The SEC 2027 G2 Computing K237 Paper 2 is a computer-based laboratory examination lasting two hours fifteen minutes, worth 80 marks and 50% of the subject. It has three compulsory task areas: Media Software (approximately 30 marks), Spreadsheets (approximately 25 marks) and Programming (approximately 25 marks). SEAB explicitly states that candidates must submit softcopies and that the allotted time includes saving the required work on their computer. A spreadsheet calculation or Scratch game that works briefly on screen but is not saved correctly is not the finished examination deliverable.
This focused guide answers what students and parents need to know before the first SEC K237 lab examination: how to prepare a 135-minute workflow, use authorised software, prevent Excel reference errors, test conditional formulas, debug Scratch logic, create a purposeful media output and verify every saved file. It follows the official SEAB 2027 K237 Computing syllabus, uses original small practice projects and avoids pretending that a commercial tutorial or online screenshot reveals confidential examination tasks. The actual lab environment and exact saving instructions are controlled by the examination centre and school.
- Subject: SEC G2 Computing K237, a new 2027 subject without an older directly equivalent national code.
- Paper 2: 135 minutes, 80 marks, 50% of final result; three compulsory tasks.
- Media Software: approximately 30 marks for visual, presentation and/or video software skills.
- Spreadsheets: approximately 25 marks for data, formulas, charting and related operations.
- Programming: approximately 25 marks for block-based logic, including Scratch concepts.
- File submission: save and submit the required electronic work in the approved examination environment; saving time is inside 135 minutes.
- Boundary: the microcontroller/MakeCode-related unit is assessed in Paper 1 only.
- First action: complete one small task in each category, save it to a designated practice folder and reopen the results to check correctness.
Why K237 is new, and why old exam tips can mislead
The Singapore-Cambridge Secondary Education Certificate starts in 2027 and assesses subjects at their registered G1, G2 or G3 levels under Full Subject-Based Banding. SEAB labels K237 Computing as a new G2 syllabus with no corresponding earlier national code. A website offering ‘ten years of exact K237 SEC past-year papers’ cannot be describing ten genuine past K237 examinations in this new format.
There are helpful educational resources from older computing courses, but they should not replace the current published specification or its official specimen papers. In particular, K237 is not G3 Computing K349. The G3 route may use different programming environments and assessment expectations. A K237 student should practise the software and task skills actually prescribed for the G2 examination.
The K237 qualification has two equal-weight papers. Paper 1 is a 90-minute electronic theory and applications examination worth 70 marks; it covers all six modules. Paper 2 is the lab-based practical worth 80 marks and covers Modules 4–6: Spreadsheets, Media Software and Programming. This article owns Paper 2 study and links back to the full subject guide for Paper 1.
Paper 2 exact structure: what is compulsory, what is approximate
| Task area | Published mark allocation | Main deliverable behaviour |
|---|---|---|
| Media Software | Approximately 30 of 80 marks | Create or edit requested visual, slide, graphic or video outputs |
| Spreadsheets | Approximately 25 of 80 marks | Manipulate data; compute, filter, chart and format according to specifications |
| Programming | Approximately 25 of 80 marks | Implement, test and debug the required block-based behaviour |
| Across tasks | 80 marks total; 135-minute sitting | Save the required softcopy files following examination instructions |
SEAB’s media, spreadsheet and programming split is approximate. The real tasks and subparts determine where marks are assigned. Students should not attempt to predict which particular software feature carries the most marks based on an online ranking. Prepare the underlying operations and practise producing complete, verifiable outputs.
All Paper 2 tasks are compulsory. A student who is strong at Scratch but refuses to practise spreadsheet formulas risks leaving a substantial part unattempted. Likewise, someone skilled at graphics but unable to trace variables and loops may lose core programming marks. Balanced preparation should reflect the actual weak area without abandoning the others.
