How to Learn Advanced English Vocabulary (Chinese Edition) | Lesson No.037 | Master Quantity, Proportion, Percentage, Ratio, Rate and Share Without Comparing Different Things as If They Were the Same | 第037课：掌握数量、比例、百分比、比率与占比，避免把不同指标当成同一种量

Series ID: EDKS-ADV-VOC-ZH-0037 · Advanced English Vocabulary (Chinese Edition) · Lesson No.037 · C1 → C2 · 简体中文辅助

A percentage is not automatically a rate. A ratio is not automatically a proportion. A share needs a whole. And every number needs a denominator somewhere, even when the sentence hides it.
percentage 不一定是 rate；ratio 不一定是 proportion；share 必须有整体；而每一个数字背后，都有一个比较基准或分母，即使句子没有把它写出来。

Advanced learners often understand the arithmetic but choose the wrong English relationship word. A report says 30% of students; the learner rewrites it as a rate of 30%. A company’s market share becomes its ratio. A teacher–student ratio becomes a proportion of teachers. These sentences can contain correct numbers and still express the wrong relationship.

Cambridge defines a ratio as a relationship between two groups or amounts, commonly expressed as how one compares with the other. It defines proportion as the amount of a part compared with a whole or another amount, and percentage as an amount expressed out of 100. Those definitions overlap enough to confuse learners, but the hidden architecture differs: part-to-whole, part-to-part, per-unit, per-time, share-of-total, absolute amount.

Before choosing the noun, identify the quantity, the comparison base and the unit.

选词之前，先找出：量是什么、跟什么比较、单位是什么。

This lesson owns the Mandarin-supported C1–C2 lexical system for quantity, amount, number, proportion, percentage, ratio, rate, share, fraction, per-unit and per-capita language. It coordinates but does not replace Lesson No.029 on degree, Lesson No.031 on comparison, Lesson No.034 on thresholds, Lesson No.035 on probability and Lesson No.036 on change over time.

---

Part I — Build the quantity architecture | 第一部分：先建立“数量关系架构”

1. Quantity is the broad parent concept | quantity = 广义“量”

Quantity is a broad formal noun for how much or how many of something exists. It can refer to countable or measurable things depending on context: a large quantity of water, quantities of data, production quantity. It does not by itself tell the reader whether the quantity is a count, volume, percentage, rate or ratio. Use it when the exact measurement type is secondary or already known.

2. Number usually answers “how many?” | number = 可数对象数量

Number is natural for countable units: number of students, number of errors, number of applications. “The number of applicants increased” refers to the count. It does not tell us the proportion of all eligible people who applied. Ten thousand applicants may be a large number but a small proportion if the eligible population is enormous.

3. Amount usually answers “how much?” | amount = 不可数或总量

Amount is common for uncountable quantities such as money, time, water, energy or information. In everyday English people sometimes use amount with count nouns, but careful formal writing normally prefers number of people/items. This distinction is simple but useful because quantitative accuracy begins before percentages appear.

4. Total names the whole count or amount | total = 总数/总量

A total is the complete amount after components are combined: total enrolment, total cost, total errors. Part-to-whole language requires knowing the total. If 30 out of 100 students pass, 30 is the part and 100 is the total. The proportion is 30/100; the percentage is 30%.

5. Count is both a process and a result | count

A count can mean the act/result of counting: case count, headcount, word count. Counts are absolute numbers. A count can rise while its proportion falls if the total grows faster. This distinction becomes essential in population, education and business reporting.

6. Proportion usually places a part against a whole | proportion = 部分占整体

Cambridge’s core use is part compared with the whole: “A large proportion of students completed the task.” The denominator is the whole group. A proportion can be expressed as a decimal, fraction or percentage. The English noun describes the relationship; the numerical format describes how it is written.

7. Proportion can also compare two quantities | broader proportion sense

Cambridge also records a broader comparison sense, including “the proportion of women to men.” This overlap with ratio is real. For learners, a useful default is: proportion often foregrounds a part relative to a whole; ratio often foregrounds the relationship between two quantities. Domain usage can vary, so follow conventional phrasing.

8. Percentage is a part expressed per hundred | percentage = 每一百中的多少

A percentage expresses a quantity out of 100. If 30 of 120 students pass, the proportion is 30/120 = 0.25 and the percentage is 25%. The count, proportion and percentage describe related facts but are not interchangeable nouns.

9. Percent and percentage have different grammar | percent vs percentage

“Twenty percent of students passed.” “The percentage of students who passed was 20%.” Percent normally follows a number; percentage is a noun that can be modified: high percentage, small percentage. Avoid “20 percentage of students”.

10. Percentage point is not percent | percentage point

If a rate rises from 20% to 30%, it rises by ten percentage points but by 50% relative to the original 20%. Both statements can be mathematically correct, but they answer different questions. Advanced writing often reports both when framing could otherwise mislead.

