Paper 4 — Revision Sweep

Marking Scheme

Section — Multiple Choice — Remember

Q1. (c) Private information belongs with the grown-ups who look after you. (d) is the tempting one, because a class group feels safe and familiar, but anything posted there can be screenshotted and passed on to strangers. [1]

Q2. (c) Editor means change my work. (b) Commenter is the tempting one, because a commenter does add something, but only feedback beside your words, never inside them. [1]

Q3. (b) Restricted keeps out everyone you did not name. (a) is the setting that caused the moustaches, because the link keeps working for everyone it reaches. [1]

Q4. (d) Nobody real ever needs your password, not for coins and not for prizes. (b) is the clever-sounding trap: by the time you change it, whoever took the old one has already used it. [1]

Q5. (a) A footprint is the trail you leave behind you, and every step of it carries a date. (c) is the near-miss: an account is one place the trail collects, but the trail also holds searches, login times and other people's copies. [1]

Q6. (b) Saving proof and showing a trusted adult is what got the picture taken down. (a) is what somebody in the chat was already suggesting, and passing it on only makes the crowd bigger. [1]

Q7. (d) A credit names whoever made it. (b) is the exact mistake in the story: putting your own name on somebody else's work makes it worse, not better. [1]

Q8. (a) Gave and Took, so Anaya could see what the screen handed her and what it cost her. (b) is exactly the answer Anaya jumped to, but Ms. Rao never said videos were bad; she pointed at autoplay instead. [1]

Q9. (d) A crowd needs to be many AND different, which is why Anaya's eleven lines counted as one opinion. (c) is tempting because grown-up work is usually paid, but the library list was free. [1]

Q10. (c) Data is the raw material. (b) Labels is the near-miss: a label is what you add to data, not the name for the raw facts. [1]

Q11. (a) Colour is formatting, not a use of data. The five real uses are to decide, spot a pattern, predict, keep a record and check a change worked. [1]

Q12. (b) Column letter first, then the row number. (a) is the classic mix-up; if it were true the address would be written 5C. [1]

Q13. (b) The letters make the whole entry text, and SUM skips text without complaining. (c) is the tempting guess, but Sheets does not think anything is wrong, which is why the money goes missing quietly. [1]

Q14. (c) Formatting changes how a cell looks, never what it is worth, so 40 still adds up as 40. (a) is the tempting one, because bold red text looks like a heading rather than money. [1]

Q15. (a) Z to A means largest first for numbers. (b) is the easy slip: A to Z sounds like first, but it puts the smallest number on top. [1]

Q16. (c) Gaps in the row numbers mean the rows are still there but out of sight. (a) is what the topic exists to prevent: deleted rows would renumber themselves 2, 3, 4 with no gaps. [1]

Q17. (b) A function is not a picture of an answer, it is a machine that keeps answering, so the total changes on its own. (d) is what a calculator would need, and that is exactly why the sheet is safer than one. [1]

Q18. (d) The slash divides. (a) is the maths-book sign, and like the multiplication cross it is not a keyboard key, so Sheets will not accept it. [1]

Q19. (c) Every slice of a pie is a share of the same total, and together they make 100%. Bars and columns in (a) and (b) compare separate amounts instead. [1]

Q20. (b) Bar heights are easy to judge against a scale, so you see both which is bigger and the size of the gap. (a) hides small gaps, because slices of nearly equal size look the same. [1]

Section — Multiple Choice — Understand & Apply

Q21. (d) Small safe pieces add up. Name plus school plus class narrows it to one child in one building. Option (b) goes too far — a class on its own describes millions of children. [2]

Q22. (d) Restricted keeps the audience to the class she names, Viewer lets them read, and only Vihaan gets Editor. Option (b) gets the roles right but hands the link to anyone it is ever forwarded to. [2]

Q23. (c) A badge names your school, which is private, and the asking is itself the warning sign. Option (b) feels fair, but whatever he sends back cannot be checked, and yours is already gone. [2]