The exam is a product-delivery exercise, not merely a demonstration of clicks
In Paper 2, candidates produce computational artefacts. An artefact is more than a visible application window: it is the requested file or set of files showing a correctly implemented result. This is why the syllabus mentions softcopy submission and saving within the allotted time. The behaviour must exist in the saved work, not only in the student’s memory of what they clicked.
An effective workflow starts by reading the actual requirements. What should be created? Which input data are supplied? What filenames and formats are needed? How will an assessor see the result? What changes should the artefact make when the input changes? When these questions are answered, software operations become purposeful rather than a hurried sequence of menu clicks.
The first fifteen minutes: build a task map before editing
The following is a practice routine, not a prescribed national minute allocation. At the start of a mock lab, read all three tasks and create a brief checklist. Identify which deliverables are straightforward, which include several dependent stages and which may require saving in a particular form. Confirm the location in which authorised examination work should be stored.
| Preparation question | Why it matters | Common preventable mistake |
|---|---|---|
| Which files must be supplied? | Marks require the requested deliverables | Submitting only a preview or one of several files |
| What is the required naming scheme? | Files may be identified by exact names | Saving an ambiguous ‘final2’ instead of the specified file |
| Which output format is required? | Editable source and exported media differ | Saving the wrong extension or an unopenable file |
| Which inputs must remain unchanged? | Given data and templates may be assessed | Replacing source data while experimenting |
| Which tasks depend on other tasks? | Order affects time and correctness | Creating a chart before fixing underlying data |
| Which checks fit the remaining time? | Testing catches hidden errors | Making cosmetic changes until the final seconds |
Do not write unauthorised notes in a real examination. Use the task paper or the permitted working arrangements as instructed. At home, a checklist can train the mental sequence, but the official centre’s rules for scratch work, file paths and devices must always prevail.
Saving is part of the assessed workflow
SEAB explicitly includes saving time within the two-hour-fifteen-minute duration. Therefore a student should get into the habit of saving early, after major milestones and again before finishing. File operations must follow the specific examination instructions; personal cloud services, USB drives and external accounts should not be assumed permissible.
A robust practice folder might contain a media working file, a requested export, a spreadsheet workbook and a Scratch project, each named consistently. After saving, close or reopen copies when permitted to verify that the work survives. Reopening does not mean launching a remote application or uploading to an unauthorised service in an actual exam; follow the local approved environment.
A five-step saved-file verification routine
- Confirm the required output name and its full location before beginning.
- Save an initial version early and save again after significant changes.
- Check the extension and content against the task requirements.
- When permitted, reopen or inspect the saved output to confirm it contains the finished work.
- Verify that every requested file exists before using remaining time for optional refinements.
A file that looks complete on screen may still contain an unsaved change. A workbook can display a calculated value while the saved copy contains an earlier formula; a video editing workspace can look polished while the required rendered video file has not been exported. These are different failure modes requiring separate checks.
Original three-task practice set: a community reading fair
Imagine a fictional school club wants to promote a reading fair. Your practice requires a one-page visual promotion, a spreadsheet of booth materials and a Scratch quiz that asks visitors about books. This three-task project is deliberately original, not an SEC specimen. Its purpose is to integrate the three examinable skills while providing visible evidence that each output matches a clear specification.
For media, the product must show the event title, date, venue and a call to action in a readable layout. For spreadsheets, the costs should update when quantities change and a chart should summarise categories. For programming, the quiz should initialise its score, respond to correct and incorrect input and display a consistent result. A full practice includes saving each output correctly.
Task 1 Media Software: communication before effects
A media output should make its message understandable to the intended audience. Before choosing colours, transitions or filters, decide what a viewer needs to know. In the fictional reading-fair poster, the event name, relevant date, location and action should be legible. A student who adds five decorative shapes but hides the venue has not fulfilled the communication task.
The syllabus includes vector and raster graphics, presentations, video editing and related functionality. Their tools are not interchangeable. Vector paths and nodes are useful for editable shapes; raster operations work with pixels. A video editing task also needs time-based sequence, transitions and export considerations.