11. Fraction is a part of a whole or a mathematical expression | fraction

“Only a small fraction of applicants were selected” is natural general English meaning a small part. In mathematics, fraction has a formal numerator/denominator structure. The phrase a fraction of often emphasises that the share is small, even without an exact number.

12. Share is a part owned, contributed or belonging to a whole | share

Share often appears where the whole is a market, budget, workload, vote, responsibility or total: market share, share of revenue, share of emissions. Unlike percentage, share does not require expression out of 100, though it is often reported as a percentage. “The company’s share rose to 30%” combines the conceptual share with percentage format.

13. Market share is not market size | share vs total

A company can increase sales while losing market share if the total market grows faster. Conversely, market share can rise while absolute sales fall if competitors shrink even more. Share requires a denominator: company sales divided by total market sales under the defined measure.

14. Ratio compares two quantities | ratio = 两个量之间的关系

Cambridge defines ratio as the relationship between two groups or amounts. “The student–teacher ratio is 15:1” means fifteen students for each teacher. Neither side must be part of a single whole in the way a percentage usually is. Ratios can compare different units or categories.

15. Ratio order matters | A-to-B is not B-to-A

A student-to-teacher ratio of 15:1 is not the same expression as a teacher-to-student ratio of 1:15, though they describe the same relationship from opposite directions. Always name the order before giving the numbers.

16. Rate usually contains “per” something | rate = 每单位的发生/变化

A rate commonly relates one quantity to another unit such as time, population, distance or exposure: kilometres per hour, errors per thousand transactions, births per 1,000 people per year, interest rate. The denominator or unit is central. Calling every percentage a rate hides whether there is a per-time or per-population relationship.

17. Some conventional rates are percentages | rate can be percentage-form

Interest rate, unemployment rate, tax rate and response rate are often expressed as percentages. That does not make rate and percentage synonymous. Rate names the metric/relation; percentage describes one numerical representation. “The unemployment rate is 4%” is natural because the metric is conventionally called a rate.

18. Per-capita language divides by population | per capita

Per capita means per person in a population. Total spending can rise while spending per capita falls if population rises faster. The phrase is especially common in economics, public policy and demography. It is not interchangeable with percentage of population.

19. Per-unit language reveals the denominator | per unit / per user / per hour

Cost per student, errors per page, requests per second, revenue per user: these metrics normalise totals by a meaningful unit. Two organisations with different sizes may be more fairly compared using per-unit measures, but only if the denominator itself is appropriate.

20. Density is amount per space/volume/area in many contexts | density

Population density, word density and energy density each relate an amount to some unit of space, text or volume. Lesson No.026 owns lexical density as a language concept; this lesson simply places density inside the wider family of normalised quantities.

21. Frequency counts occurrence within a period or set | frequency preview

Frequency answers how often something occurs and will be owned deeply by Lesson No.040. Here it is useful as contrast: a rate may include frequency but also a denominator/time base; a percentage may show the share of observations belonging to a category. Do not collapse these metrics.

22. Average is not a proportion | average preview

An average summarises central value, not part-to-whole relationship. “Average class size is 20” and “20% of classes are large” say entirely different things. Lesson No.038 will own average/median/mode and typicality.

23. Denominator is the hidden architecture | denominator

When a quantitative sentence is unclear, ask “out of what?” A percentage without a denominator can mislead. “50% passed” means little if the reader does not know which group is included. “Cases doubled” also needs a baseline. The denominator often carries the real meaning of the statistic.

24. Numerator is the counted/measured part | numerator

In a fraction or rate, the numerator is the quantity being counted relative to the denominator. In general communication you may not say “numerator,” but thinking in numerator/denominator terms prevents errors: What is counted? Compared against what? Over what period?

25. Unit is part of meaning | units cannot be stripped casually

100 dollars, 100 students, 100 cases per year and 100 cases per 100,000 people are not the same kind of number. Units belong to the proposition. A precise learner carries the unit through paraphrase and summary instead of treating the number as self-explanatory.

26. Scale changes interpretation | scale

A rate per 100 people and a rate per 100,000 people may contain the same underlying relationship but look numerically very different. Always retain the scale. “50 per 100,000” cannot be paraphrased “50%”.

27. Base names the reference quantity | base

A percentage increase is calculated relative to a base value. If revenue rises from 10 to 15, the absolute increase is 5 and the relative increase is 50%. The same +5 change from 100 to 105 is only 5%. The base changes the percentage.

28. Absolute difference and relative difference are different | absolute vs relative

Absolute difference keeps original units; relative difference compares the change with a baseline. Reporting only one can distort interpretation. Advanced writing should often give both when audience decisions depend on scale.