Q24. (b) A download is a new copy that belongs to somebody else, so his delete button cannot reach it. Option (a) sounds sensible but nothing quietly finishes deleting somebody else's copy later. [2]

Q25. (c) Searches are recorded automatically, so they are a passive footprint. Options (a), (b) and (d) are active — she decided each one before it existed. [2]

Q26. (a) The audience is what keeps a mean post alive, so laughing joins the crowd. Option (c) is the excuse children use most, but Farhan counts the laughs, not the words. [2]

Q27. (c) A private message tells the target he is not alone, which is the thing only he can feel. Option (b) feels tidy and safe, but leaving protects you and leaves Farhan inside with the crowd. [2]

Q28. (b) Somebody made the picture, so name them. Option (c) is the exact mistake to avoid — putting your own name on someone else's work makes it worse, not better. [2]

Q29. (d) Choosing is easy before you start and very hard once autoplay is running. Option (c) misses the point — the length is not what saves her, the deciding-first is. [2]

Q30. (a) One person repeating themselves is not a crowd, so one entry each stops the flooding. Option (c) throws away the reason crowdsourcing works — 200 children walk far more road than a few teachers. [2]

Q31. (b) A label and context let the number answer a question. Option (a) is tempting because sheets feel organised, but a lonely 34 in cell B7 is still 34 of nothing. [2]

Q32. (b) Four days sit at 2 or 3 minutes, and Thursday is a single odd day worth mentioning. Option (d) is the dangerous one — deleting a number you dislike destroys the record you may need later. [2]

Q33. (b) B4. Tickets live in column B and 4C is in row 4. Option (a) A4 is the same row but holds the words 4C, not the number. [2]

Q34. (c) Plain text, because a room code only looks like a number and is never added up. Option (d) Number is exactly what caused the problem — numbers drop leading zeros. [2]

Q35. (d) A header only works when it differs from the rest, so the data must go back to plain. Options (a) and (c) change every row equally, so row 1 still looks exactly like the data. [2]

Q36. (c) 39. The names never moved, so Ishan keeps row 1 while the smallest mark slides up to meet him. Option (a) 62 assumes the rows stayed joined, which is exactly what a one-column sort breaks. [2]

Q37. (b) Three — rows 4, 5 and 7 are missing from the run. Option (d) is the fear this topic exists to remove: gaps prove the rows are still there, because deleting would renumber them with no gaps. [2]

Q38. (c) 215, because the largest is 310 and the smallest is 95, and 310 − 95 = 215. Option (d) 405 adds them instead of subtracting, and (a) 145 uses 240 as the largest by mistake. [2]

Q39. (a) 270, because 9 × 30 = 270 and the formula reads whatever is in C3 right now. Option (b) 360 is the old answer, 9 × 40, which is what a typed formula like =9*40 would still be showing. [2]

Q40. (a) 40%, because 10 out of 25 is 10 ÷ 25 = 0.4. Option (b) 10% simply repeats the count of children, which is the mistake a percentage question is built to catch. [2]

Section — Multiple Choice — Think & Decide

Q41. (b) The rule was never never share. It is share with family, teachers, doctors and anyone who genuinely needs it, and never with a stranger offering a prize. A hospital needs the date of birth for the right dose and the address to send results. What decides it is who is asking, why, and whether a parent is there. (c) reaches the same answer for a reason that would also excuse a fake hospital website. Marking guidance: Tests the rule as a test of who and why, not as a list of forbidden words. (c) is right by accident, which is the distinction being examined. [3]

Q42. (c) Restricted is the tightest audience and is correct for anything private, but safest is not the same as best. A notice the whole school must read, or a lost-and-found list for parents, fails at its job if six named people can open it. The rule is the smallest audience that still does the job, and sometimes that audience is genuinely large. Marking guidance: Tests matching a setting to a purpose. (a) is the tidy rule most children want; the question is whether they can find the case where it fails. [3]