Vector versus raster: avoid using the wrong editing model
| Property | Vector | Raster |
|---|---|---|
| Underlying data | Shapes, paths and nodes | Pixel grid |
| Useful example | Logo, icon or scalable diagram | Photograph or textured image |
| Typical edit | Modify a node, shape, fill or group | Crop, resize or adjust pixel appearance |
| Common error | Ungrouped object accidentally moved | Image enlarged far beyond available resolution |
| Assessment consequence | Requested structure or visual consistency missing | Poor clarity, crop or export |
For a vector logo made from two overlapping shapes, the student may need to group objects after arranging them. If an examination task requires individual edits later, grouping too early can make changes awkward. In a raster photo, cropping can remove unwanted space, but it may also cut off the meaningful subject. The correct operation depends on the instruction and message.
Original vector task: a simple reading-fair icon
Create an editable symbol consisting of an open-book shape and a smaller bookmark shape in a permitted practice application. Align the shapes, select legible contrast and group them only when the individual edits are finished. Resize the icon to test whether its proportions remain coherent. If a required part disappears behind another object, inspect stacking order rather than redrawing everything.
The point is the sequence of operations: create, edit, arrange, verify, save. Merely clicking a familiar icon does not demonstrate why the structure is appropriate. Students should be able to explain what grouping changes and why a vector output may suit a logo better than an enlarged small photo.
Raster editing: crop, dimensions and truthful results
The syllabus includes raster editing features such as brightness, contrast, cropping and resizing. These affect the image’s content and presentation. An image can become brighter yet less readable if highlights are lost; cropping can improve focus or remove essential context. The exam assesses execution of requested operations, not an unrestricted makeover.
A practice task might ask you to crop an event photograph to a specified aspect ratio and improve legibility for a poster. First retain the main subject, then adjust the crop and verify the final dimensions. If the crop cuts off part of the event sign, reconsider. Editing should support the task, not simply create a dramatic appearance.
Presentations: a storyboard is an efficiency tool
K237 covers designing presentations and multimedia narratives, including slide masters and visual hierarchy. Students should be able to organise information logically across slides, maintain consistent typography and use effects purposefully. A slide master can establish recurring layout and style, reducing the need to manually correct every slide.
Original three-slide announcement
A fictional assignment asks for three slides: first the reading fair’s purpose, second the activity highlights, third the registration instructions. A coherent solution keeps the main message in a consistent place and avoids competing text blocks. If the final slide gives the wrong day or omits how to register, decorative quality will not repair task incompleteness.
Check whether a master element belongs on every slide and whether any page needs a different layout. A blanket style change may unexpectedly reduce contrast or reposition images. Verify the visible result, then save the requested editable or exported file as instructed.
Video editing: the timeline and export are separate stages
Video editing combines clips, still images, text, timing and possibly audio. A storyboard helps establish sequence before assembly. Students should know that a project file can preserve editing decisions, while an exported media file is a rendered output. An examination may require one, the other or both; the actual task wording controls submission.
For a short announcement, a useful sequence could show the event identity, explain who it is for, provide essential details and end with a call to action. Distracting transitions are not automatically better. If audio is included, clarity and suitable levels matter, but the exam’s allowed resources and copyright instructions must be respected.
Video verification checklist
- Does the clip have the intended sequence from start to finish?
- Are important words on screen long enough to read?
- Are transitions coherent rather than distracting?
- Does any audio obscure speech or important information?
- Has the correct requested frame size or output format been used?
- Does the saved/exported file actually open and play where permitted?
A student should not use copyrighted music or media without appropriate rights merely to make a practice export exciting. For national assessed work, the provided resources and copyright requirements are definitive. Outside exams, learning to attribute and license media properly is part of responsible computing.
Task 2 Spreadsheets: formulas must recompute when data change
A spreadsheet is not simply a grid containing neat numbers. A working computational spreadsheet applies formulas to source data and changes results when the data change. If the exam asks for a calculated subtotal, typing a correct amount manually is not equivalent to entering the requested formula. The method is part of the deliverable.
