one fact, no explanation, one mark. 'Explain' wants a reason with a consequence attached, usually two marks: point plus justification. 'Describe' wants a sequence or a set of features written out step by step. Students who write an 'explain' answer at 'state' length, or a one-liner for a 4-mark 'describe' question, cap their own score before the examiner even judges accuracy.

Take a classic: "Explain why a solid-state drive is more suitable than a hard disk drive for a laptop." A 'state' answer says "SSDs are faster." That's one mark at best. The 'explain' version needs the mechanism: "SSDs have no moving parts, so they are less likely to be damaged if the laptop is dropped, which matters because laptops are portable and get moved around." Two ideas, two marks, cause linked to effect.

The keywords the mark scheme actually wants

For binary and data representation questions, examiners look for precise terms: 'most significant bit', 'overflow', 'two's complement', 'mantissa', 'exponent'. Saying "the number gets too big" instead of "overflow occurs because the result exceeds the number of bits available" will not get the mark even if your understanding is correct in spirit.

Same pattern in networking topics. "The router directs packets" is vague. "The router uses the destination IP address to forward packets between networks" hits the specific vocabulary the scheme checks for. For security questions, don't just say "hackers can steal data" — name the attack type: SQL injection, phishing, brute force, and pair it with the matching defence (parameterised queries, two-factor authentication, account lockout policies).

Pseudocode and algorithm questions reward exact syntax

Cambridge IGCSE uses its own pseudocode style (IF/THEN/ENDIF, FOR/TO/NEXT, DECLARE, INPUT, OUTPUT) and marks are given line by line for correct structure, not just correct logic. A working Python-style solution written in an exam that expects Cambridge pseudocode syntax can lose marks even though it would run fine on a real machine. Learn the exact keyword set your specification uses and write in it consistently, including indentation for nested loops and conditions, because examiners follow indentation to track scope.

For trace tables (a staple in both exam boards), write out every variable's value at every iteration, including the ones that don't change. Skipping a row because "it stayed the same" is a common way to lose a mark on a 4-mark trace question, since the scheme checks the full row against the expected output.

Practical revision that maps to the mark scheme

• Get past papers with the official mark scheme (not just the paper) directly from the exam board site and mark your own answers against it line by line.
• Build a glossary of 3-5 words per topic — cache, fetch-decode-execute, normalisation, checksum — and drill definitions until you can write them from memory in the exact phrasing the scheme rewards.
• Practice writing 'explain' answers as two sentences: fact, then consequence. Force the habit before the exam so it's automatic under time pressure.
• For programming papers, hand-trace your own code before typing it, since many marks are for showing working, not just final output.

Where students still lose easy marks

Units. "Storage capacity increases" without saying from what to what, in what unit (GB, TB), rarely gets full marks on a calculation question. Also watch number base questions — convert to decimal and back, and show every step, because method marks exist even if the final answer is wrong.

Computer Science at this level rewards precision over cleverness. The syllabus content is not deep, but the phrasing discipline is unforgiving, and that's exactly the skill worth practicing before results day matters more than the day before.

If you want structured walkthroughs of these exact topics — binary, networking layers, pseudocode syntax, security threats — check the related Computer Science and Tutoring segments on Korra Studio.