29. “Twice as many” is not “200 percentage points more” | multiplicative language

If Group A has 20 cases and Group B has 40, B has twice as many cases, or 100% more cases. Percentage-point language applies when the quantities themselves are percentages/rates expressed in percentage terms, not to raw counts.

30. Part I checkpoint: identify five things | 第一部分检查点

Before choosing a quantitative noun, identify: (1) what is being counted/measured, (2) the whole or comparison quantity, (3) the unit, (4) the time period if any, and (5) the numerical format. Then choose count, amount, proportion, percentage, ratio, rate, share or another metric. The word should reveal the relationship, not hide it.

Part II — Learn the denominator families | 第二部分：掌握不同“分母家族”

31. Part-to-whole metrics ask “what share of the whole?” | 部分占整体

If 30 out of 120 students pass, the part-to-whole relationship is 30/120. We can express this as a fraction (one quarter), a proportion (0.25 or 25% depending convention), a percentage (25%), or a share (25% of the cohort). The vocabulary changes packaging, but the denominator remains the whole group of 120. If the denominator changes, the meaning changes even when the numerator stays 30.

32. Part-to-part metrics ask how two categories compare | 部分与部分比较

Suppose a class has 30 boys and 20 girls. The boys-to-girls ratio is 30:20, or 3:2. The proportion of boys in the class is 30/50 = 60%. These are related but not identical descriptions. The ratio compares boys with girls; the proportion compares boys with the total class. A learner who writes “the proportion of boys to girls is 3:2” may be understandable, but ratio is the clearer conventional noun for the part-to-part relation.

33. Rate metrics ask “per what unit?” | rate 的分母通常有单位

A rate often normalises a count by time, population, distance, exposure or another meaningful unit. Examples: 60 kilometres per hour, 5 errors per 1,000 transactions, 30 cases per 100,000 people per year. The numerator and denominator may have different units. This is one reason rate cannot be reduced to “percentage”.

34. Rates can be expressed per 1, per 100, per 1,000 or per 100,000 | scale convention

The scaling constant is chosen for interpretability. A rate of 0.0005 per person is the same underlying rate as 50 per 100,000 people, but the second is easier to read in some domains. Preserve the scale because dropping “per 100,000” can transform a rare-event rate into an absurdly large percentage.

35. Time can appear in the numerator, denominator or both | time architecture

Revenue per month, cases per year, kilometres per hour and growth per quarter are not interchangeable. A rate should tell readers whether time is part of the denominator or merely the reporting period. “100 cases in 2026” is a count observed during a year; “100 cases per 100,000 person-years” is a normalised rate.

36. Exposure denominators matter | per exposure / per transaction

Ten failures among ten machines is very different from ten failures among one million transactions. Technical systems therefore use denominators such as per operating hour, per kilometre travelled, per device-year, per transaction. The denominator defines opportunity for the event to occur. Comparing raw counts across different exposure levels can be misleading.

37. Response rate is a conventional part-to-whole metric called a rate | conventional naming matters

Some metrics are called rates even though they are essentially proportions: response rate, completion rate, pass rate, unemployment rate. Learn conventional names rather than trying to force every metric into a pure mathematical taxonomy. The key is to understand the relation behind the label so you can interpret and paraphrase it correctly.

38. Conversion rate is another proportion-like rate | conversion rate

If 50 of 1,000 visitors purchase, the conversion rate is 5%. The denominator is eligible/observed visitors under the chosen definition. Changing the denominator from all visitors to qualified leads can change the rate dramatically even when the number of purchases is identical.

39. Pass rate is not the same as average score | rate vs average

A class can have a high average score but low pass rate if scores are polarised, or a modest average with high pass rate if most students cluster just above the threshold. One metric describes central value; the other describes the share meeting a criterion. Lesson No.038 will own this central-value distinction.

40. Market share is a part-to-whole business metric | market share

Market share expresses a company’s sales, revenue, users or another defined quantity as part of the total market under that same definition. A “30% market share” is meaningless until the market boundary, period and metric are specified. Unit sales share and revenue share can differ.

41. Vote share, audience share and budget share follow the same logic | share family

Vote share is votes for a candidate/party divided by a defined vote total. Audience share is a broadcaster’s audience relative to a defined total audience. Budget share is one category’s spending relative to total spending. The noun share foregrounds allocation within a whole.

42. Share can exceed 100% only if the metric is not part-to-whole | check the architecture

A genuine part-of-whole share should normally lie between 0% and 100% under a single non-overlapping classification. If categories overlap, summed “shares” may exceed 100%. For example, respondents can select multiple reasons. The English description should warn readers when percentages are not mutually exclusive.