Original data table for a reading fair
| Cell/column | Content | Purpose |
|---|---|---|
| A Item | Bookmarks | Name of purchased item |
| B Quantity | 25 | Units ordered |
| C Unit Price | $0.80 | Cost for one item |
| D Subtotal | =B2*C2 | Calculated amount |
| E Category | Stationery | Used for categorisation |
In this example, D2 should calculate $20.00. If B2 changes from 25 to 30, the subtotal should become $24.00 without manual correction. This is the simplest test of whether the cell contains an appropriate formula. A student who cannot predict the output after changing the quantity may be using software mechanically rather than reasoning.
The same principle applies to charts and conditional formatting. The result should be tied to relevant data and update appropriately. Test with different values before final submission; a spreadsheet that appears correct on one provided sample can contain a formula error hidden until the input changes.
Relative and absolute references: one common mistake, many lost marks
A relative reference such as B2 changes when a formula is copied to another row. An absolute reference such as $H$1 remains anchored. The correct choice depends on which values vary by row and which are shared. Students often know how to type dollar signs but fail to explain why they belong in one location.
Worked Excel practice: a shared discount
Suppose H1 stores the discount rate 0.10, and D2 contains an item’s original price. An appropriate discounted-price formula in E2 is =D2*(1-$H$1). When copied to E3, the row-specific price reference becomes D3, while $H$1 remains fixed. If H1 changes to 0.15, all correctly linked discounted prices should update.
If the rate is typed separately into every row, the worksheet may become inconsistent when the rate changes. If the H1 reference is left relative, copying can point to H2, H3 and so on. The repair is to describe the business rule before writing the formula: each row has a different price but shares one discount rate.
IF statements: test the boundary, not only an easy case
The syllabus includes IF and other logical functions. They turn rules stated in ordinary language into conditions. A formula can work for values clearly above and below a threshold yet fail exactly at the boundary if the comparison symbol is wrong. Boundary testing is therefore a core practical habit.
Worked Excel practice: attendance eligibility
A fictional club awards a certificate for at least eight attendances. If B2 contains attendance, an appropriate Excel formula is =IF(B2>=8,”Eligible”,”Not yet”). Test 7, 8 and 9. The first should be ‘Not yet’ and the other two ‘Eligible’. If the formula uses > rather than >=, exactly eight will be wrongly rejected.
Now add a second condition: membership dues are paid. A logical AND can combine the requirements if the task demands both. But do not invent a rule not present in the actual exam; the worksheet should implement only the conditions provided. The student must also know how the payment status is represented in the data.
COUNTIF and SUMIF: why the criterion must match the data
Conditional functions are useful when only some records count toward a total. A club may want the number of booths labelled ‘Books’ or the sum of expenses in the ‘Printing’ category. The formula must refer to the correct range and criterion. If data use ‘Books’ but the criterion is ‘Book’, an exact matching rule may produce an unexpected result.
Original conditional count example
Suppose A2:A6 contains five booth categories: Books, Games, Books, Food, Books. The formula =COUNTIF(A2:A6,”Books”) returns 3. If the cell range ends at A5, it may return 2, demonstrating a range-selection error rather than misunderstanding of the category. Before finalising, identify where the dataset starts and ends.
For SUMIF, confirm which column contains the criteria and which contains the values to add. The syntax can appear memorised while the actual chosen ranges are wrong. A manual sum of a tiny test subset can validate the spreadsheet during practice.
LOOKUP functions: plausible wrong answers are dangerous
K237 lists lookup functions including VLOOKUP, HLOOKUP, INDEX and MATCH. A lookup retrieves data based on a specified key. It is easy to obtain a plausible-looking result from the wrong column or an unintended approximate match. A test should include a known key and a key not found in the data.
