Selecting from a data file when half of it is a trap
In Paper 3 Part B, Sze-chit is writing Task 4: a 250-word email to the operator of his school tuck shop proposing three changes to the menu and justifying each. The data file contains six items: (1) a survey of 300 students carried out last month; (2) a letter from a nutritionist dated 2021; (3) the tuck shop's current price list; (4) a school newsletter from 2019 about a "healthy eating week"; (5) a WhatsApp message from the PE panel about an inter-school swimming gala; (6) a school policy memo requiring every item sold on campus to carry a nutrition label. The recording, heard once, included three student quotations and a figure for the percentage of Form 1 students who buy fried food daily - Sze-chit missed the figure while writing. He has 30 minutes. His plan is to copy whole sentences from the survey and the nutritionist's letter, and to fill the gap in his notes with a figure that "sounds about right".
1. Which data-file items should Sze-chit use, which should he leave out, and on what criteria?
Read the worked answer
Use
- (1) Student survey, last month - current, directly about the people the tuck shop serves, and it yields quantities (300 students, proportions) that can justify each proposal.
- (6) Policy memo - this is a constraint, not evidence. Proposals that break it are unusable, so it should shape every recommendation (for example, any new item must be able to display a label).
- (3) Price list - used narrowly, to keep proposals feasible (a healthy option priced at the level students already pay) rather than to describe the shop.
Use with care
- (2) Nutritionist's letter - the principle (fried food frequency, sugar content) is still valid, but it is four years old. Cite it as background, not as a claim about the school now.
Leave out
- (4) 2019 newsletter - superseded, and a school newsletter is promotional rather than evidential.
- (5) Swimming gala message - irrelevant to menu design.
The criteria behind the decisions
Four filters, applied in this order: relevance to the purpose of the email, currency, authority/provenance, and whether the item constrains or merely illustrates. Integrated tasks credit selection, so a script that uses everything is not better than one that uses three items well; a brief acknowledgement of why an item was not used can itself show understanding.
2. Sze-chit missed a spoken figure. What should he do, and why is inventing a number the worst option available?
Read the worked answer
What to do
Write around the gap and keep going. Two workable moves:
- Generalise accurately: "the recording suggested that a substantial share of the youngest students buy fried food every day". This is faithful to what was heard without asserting a number he does not have.
- Substitute a figure he does have, with its source named: "According to last month's survey of 300 students, ...". A sourced, accurate number is worth more than an unsourced, invented one.
Why inventing is the worst option
- The mark scheme credits specific content points drawn from the data file and the recording. A fabricated figure matches no content point, so it cannot score - the effort is wasted as well as wrong.
- It corrupts the credibility of the whole email. A tuck shop operator who checks one number and finds it false has reason to distrust the rest, and examiners read for internal consistency too.
- It cannot be corrected later. "42%" written into a sentence will still be there at the end of the paper, whereas a general statement can be tightened in the final minutes if memory returns.
Practical discipline
The recording is played once, so the loss is permanent. He should leave a visible gap marker in his notes, keep writing the other tasks, and never stop the clock to hunt for it. Losing one content point is survivable; losing five minutes of writing time for the other tasks is not.
3. Explain why copying whole sentences from the data file costs marks, and demonstrate the transformation with one item.
Read the worked answer
Why lifting fails
- The data file is written in note and bullet form, not in the grammar of an email. Pasted into a sentence, it produces fragments, missing articles and wrong tenses - so lifted text actively creates language errors.
- Lifting does not show comprehension. The task exists to test whether he can read a source, judge it and reshape it for a new reader; unchanged text gives the marker nothing to credit.
- Long lifts distort the register. An email to the tuck shop operator should be polite and direct; a survey bullet is neither.
Transformation demonstrated
Source (bullet): "300 students surveyed - 68% want more fruit - price main complaint."
Written into the email:
In last month's survey of 300 students, more than two-thirds said they would buy more fruit if the price were lower, which suggests that cost rather than preference is the obstacle.
What changed, and why each change matters
- A reporting frame (In last month's survey of 300 students) tells the reader where the number comes from.
- The proportion is expressed in words (more than two-thirds) rather than as a pasted percentage, which reads naturally.
- The which suggests clause turns a statistic into an argument, which is what the task is asking for.
- The vague note price main complaint becomes a specific claim the operator can act on.
Rule: take the number and the finding, drop the wording.