JCE Computer Studies common mistakes
JCE Computer Studies · Understand the traps and how to avoid them.
When converting between binary, octal, hexadecimal and decimal, candidates convert the digits left-to-right (most significant first) and treat each hex digit as a single decimal digit, e.g. writing 0x2A as 2×10 + A instead of 2×16 + 10.
School arithmetic trains left-to-right reading and base-10 place values, so candidates reuse the decimal algorithm without rebuilding the place-value weights for the new base. Hexadecimal letters A–F are also treated as symbols rather than as the values 10–15.
What to do instead
Always convert right-to-left using powers of the base and write the place-value weights above each digit before converting (…16², 16¹, 16⁰). Convert any hex letter to its decimal value first. Check the answer by converting back the other way, and memorise the 4-bit nibble table (0000–1111 ↔ 0–F) so binary↔hex becomes direct grouping rather than arithmetic.
Candidates confuse the roles of RAM and ROM, or of the CPU, ALU and control unit, when describing the fetch–decode–execute cycle or the function of a component in a system diagram.
The topics are taught as vocabulary lists rather than as a working pipeline, so the terms are stored as isolated labels. Candidates also carry over loose everyday usage ("memory" meaning storage) into technical answers.
What to do instead
Draw the cycle as a labelled loop: PC → MAR → memory → MDR → CIR → decode → execute → PC increment, and state which register or bus is used at each step. For each component, write one sentence on what it stores or does and one on what it connects to. Practise explaining why RAM is volatile and ROM is not by reference to how each is written during normal operation.
When asked to trace or write an algorithm (loop, nested loop, or a search/sort), candidates step through only one or two iterations and then state the final output, missing off-by-one errors, the last pass, or the swap that occurs on the final comparison.
Tracing is mentally demanding and candidates shortcut to pattern-matching the intended result. They also confuse counting-loop bounds (does it run n or n−1 times?) and forget that a swap changes the values used in the next comparison.
What to do instead
Use a trace table with one column per variable plus an output column, and fill one row per iteration or comparison, no exceptions. Mark the loop condition at the top of each row and stop only when it evaluates false. For sorting, trace the actual array after every swap. Verify the count of iterations against the loop bounds before writing the answer.
Candidates write pseudocode or program code with mismatched constructs: a WHILE loop that never changes its condition variable, IF without ENDIF, a FOR loop with the wrong increment direction, or variables used before assignment.
Candidates copy the visual shape of code from memory rather than its grammar, and they draft directly in the answer space without planning the control flow. They also assume a variable starts at zero when the language or pseudocode standard does not guarantee initialisation.
What to do instead
Before writing, list the inputs, the variables and their initial values, and the loop's termination condition. Use a fixed indentation pattern and always close every block you open. After writing, dry-run the code with a small concrete input (n=3) and check the loop actually terminates. Learn the specific pseudocode conventions your syllabus specifies and use only those.
On database questions, candidates draw an entity-relationship or table design that stores repeating groups in one field (e.g. a single 'Subjects' column holding several values) or omits a primary key, then cannot answer the normalisation follow-up.
Candidates model the data as it appears on a paper form or spreadsheet rather than as relations. Normalisation is memorised as definitions (1NF, 2NF, 3NF) without practising the decomposition step, so the abstract rule is not linked to a concrete table.
What to do instead
Apply the test 'one value per cell, every row identifiable by a key' to every table you draw. For each repeating group, create a separate table with a foreign key back to the parent. Practise taking one messy table and writing the 1NF, 2NF and 3NF versions side by side, naming the key and the partial/transitive dependency removed at each step.
Candidates answer network questions by giving a device or layer name only ("a router", "layer 3") without stating what it does to the data, or they mix up the roles of switch, hub, router and gateway and the TCP/IP versus OSI layer numbering.
Device names are learned as a list, and the OSI model is memorised as seven words rather than as a set of responsibilities. The two models overlap in teaching materials, so layer numbers get attached to the wrong stack.
What to do instead
For every device, learn the triple: what address it reads (MAC vs IP), what it does with the frame or packet, and which layer it operates at. Build a one-page table comparing hub, switch, router and gateway on those three points. When a question asks for a layer, always add the protocol data unit (bits, frame, packet, segment) as a check.
On security and ethics questions, candidates list a control by name ("use a firewall", "encrypt it", "get consent") without saying what threat it addresses, how it works, or who is responsible, so the answer earns no applied marks.
These topics are often revised as slogan lists close to the exam. Candidates also answer the ethical scenario as a personal opinion rather than applying a named principle or a legal/ethical framework.
What to do instead
Structure every security answer as threat → mechanism → effect (e.g. 'phishing via email → staff training plus sender authentication → fewer credentials captured'). For ethics scenarios, name the principle at stake (privacy, accuracy, ownership, accessibility, consent), identify the stakeholders, and state the action plus its justification. Learn two or three named laws or codes relevant to your jurisdiction and cite them by function.
Candidates calculate file size or data capacity by multiplying the wrong units, e.g. treating 1 KB as 1000 bytes when the question uses binary prefixes, or forgetting to multiply by colour depth and frame rate for an image or video.
The decimal and binary prefix conventions are both in circulation, and candidates apply a memorised formula without checking which convention the question states. Multi-factor formulas (resolution × colour depth × frames) are partially applied under time pressure.
What to do instead
Write the formula in full before substituting: image bits = width × height × colour depth; video bits = that × frame rate × seconds; sound bits = sample rate × bit depth × channels × seconds. Convert to bytes, KB, MB only at the end, and state explicitly whether you used 1000 or 1024, following the question's wording. Sanity-check the magnitude (a small icon should not be megabytes).
When asked to select and justify an application package feature or a spreadsheet formula, candidates describe the menu path or give a generic formula (e.g. 'use SUM') instead of the exact function, cell references, or feature that solves the stated problem.
Practical software skills are learned by clicking, so the knowledge is procedural and hard to verbalise in a written exam. Candidates also confuse relative and absolute references when a formula must be copied down a column.
What to do instead
Practise writing formulas in text form with exact syntax and references, including the $ signs, and state what happens when the formula is filled down. For each major feature (mail merge, conditional formatting, pivot table, lookup, goal seek) write one line on the problem it solves and the data it needs. Read the question's data layout and name the actual cells rather than describing them.
On emerging-technology and social-impact questions, candidates give unbalanced answers that list only benefits (or only risks) and make unsupported claims about AI, cloud computing or automation without linking them to a named example or stakeholder.
These topics attract opinion and headlines rather than textbook facts, and candidates reproduce enthusiasm or alarm instead of analysis. The command word ('discuss', 'evaluate') is also overlooked, so a one-sided list is produced where a judgement is required.
What to do instead
For any 'discuss' or 'evaluate' question, write a benefit and a corresponding drawback for the same stakeholder group (users, workers, organisations, society), then add an explicit judgement sentence. Anchor each claim to a concrete mechanism (e.g. 'training data bias → skewed decisions for under-represented groups'). Check the command word before writing and match the structure to it.