Worked exact-match formula
Imagine product codes in A2:A6 and their prices in B2:B6. The target code is entered in D2. An Excel formula =VLOOKUP(D2,$A$2:$B$6,2,FALSE) requests an exact match in the first column and returns the second column. If the table moves, the range and column index must be reconsidered. Copying the formula without checking the new data layout is unreliable.
The student should understand why FALSE matters for exact matching and why the desired lookup value must be in the table’s appropriate position for VLOOKUP. A different table layout may be better solved with another prescribed function. It is the problem that determines the tool.
Charts: a correct image must represent correct data
The Computing syllabus includes bar, column, line, pie and combination charts. These are not interchangeable decorations. Use a line chart for an appropriate sequence over time, a bar chart for comparison across categories and a pie chart when proportions of a meaningful whole are to be displayed. Labels, units and correct series selection matter more than a colourful palette.
An original chart sanity check
Suppose a worksheet tracks fair attendance over four hours: 12, 18, 25 and 20 visitors. A line chart may show changes over time. If the data range accidentally includes the item-price column instead of visitor counts, the chart may still look smooth while communicating the wrong quantity. The correction is to trace every plotted series back to its original cells and axis labels.
When asked to make a combination chart, confirm which measure belongs to the column series and which to the line series according to the task. Do not place two unrelated quantities on one axis without considering units and interpretation.
Sorting, filtering and validation: keep each row intact
A common spreadsheet disaster occurs when a student sorts only one column and detaches prices from product names. Sort the entire intended dataset so each record remains together, according to the application’s proper procedure. Filtering should restrict the visible set without deleting other records unless deletion is explicitly required.
Data validation can reject entries outside specified conditions. If a fictional worksheet allows attendance from zero to twenty, test invalid values such as −1 and 21 as well as valid boundaries 0 and 20. A custom error message may improve usability, but it does not replace correctly configured validation.
Task 3 Programming: understand the game’s state
The practical programming component is based on the prescribed block-based concepts including Scratch. A student can assemble blocks that appear plausible while the game behaves incorrectly because variables are not initialised, events repeat or a loop never reaches its intended condition. The best debugging skill is tracing what happens at each step.
Before moving blocks, write the expected sequence: start the game, set score to zero, position a sprite, respond to an input, update the score once per valid event, and finish or restart according to the rule. A program is correct only when its actual behaviour matches the task under several tests.
Variables: what is the initial state?
A variable stores a value used by the program, such as score, level or lives remaining. At the start of a new run, initialise it as required. If a previous run leaves a score at 12 and the next run begins with that score rather than zero, the program’s state has not been reset correctly. This can produce a visually impressive game with flawed assessment logic.
Original Scratch game: catch a falling book
Imagine a sprite representing a book falling from the top of the stage. A player sprite catches it to earn a point. The game should set score to 0 when started, position the objects, detect a valid catch, increase score by 1 and reset the falling object. Each of these is a separate logical responsibility that can be tested.
A useful first test is one catch. Then test repeated catches, missing the book, restarting and contact that lasts for several animation steps. If the score increments ten times during one catch, the repeated event logic needs repair. Do not simply reduce the final displayed score; fix the underlying condition.
Conditions and loops: the source of many hidden bugs
A condition tests whether something is true, while a loop repeats instructions according to its rule. A forever loop continually checks or updates state; a fixed repeat loop runs a specified number of times; a repeat-until loop depends on a stopping condition. Students should predict execution, not just recognise the block colours.
Worked debugging thought experiment
A game says, ‘Increase score when touching the player.’ The code checks touching continuously and increases score every time the condition is true. A single long collision causes multiple increments. The conceptual repair is to register a catch once and ensure the condition becomes false or the object is reset before another increment. An appropriate wait or state mechanism can achieve this depending on the project.
Test the repaired code under the same long-contact scenario and again after restarting. The result should be one point for one valid catch according to the fictional rule. This is a general debugging pattern: identify the difference between one event and repeated detection of a continuing state.