Q43. (a) A photo is a picture somebody chose, and any picture on the internet can be chosen. A friend count is a number the app shows, and anybody who sends enough requests collects one. Both are things you can see and neither is anything you can check. (b) and (d) each rescue one of the two, which is the half-answer this question is built to catch. Marking guidance: Tests knocking down both pieces of evidence rather than one. The replacement test must rest on real-life confirmation. [3]

Q44. (d) The choice is made for her. The comment sits in her own account, so it has a delete button. The login record is passive and belongs to the app, so there is no button at all. The cousin's photo sits in his account, so the only route is asking him. (c) picks the right item for a wrong reason — deleting never reaches a copy somebody else already holds. Marking guidance: Tests who owns each trace. (a) is a sensible priority that turns out to be impossible, and (c) is the right answer with the wrong mechanism. [3]

Q45. (c) Loyalty does not mean helping a friend keep doing harm, and the price of that secret is paid by somebody who did nothing. This route is kind to both: the friend gets the chance to fix it, the target learns somebody noticed, and the adult step still happens if the post stays. (d) reaches the right end but skips the step that lets the friend put it right. Marking guidance: Tests deciding in favour of the target while still treating the friend fairly. (d) is defensible and worth discussing, but it gives up on the friend too early. [3]

Q46. (b) Permission is not a vote, because what is being decided is not shared property — it is each person's own face. Four people cannot give away the fifth person's yes any more than they could give away her password. (c) is nearly the right idea: cropping does solve it, but only if she agrees to the crop, because it is still her photo being edited. Marking guidance: Tests consent as individual rather than pooled. (c) is the practical answer done without asking, which is the same fault in a smaller size. [3]

Q47. (c) As clock time the hours really are equal, and that half of the claim is true. What differs is what each hour takes. The 4 pm hour competes with play or homework, which can be moved. The 10.30 pm hour competes with sleep, which cannot be borrowed back, and the cost lands next morning. Same length, different price. (b) picks a real cost, but one that can be rescheduled. Marking guidance: Tests conceding the true half before rejecting the conclusion. (b) is a genuine competing position that a child can argue well. [3]

Q48. (d) Crowdsourcing works because many different people cover ground one person cannot, and small mistakes cancel out. Four ratings could be four friends of the author. The 5.0 is fragile — one 2-star rating drags it a long way, while the 3.9 would barely move. (b) reaches the same choice but claims size alone settles it; 200 ratings from one class is big and still leaning. Marking guidance: Tests preferring the larger sample for the right reason. (b) is the same verdict without the condition that would undermine it. [3]

Q49. (b) Every number needed to compare July reading is present, so who read most is answerable and the total is 23. Improvement is a comparison between two times and June was never recorded, so no amount of sorting can recover it. (a) is the subtle trap — the sheet counts books, not pages, and a column that was never collected cannot be worked out later. Marking guidance: Tests that unrecorded facts cannot be recovered by organising. (a) and (c) both invent a column the sheet never had. [3]

Q50. (c) The three numbers sit within two of each other, so about 23 is sensible and cooking 25 leaves a margin. Three repeats is enough to call it a pattern where two would not have been. The break must be far outside the ordinary wobble, and fixing that boundary before Monday matters — otherwise any number can be explained away afterwards. (d) makes wobble itself count as a break. Marking guidance: Tests predicting from a small pattern and fixing the test in advance. (a) invents a trend the three numbers do not show. [3]

Q51. (a) Everything from the old row 3 downwards shifts one row down, and the new empty row becomes row 3. The danger is that B3 now holds nothing, so any note that said Bhavya is in B3, or a total typed by hand, is wrong and does not look wrong. Sheets adjusts its own formulas when rows are inserted, but it cannot fix a sentence a person wrote. Marking guidance: Tests tracing a row insert, plus the point that a written-down address goes stale silently. [3]

Q52. (d) The gap is exactly ₹250, which points at one whole entry rather than a scatter of small errors, and SUM skips text without complaining. The clue is on screen: that entry hugs the left edge while real numbers sit right. The fix is to retype 250 with digits only and add the ₹ through Format, Number, Currency. (a) would produce the same shape of gap, which is why the question rules it out. Marking guidance: Tests reading an exact gap as evidence. (a) is genuinely tempting until the child re-reads the conditions given. [3]

