The stem has two halves, and a wrong choice satisfies only one
Every question in this family looks like one of these:
Which choice most effectively uses data from the table to support the claim?
Which choice most effectively uses data from the graph to complete the statement?
Which choice best describes data that would support the researcher's conclusion?
Two conditions must hold at once. The choice must (1) report the figure accurately and (2) make the specific claim in the text more likely to be true. Test writers build the distractors by satisfying exactly one of the two. Most students check only the first.
Học sinh Việt vào dạng này thường làm đúng một việc: dò xem con số trong phương án có khớp bảng không. Khớp thì chọn. Nhưng ba trong bốn phương án đều khớp bảng — đề cố tình làm vậy. Việc thứ hai mới quyết định: câu đó có đẩy luận điểm đang bàn lên không. Đọc luận điểm trước, gạch chân nó, rồi mới đọc bốn phương án.
Read the figure before you read the claim
Spend eight seconds on the table before anything else. Name three things out loud:
- What is each column measuring? A count, a percentage, a rate per unit, an index?
- Which way is the ordering variable moving? Down the rows, is the first column rising or falling?
- Where does the pattern break? Almost every question is built on the one row where the trend stops, jumps, or reverses.
A researcher testing seed coatings planted four batches of the same variety and recorded the proportion that germinated. The results appeared to show that thickening the coating helps only up to a point, after which extra coating buys nothing.
| lcc@ Batch | Coating thickness (mm) | Germinated (%) |
|---|---|---|
| None | 0.00 | 41 |
| Light | 0.02 | 63 |
| Medium | 0.05 | 78 |
| Heavy | 0.09 | 79 |
Which choice most effectively uses data from the table to support the researcher's conclusion?
- Seeds with no coating germinated at 41%, the lowest rate of the four batches.
- Germination rose from 41% with no coating to 78% at 0.05 mm but only to 79% at 0.09 mm.
- The heaviest coating produced the highest germination rate, 79%.
- Germination fell when thickness rose from 0.05 mm to 0.09 mm.
The claim has two parts: coating helps, and then stops helping. (A) reads the table correctly and supports the first part only — it says nothing about a ceiling. (C) reads the table correctly and cuts against the claim, since it presents the heaviest coating as the winner. (D) misreads: 79 is higher than 78, not lower. Only (B) puts the large early gain and the trivial late gain side by side, which is exactly the shape the conclusion describes. Answer (B).
Phương án chỉ trích một dòng duy nhất trong bảng gần như không bao giờ là đáp án. Một con số đơn lẻ không so với gì cả thì không chứng minh được xu hướng, ngưỡng, hay khác biệt — mà luận điểm trong đoạn văn luôn là một trong ba thứ đó. Muốn ủng hộ một so sánh thì phải có hai số.
Four ways a choice dies
1. Đúng số, thiếu vế. Trích một dòng, hoặc chỉ nói được nửa đầu của luận điểm.
2. Đúng số, ngược chiều. Con số chính xác nhưng lại làm luận điểm yếu đi.
3. Sai số. Đảo chiều so sánh, đọc nhầm cột, hoặc đổi đơn vị.
4. Số không có trong bảng. So sánh theo thời gian khi bảng chỉ có một thời điểm; tách
nhóm nhỏ khi bảng chỉ có tổng.
A forestry team measured how much annual growth each age class of a conifer lost during a single drought year. The team argued that drought sensitivity does not build up gradually with age; it rises sharply only after a tree passes about sixty years.
| lcc@ Age class | Trees measured | Growth lost in drought year (%) |
|---|---|---|
| Under 20 years | 118 | 4 |
| 20–59 years | 142 | 6 |
| 60–99 years | 96 | 19 |
| 100 years and up | 61 | 24 |
Which choice most effectively uses data from the table to support the team's argument?
- Trees under 20 years lost 4% of their annual growth.
- Growth loss increased steadily across the four age classes, from 4% to 24%.
- Trees 100 years and older lost the most growth, 24%.
- Loss stayed below 10% in both classes under 60 years but jumped to 19% and 24% in the two classes above 60.