Broadcasting messages: coordinate without guessing order
Scratch-style messages can coordinate behaviour among sprites. If a player starts an action and another sprite must react, a broadcast can help separate responsibilities. But a student should still know when the message is sent, which receiver acts and whether repeated broadcasts cause duplicate behaviour.
In a simple original quiz, a broadcast might signal that a new round begins. Receivers reset their positions or display state. If multiple messages unintentionally trigger the same score update, the game may behave incorrectly. Trace events across sprites rather than debugging only the block currently visible.
Lists: index errors in small tasks
The syllabus covers list or array operations such as accessing items by index, adding or removing items and checking whether an item is present. A student should know how a list changes after deletion and whether an index still points to the intended entry. A complex game is not needed to teach this concept.
Original list exercise
Suppose a list contains [Book A, Book B, Book C]. Removing the second entry leaves [Book A, Book C]. If the program next requests item number three, that may no longer exist in the two-item list. A valid program needs to account for the new length or select the intended item differently. The learning is about state change and valid positions.
Test the empty-list case and the one-item case as well as the typical three-item example. Boundary behaviour often exposes a bug that a normal run hides.
Custom blocks: reducing repetition without hiding logic
A custom block can group repeated instructions into a named operation. For example, a quiz may have a ‘Start New Round’ routine that resets certain variables and positions. This can make the main sequence easier to understand when used appropriately. However, the name does not guarantee the instructions inside it are correct.
When testing, call the block under known conditions and inspect all intended changes. If a routine resets score when it should preserve score between rounds, the block is logically flawed even though it executes without a visible error.
Paper 1-only content: microcontrollers are not a Paper 2 build requirement
The K237 syllabus includes MakeCode and microcontroller concepts, but expressly limits that related subtopic to Paper 1 assessment. Students can and should understand the prescribed concepts for the electronic theory paper. They should not devote the whole 135-minute lab rehearsal to physical microcontroller construction while neglecting media, spreadsheet and Scratch practical skills.
This distinction is especially important for parents choosing enrichment resources. A robotics activity may develop computational thinking, but that does not make it the correct direct practice for K237 Paper 2. Align resources with the actual examination scheme rather than the broad marketing word ‘Computing’.
The three-stage test: unit tests, integration and saved-output checks
| Testing level | Example | What failure reveals |
|---|---|---|
| Unit | Change one spreadsheet quantity and verify subtotal | Formula or reference error |
| Unit | Trigger one catch in Scratch | Event or state error |
| Unit | Reopen one image export | Media format or file-save issue |
| Integration | Update data then inspect linked chart | Incorrect series or range |
| Integration | Restart game after several events | Variables not reset |
| Final delivery | Inspect all saved required files | Missing, wrong-named or incomplete output |
A program or workbook can pass a simple test and fail a second condition. Testing at several levels is not wasted time; it is part of completing the task. Do not only test the easiest scenario because it produces a gratifying result. Try boundary values and altered input to discover hidden assumptions.
A realistic 135-minute lab rehearsal strategy
No universal minute split is prescribed beyond the total duration, and the specific task complexity can vary. A practice allocation might reserve a short initial reading period, protect time for all three compulsory areas and leave a final verification window. The student should adapt based on their timed diagnostic evidence. A fast programmer may need more spreadsheet testing; a capable media editor may need to protect programming time.
| Rehearsal segment | Purpose | Common risk |
|---|---|---|
| Initial task review | Identify output requirements, folder and dependencies | Starting attractive work before reading constraints |
| Media production | Create and export the required visual result | Polishing effects but missing the requested format |
| Spreadsheet production | Implement accurate formulas, charts and validation | Hardcoded values or wrong range |
| Program development | Build, run, trace and debug the required logic | Code looks right but fails after restart |
| Final verification | Check files, names, locations and current content | Missing or unsaved deliverable |
During practice, record actual time and identify where the learner stalls. The next lesson should target that point. If the learner repeatedly runs out of time because they are searching menus, short software fluency exercises may help. If they spend the time debugging incorrect logic, teach computational reasoning. If they forget to save, make output verification an explicit routine.