Q53. (b) Formatting changes how a cell looks and never what it holds, so MAX would have found that same number if every cell were plain black. The colouring is not useless — it helps a human eye land on the number quickly — but it helps the reader, not the function. If the range were wrong, no amount of red would rescue it. Marking guidance: Tests that a function reads values, not appearance. (d) is the half-belief that colour helps a function as long as it is used sparingly. [3]

Q54. (c) Z to A on runs means largest first: 34, 28, 22, 19. A to Z on balls means smallest first: 10, 12, 30, 40. The two orders nearly reverse — Pooja is second on runs and last on balls. (d) is the real trap: sorting cannot decide what best means. Nithya scored most, Rehan used fewest balls, and both sorts answer different questions correctly. Marking guidance: Tests both sort directions and the point that the sort column is a choice about the question. (a) and (b) each reverse one of the two. [3]

Q55. (a) There is no disagreement, only two questions answered at once: 7 + 4 + 6 = 17 on screen, and 7 + 4 + 6 + 5 + 9 = 31 for the column. Filtering hides rows from you without hiding them from the function, and the missing 14 is exactly 5 + 9. To total only what is visible, add the three by hand or SUM just those cells — and label it, because the next filter change makes that label untrue. Marking guidance: Tests both totals and the gap between them. (b) is the instinct to blame the formula whenever two numbers disagree. [3]

Q56. (d) Checking first: 45 + 52 = 97, + 38 = 135, + 61 = 196, + 47 = 243, + 55 = 298, + 46 = 344, so the formula added its range perfectly. B9 is outside that range, so the missing score is 396 − 344 = 52. (b) gets the number right and the diagnosis wrong: the formula is obeying an instruction that is out of date, and =SUM(B2:B9) fixes it. Marking guidance: Tests verifying the total before blaming anything. (b) is the same arithmetic with the wrong conclusion, which is the distinction being examined. [3]

Q57. (c) The letters make that cell text, and MIN skips text entirely, so among the three real numbers 95 genuinely is smallest. The function is right about what it can see and wrong about the question asked — which is worse than an error message, because an error shouts and gets fixed while 95 looks ordinary and ends up in a report. Rs 80 sits against the left edge; that is the clue. Marking guidance: Tests why a plausible wrong answer is more dangerous than a visible error. (d) is the read-as-zero belief this topic exists to correct. [3]

Q58. (b) Anaya's cell reads the new price and shows 18 × 25 = 450. Kabir's numbers were typed by hand and know nothing about any cell, so it is a sum frozen on the day he typed it. Both showed 360 before the change, which is exactly what kept the fault invisible, and they now differ by ₹90 with nothing on screen to say which is right. Marking guidance: Tests what a typed formula actually knows. Credit is for noticing the two agreed before the change, which is why nobody spotted it. [3]

Q59. (d) The first fault is meaning — four schools' totals are separate amounts, and a pie would claim the 195 trees are one whole being divided. The second is eyesight: the shares are about 25%, 26%, 24% and 26%, and slices that close cannot be told apart. A bar chart fixes both, and against a scale the best school clearly beat the worst by 5 trees. Marking guidance: Tests finding two genuinely different faults. (c) accepts the misuse and tries to fix it with size, which does not address either fault. [3]

Q60. (a) The whole is 18 + 6 + 9 + 3 = 36, so cricket is 18 ÷ 36 = 50%, dance 25%, chess about 17% and art about 8%. Everything else together is 6 + 9 + 3 = 18, exactly equal to cricket, so the claim fails by the narrowest possible margin — a tie, not a win. A pie shows it at a glance: cricket fills half the circle and the other three fill the rest. Marking guidance: Tests comparing one category against the sum of the rest. (c) compares cricket with each club one at a time, which is the commonest misreading. [3]