MATSEC IM Information Technology common mistakes
MATSEC IM Information Technology · Understand the traps and how to avoid them.
When asked to design a database table or show a schema, the candidate lists field names as prose (e.g. 'Student: name, ID, course') without stating data types, primary key, or foreign keys, so the answer cannot be marked as a design.
The candidate confuses a descriptive list with a formal schema and has not practised converting a scenario into tables with keys and types under exam timing.
What to do instead
Adopt a fixed table-drawing ritual: write each field on its own line with name, data type, size where relevant, then underline the primary key and draw an arrow for each foreign key. Before writing, circle every noun in the scenario as a candidate entity and every relationship verb. Practise three past-paper scenarios this way until the format is automatic.
The candidate answers a 'describe two advantages of a relational database over a flat file' question with vague benefits such as 'it is better organised' or 'it is faster', with no mechanism or context.
Advantages are memorised as slogans rather than understood as consequences of normalisation, constraints, and separation of data from programs, so the candidate cannot unpack them into mark-earning points.
What to do instead
For each advantage, learn a two-part template: the mechanism plus the consequence. For example: 'Data is stored once in related tables instead of duplicated in every file, so an update need only be made in one place, removing the risk of inconsistent copies.' Build a personal list of five such mechanism-plus-consequence sentences and test them against past questions.
In a question about the fetch-execute cycle or CPU operation, the candidate writes that the 'CPU does the processing' or lists components (ALU, control unit, registers) without describing what each does or the order of events.
Component names are learned as a vocabulary list, and the candidate has never rehearsed the cycle as a sequence of transfers between named registers.
What to do instead
Write the cycle as a numbered sequence naming registers explicitly: program counter holds the address, the address is copied to the memory address register, the instruction is fetched into the memory data register and then the instruction register, the control unit decodes it, the ALU executes it, and the program counter is incremented. Draw it from memory daily for a week, then answer past questions using only the numbered sequence.
When converting between number bases or performing two's complement, the candidate shows only the final answer and loses method marks, or makes a single place-value slip that ruins the whole result.
The candidate does the working mentally to save time and has not built the habit of column-by-column place-value working with a check.
What to do instead
Always show place-value columns (128, 64, 32, 16, 8, 4, 2, 1) and convert back to the original base as a self-check before moving on. For two's complement, invert all bits and add one, then verify by adding the result to the original to get zero. Do ten conversions per day with visible working until the check becomes reflexive.
The candidate describes a logic circuit or Boolean expression only in words ('it turns on when either switch is on but not both') and never writes the truth table, Boolean expression, or gate diagram the question asks for.
The candidate understands the behaviour but cannot map everyday language onto AND, OR, NOT, XOR, NAND, and NOR, and so avoids the formal representation.
What to do instead
For every verbal description, first build a truth table with one column per input and one for the output, then read the expression off the rows where the output is 1 as a sum of products, then simplify using Boolean identities. Practise translating five verbal scenarios per session into truth table, expression, and gate diagram, and check the diagram against the table.
In a networking question, the candidate gives a definition of a LAN or WAN but cannot apply it, for example failing to say why a school with two buildings should use a particular topology or medium, or omitting the device (switch, router, access point) that makes the design work.
Networking is revised as definitions and acronym expansions rather than as design decisions with trade-offs, so the candidate has no framework for choosing between options.
What to do instead
Build a comparison grid for topologies (star, bus, ring, mesh) and media (twisted pair, fibre, wireless) with columns for cost, speed, fault tolerance, and ease of expansion. For each past-paper scenario, write a recommendation sentence of the form 'choose X because Y, accepting the drawback Z'. Memorise which device operates at which layer and what it forwards.
The candidate writes an algorithm or pseudocode that is logically close but uses undefined variables, mixes pseudocode with real language syntax, or omits initialisation and loop termination, so it cannot be traced or marked.
The candidate writes code as if to a compiler rather than to a human marker, and has not been taught a consistent pseudocode convention with declared variables.
What to do instead
Adopt one pseudocode style and stick to it: declare and initialise every variable at the top, use explicit START and END, indent consistently, and state loop conditions that provably terminate. After writing, trace the algorithm by hand with a small sample input, recording each variable's value line by line, and fix anything that does not match the expected output.
When asked to test or debug a program, the candidate says 'check for errors' or supplies only normal-case test data, missing boundary values, invalid input, and expected outputs for each test.
Testing is treated as an afterthought rather than a planned activity with categories of test data, so the candidate cannot produce the table a marker expects.
What to do instead
Learn the three categories explicitly: normal, boundary (including just inside and just outside each limit), and erroneous or extreme data. Present tests as a table with columns for test data, type, and expected result. For any program in the question, produce at least one test of each type and state what a failure would reveal about the logic.
In a spreadsheet or database practical question, the candidate writes a formula that works for one row only (relative references where absolute are needed) or describes a query in English instead of writing the criterion, sort, or function required.
The candidate has not distinguished relative from absolute referencing or practised the syntax of lookup, conditional, and aggregate functions, and confuses describing a query with specifying one.
What to do instead
Memorise the rule: references that must stay fixed when copied need dollar signs on both column and row. Practise a fixed set of functions (IF with nested conditions, VLOOKUP or INDEX/MATCH, SUMIF, COUNTIF, and simple queries with criteria and sorting) on a small dataset, and always state the exact cell or field names the formula or query will use.
On a social, legal, or ethical question (data protection, copyright, computer misuse, health and safety, environmental impact), the candidate gives a one-sided opinion such as 'it is illegal' or 'it is bad for privacy' without naming the principle, the obligation on the organisation, or the counter-argument.
These topics are revised as general knowledge rather than as structured arguments, so the candidate cannot produce the balanced, principle-based answer that earns the higher marks.
What to do instead
For each topic, prepare a three-part answer: the principle or law and what it requires, the effect on the organisation or individual (including the penalty or risk), and one counterbalancing benefit or limitation. Learn a small set of named principles (for example lawful and fair processing, purpose limitation, data minimisation, accuracy, storage limitation, and security) and apply them to the scenario rather than reciting them.