A paper-to-digital task translation exercise
Before opening a spreadsheet or Scratch program, rewrite the task in three short statements: input, process and required output. For the attendance example, inputs are attendance counts and the eligibility threshold; the process is an IF rule; the output is a correctly labelled status for each record. For the falling-book game, the input is a catch event, the process updates score once, and the output shows the new score.
This simple translation mirrors the way computing turns information into behaviour. It can prevent mistakes in all three practical areas. A media task also has inputs—images, text and a target audience—and produces a deliberately structured visual output. Decomposition before clicking can improve both clarity and speed.
An error ledger that improves the next lab rather than adding homework
| Error family | Evidence | Next practice |
|---|---|---|
| Requirements | Wrong file type despite good design | Write and verify the deliverable checklist |
| Software operation | Cannot locate/group the correct objects | Short targeted feature practice |
| Data | Chart uses the wrong column | Trace each plotted series to source |
| Formula | Relative reference moves unexpectedly | Explain what must be fixed/relative |
| Logic | Boundary value classified incorrectly | Test 7, 8 and 9 around threshold |
| Programming | One event triggers repeated scoring | Trace loop and reset state |
| File handling | Saved older draft instead of final | Save/reopen after a changed task |
| Time management | Third task is left blank | Practice all three under timed conditions |
Every entry should include what went wrong, its first cause and a new test. ‘Careless error’ is too vague. ‘Copied a discount formula without anchoring H1’ is actionable. The latter can be repaired by understanding shared versus row-specific values and tested on a fresh worksheet.
Seven-day emergency preparation for the K237 lab
- Day 1: verify the K237 code, 135-minute format and the software available under school instructions.
- Day 2: practise Excel formulas, relative/absolute references and boundary testing.
- Day 3: practise conditional functions, lookups, sorting, validation and charts.
- Day 4: create and save a simple vector/raster output and a presentation task.
- Day 5: practise a short storyboard, media edit and correct export.
- Day 6: build and debug an original Scratch project with several test conditions.
- Day 7: complete a timed three-task rehearsal and verify every saved file.
A seven-day plan can reduce avoidable errors but cannot substitute for sustained computing education. If the student lacks basic keyboard, file or software fluency, begin with smaller tasks. If repeated anxiety or access issues interfere, involve the school early. Do not train using unauthorised copies of confidential examination files.
A six-week full preparation plan
| Week | Core practical focus | Independent proof |
|---|---|---|
| 1 | Basic media editing and deliverable reading | Correct saved visual output from a fresh brief |
| 2 | Excel references, formulas and number formats | Worksheet recalculates after changed data |
| 3 | IF, lookup, validation, filters and charting | Boundary cases and labels are correct |
| 4 | Scratch loops, events, variables and messages | Program behaves correctly across multiple tests |
| 5 | Integrated media, spreadsheet and programming mock | Three complete outputs in one session |
| 6 | Time management, unfamiliar tasks and file verification | Fewer repeat errors and correct saved deliverables |
The schedule is a guide to structuring practice, not a guarantee of a result. A student with strong spreadsheet skills but weak media editing may need to shift time accordingly. Measure completion and correctness on unseen tasks, not hours of screen time or the complexity of the game design.
What parents should ask after a student says ‘It works’
Ask to see the saved output, not only the open application. Does the spreadsheet recalculate when the input changes? Are the chart axes and units meaningful? Can the Scratch game restart without carrying over a score? Does the exported media file actually open? Can the student explain which requirement each output satisfies?
These questions do not require parents to be programmers. They test whether the candidate can demonstrate reliable results. A student who can confidently explain a simple project is often showing more transferable competence than one who has copied a sophisticated online tutorial without understanding the behaviour.