(B) is the trap built for this item: every number in it is right, and “4% to 24%” is a true range, but steadily describes a gradual ramp — the opposite of the threshold the team claims. (A) and (C) each quote one row and cannot show where a break occurs. (D) reports the flat stretch and the jump together, which is what a threshold claim needs. Answer (D).
Chú ý cặp từ steadily (đều đặn) và sharply / abruptly (đột ngột). Cùng một bảng số, hai từ này mô tả hai luận điểm khác hẳn nhau, và đề rất hay đặt cả hai vào hai phương án. Muốn biết chọn cái nào thì nhìn khoảng cách giữa các dòng: 4, 6, 19, 24 có một bước nhảy lớn giữa dòng hai và dòng ba, nên đó là ngưỡng, không phải xu hướng đều.
A city planner compared four cities that had adopted rooftop-garden programmes at different scales. She concluded that across the four cities, denser rooftop planting goes together with cooler rooftops in midsummer.
| lcc@ City | Rooftop gardens per 1,000 buildings | Mean rooftop temp., July ($^\circ$C) |
|---|---|---|
| Aveiro | 4 | 52 |
| Bregen | 11 | 48 |
| Cortal | 26 | 41 |
| Dunmar | 38 | 39 |
Which choice most effectively uses data from the table to support the planner's conclusion?
- Dunmar, with 38 gardens per 1,000 buildings, averaged 39 $^\circ$C, while Aveiro, with 4, averaged 52 $^\circ$C.
- Rooftop temperatures in Dunmar have fallen by 13 $^\circ$C since its garden programme began.
- Cortal has 26 rooftop gardens per 1,000 buildings.
- Bregen has fewer gardens than Cortal and a lower mean rooftop temperature.
(B) is the fourth death: the table has one July, so nothing in it can tell you what Dunmar was like before the programme. Notice how easy (B) is to accept — $52-39=13$, so the number even looks as though it came from the table. (C) quotes one cell. (D) reverses the reading: Bregen's 48 $^\circ$C is higher than Cortal's 41 $^\circ$C. Answer (A).
Bẫy “số không có trong bảng” rất khó thấy vì phương án thường ghép hai con số có thật thành một phép trừ nghe hợp lý. Cách chặn: hỏi bảng này có cột nào đo cái đó không. Bảng chỉ có một năm thì mọi câu chứa since, has risen, before the programme đều chết ngay.
Count, share, and rate answer three different questions
A count answers “how many”. A share answers “what fraction of this group”. A rate answers “how many per unit of exposure”. A claim about priority, concentration, or intensity almost always needs a share or a rate; a choice that answers it with a raw count is accurate and useless.
An archive reports on its digitisation programme. Its director states that the programme has deliberately prioritised the archive's smaller language collections.
| lcc@ Language | Manuscripts held | Manuscripts digitised |
|---|---|---|
| Latin | 4,800 | 1,440 |
| Arabic | 900 | 540 |
| Greek | 600 | 420 |
| Syriac | 200 | 160 |
Which choice most effectively uses data from the table to support the director's statement?
- More Latin manuscripts have been digitised than manuscripts in any other language.
- The Syriac collection is the archive's smallest, with 200 manuscripts.
- Fewer Greek manuscripts have been digitised than Arabic manuscripts.
- The share digitised rises as the collection shrinks: 30% for Latin, 60% for Arabic, 70% for Greek and 80% for Syriac.
Every count in (A) and (C) is correct, and both point the wrong way: the biggest collection has the most digitised files, which is what you would expect if nothing had been prioritised. (B) quotes a single figure with no link to digitisation. Dividing gives $1440/4800=30\%$, $540/900=60\%$, $420/600=70\%$, $160/200=80\%$ — the shares run opposite to collection size, which is what “prioritised the smaller collections” means. Answer (D).
Khi đề nói prioritised, concentrated in, disproportionately, relative to its size — đó là tín hiệu bắt buộc dùng tỉ lệ, không dùng số đếm. Ngược lại, khi đề nói total burden, overall volume, the largest number of thì mới dùng số đếm. Nhớ cặp này thì loại được hai phương án trước khi làm phép chia nào.