Ethics, privacy and exam conditions
The SEC examination assesses the candidate’s own practical skills under controlled conditions. Do not bring unauthorised storage devices, saved projects, hidden cloud links, AI helpers or external communications into the exam. School and SEAB instructions govern permitted resources, files, software and access arrangements. Using a personal preferred tool during learning does not authorise its use during the national examination.
Outside exam conditions, general resources can help students learn concepts, but copied work should not be represented as independent performance. Respect media licences and personal data. A good computational solution is not only correct; it is also produced and shared responsibly.
How effective tutoring repairs different K237 problems
A student who knows spreadsheet functions but chooses the wrong range needs data and reference reasoning. One who builds Scratch quickly but cannot debug needs state tracing. One who can edit video but forgets the output needs a file-delivery routine. Applying identical instruction to all three can waste time. Diagnose the first failure and teach a smaller test before returning to the complete task.
At eduKateSG, the three-student tutorial philosophy is diagnosis, first-principles explanation, guided practice and independent changed-case retesting, as set out in the immutable Clementi teaching reference. Small groups can provide closer observation when properly run, but class size alone does not guarantee a grade or replace trained Computing instruction.
Frequently asked questions about the SEC 2027 G2 Computing lab
What is the SEC 2027 code for G2 Computing?
K237, newly introduced for 2027 with no corresponding old national subject code.
How long is Paper 2?
Two hours fifteen minutes, including saving the work as required.
How much does Paper 2 count?
Fifty percent of the subject. Paper 1 also contributes fifty percent.
How many marks are in the lab?
Eighty marks, allocated approximately 30 media, 25 spreadsheet and 25 programming.
Are any of the lab tasks optional?
No. SEAB states that all tasks are compulsory.
Does the exam test Scratch?
The G2 Programming syllabus includes Scratch-style block programming and game-related skills.
Does G2 Paper 2 require Python and JupyterLab?
Do not substitute G3 Computing’s different environment. K237 has its own prescribed block-based skills and authorised software arrangements.
Are microcontrollers part of the laboratory paper?
The K237 microcontroller subtopic is assessed in Paper 1 only.
What is the most common Excel error?
There is no official ranking. Typical practice issues include wrong cell references, ranges, comparison operators and hardcoded results.
Should I use an absolute reference for every cell?
No. Fix a reference only when the corresponding value should remain anchored as the formula is copied.
Why test the value equal to a threshold?
Because a formula using greater-than instead of greater-than-or-equal can fail exactly at the boundary.
Can a worksheet appear correct but be wrong?
Yes. It may contain hardcoded values, wrong references, inappropriate chart data or an unsaved latest version.
What if the Scratch game scores several points per collision?
Trace whether repeated contact is counted every frame; fix the event state or reset logic and test again.
Does SEAB allow personal cloud drives for submission?
Do not assume so. Follow the official examination folder and saving instructions, not a personal workflow.
Are media exports necessary?
Save exactly the file formats requested in the actual question, which may include editable sources and/or output files.
Can students use unofficial K237 ‘ten-year series’ books?
K237 is new for 2027. Use current SEAB syllabus and released specimen resources, and check other materials carefully.
How do I know practical preparation is improving?
The student independently completes new tasks, verifies behaviour and saves the required outputs correctly within a realistic time.
Official K237 sources and related eduKate guides
- SEAB official 2027 K237 G2 Computing syllabus, Paper 2 and software skill lists
- SEAB K237 subject directory and official specimen papers
- Mastering G2 Computing K237: all six modules and both papers
- How G2 Computing Paper 2 Lab Practical Works
- 2027 K237 new syllabus: spreadsheet, media and Scratch specification
- How to Prepare for K237 SEC Examinations
- K127 versus K237 versus K349: computing subject choices
Fact-checked 11 November 2026 against the published SEAB 2027 K237 syllabus. Every worksheet, project, score counter, practice timeline and file requirement illustration is original and hypothetical, not a disclosure of actual national paper tasks. Follow the exam centre’s official permitted software and submission instructions.