The claim's direction word decides the answer
The verb and adverb in the claim tell you the shape of the evidence you need.
only after / not until / beyond a certain point $\rightarrow$ a flat stretch and a jump.
steadily / progressively $\rightarrow$ roughly equal steps across every row.
despite / even though $\rightarrow$ two figures pulling in opposite directions.
concentrated in $\rightarrow$ a share or rate, contrasted with the rest.
emerges over time / only in the long run $\rightarrow$ two time points, reversed.
In a study of study methods, one group reviewed a word list four times in one sitting (massed) and another reviewed it four times across four days (spaced). Both groups were tested twice. The authors argued that the advantage of spacing does not appear at the shorter delay and emerges only over a longer interval.
| lcc@ Method | Recall at 1 week (%) | Recall at 6 months (%) |
|---|---|---|
| Massed | 82 | 31 |
| Spaced | 74 | 58 |
Which choice most effectively uses data from the table to support the authors' argument?
- Spaced review produced 58% recall at six months.
- Massed review produced the study's highest single figure, 82% at one week.
- At one week massed review led 82% to 74%, but at six months spaced review led 58% to 31%.
- Neither method retained more than 60% of the material at six months.
The claim is a reversal across the two delays, so the evidence must contain both time points and both methods. (A) has one cell. (B) is accurate and argues for massed review. (D) is accurate — 58 and 31 are both under 60 — and compares the two methods to a threshold nobody mentioned. Answer (C).
A researcher recorded the proportion of male birds singing the local dialect at four distances from the centre of a population's range. She argued that dialect conformity holds firm across the core of the range and then breaks down abruptly near the edge.
| lc@ Distance from centre | Males singing local dialect (%) |
|---|---|
| 0 km | 96 |
| 5 km | 88 |
| 15 km | 61 |
| 30 km | 24 |
Which choice most effectively uses data from the table to support the researcher's argument?
- At 30 km, 24% of males sang the local dialect.
- Conformity fell at every step outward, from 96% to 88% to 61% to 24%.
- Conformity stayed above 85% out to 5 km but had fallen to 24% by 30 km.
- Conformity was higher at 15 km than at 5 km.
(B) is the sharpest distractor in the unit: it is a complete and accurate description of the data, and it describes a gradual decline, which is the argument the researcher is rejecting. (D) reverses two rows. (A) is a single point. (C) reports the flat core (96 and 88, both above 85) and the collapse at the edge. Answer (C).
When the text carries more than one claim
Harder items put two researchers, or a hypothesis and an objection, into the same short text. The stem then names one of them. Every choice will support somebody.
At an excavated wetland, one specialist argues that people burned the land to clear it for the first cereal crops, so burning should peak as farming begins. A second specialist argues that cereal farming kept expanding after burning was given up.
| lcc@ Layer | Charcoal (fragments per litre) | Cereal pollen (% of total) |
|---|---|---|
| 1 (oldest) | 0.2 | 1 |
| 2 | 1.9 | 3 |
| 3 | 3.4 | 22 |
| 4 (youngest) | 0.6 | 31 |
Which choice most effectively uses data from the table to support the second specialist's argument?
- In layer 4 charcoal falls to 0.6 fragments per litre while cereal pollen rises to 31%.
- Charcoal peaks in layer 3 at 3.4 fragments per litre, the layer in which cereal pollen first exceeds 20%.
- Cereal pollen increases from layer 1 to layer 4.
- Charcoal is lowest in layer 1, where cereal pollen is also lowest.
(B) and (D) are accurate and both serve the first specialist: burning and farming rise together. (C) is accurate and neutral — both specialists agree farming expanded. Only (A) separates the two variables in the way the second specialist needs: burning collapses while farming keeps climbing. Answer (A).
Với dạng hai luận điểm, việc đầu tiên không phải là đọc bảng mà là khoanh tên trong câu hỏi: the second specialist, Reyes's hypothesis, the objection. Đọc nhầm chủ nhân luận điểm là mất câu, dù đọc bảng đúng hoàn toàn. Hai phương án bao giờ cũng được viết sẵn cho người kia.