43. Ratio can compare unlike units | price-to-earnings, debt-to-income

Ratios often compare quantities with different conceptual roles: debt-to-income ratio, price-to-earnings ratio, student-to-teacher ratio. The two quantities may share units or not. The noun tells readers that the comparison itself is the metric.

44. “X times as high” and ratio language need care | multiplicative comparisons

If A = 30 and B = 10, A is three times B, or the ratio A:B is 3:1. Phrases such as “three times higher” are widely used but can be interpreted inconsistently. In precise writing, “three times as high as” or “three times the value of” is often clearer.

45. Odds are a ratio, but not the same as probability | odds

Lesson No.035 introduced odds as likelihood language. Formally, odds compare favourable to unfavourable outcomes under a defined model, while probability compares favourable outcomes with all outcomes. For a probability of 0.75, odds in favour are 3:1. General learners need not calculate odds often, but should know the relationship is not identical to percentage probability.

46. Fraction can carry vague meaning in prose | a fraction of

“Only a fraction of users complained” usually means a small part, not a mathematically specified fraction. This is an example of technical vocabulary becoming qualitative in general prose. If exact magnitude matters, give the number or percentage.

47. Majority and minority are proportion words | majority / minority

A majority is more than half under the relevant definition; a minority is less than half. But “minority” can also refer to social groups and therefore carries additional meanings. “A large minority” is possible: less than half but still substantial. Do not equate minority with tiny.

48. Plurality is not always majority | plurality

In some electoral and general contexts, a plurality means the largest share among options even if below 50%. This is a valuable example of part-to-whole language: the largest category need not form a majority.

49. “Most” is not a precise percentage | most / many / several

Most usually means a majority, but natural-language quantifiers such as many, several, few, a substantial proportion are context-sensitive. If exact data exist and matter, numerical reporting is safer than relying on vague quantifiers.

50. “Almost all” is strong part-to-whole language | almost all

Almost all signals a very high proportion but not universality. It should not be paraphrased as all. If the source reports 97%, preserving the number may be more transparent than choosing a verbal category.

51. Per-capita normalisation can reverse rankings | total vs per capita

Country A may have higher total emissions than Country B but lower emissions per capita. Both statements can be true. Quantitative language should reveal whether comparison is absolute or normalised. A headline using totals and a report using per-capita values may appear contradictory while measuring different questions.

52. Per-household, per-user and per-employee are different normalisations | denominator choice

Revenue per user, revenue per employee and revenue per household each answer different questions. A denominator is never a neutral formatting choice. It encodes what “fair comparison” means for the task.

53. Weighted averages and weighted shares depend on weights | preview

Some summaries weight observations differently because groups differ in size, importance or sampling design. The word weighted tells the reader the calculation does not treat every item equally. Lesson No.038 will discuss central values; here the key lexical lesson is that weights alter the effective denominator/contribution.

54. Composition asks what the whole consists of | composition

Composition describes the mixture of categories that make up a whole: workforce composition, portfolio composition, class composition. If the share of one category rises, another category’s share must usually fall when categories are exhaustive and non-overlapping, even if absolute counts rise for both.

55. Compositional effects can mislead trend interpretation | denominator shifts

If the total group changes, percentages can shift without any individual category shrinking in absolute terms. Suppose Group A rises from 40 to 50 while Group B rises from 60 to 100. A’s count rises, but its share falls from 40% to about 33%. Count and share tell different stories.

56. Percentage change and change in percentage are different | wording precision

“The percentage changed from 20% to 30%” describes a percentage-point increase of 10. “The percentage increased by 50%” describes relative growth from 20 to 30. “A 10% increase” could mean something entirely different depending on the base. Avoid ambiguous shorthand where decisions matter.

57. Percent of a percent creates confusion easily | nested percentages

If 20% of users click and 10% of clickers buy, buyers are 2% of all users, not 10% of all users. When percentages are nested, explicitly identify the denominator at each stage. Funnel metrics frequently fail because the denominator silently changes.

58. Denominator drift is a communication failure | denominator drift

Denominator drift occurs when a paragraph begins with one reference population and later reports a percentage from another without making the switch clear. “60% of applicants passed the first stage; 80% passed the interview” is ambiguous unless readers know whether the second percentage is of all applicants or only interviewees.

59. Eligible population and observed population can differ | eligibility denominator

A participation rate might use all eligible people as denominator, while a completion rate uses only participants. Using the wrong denominator can produce a numerically neat but conceptually wrong metric. Institutional definitions should be checked rather than guessed.

60. Missing data can change the denominator | complete-case denominator

If 100 students enrol but only 80 have complete data, a reported percentage based on 80 may differ from one based on all 100. “Among students with complete data” is not a minor caveat—it identifies the denominator. Transparent quantitative English includes exclusions.

61. Population, sample and subgroup denominators are distinct | population vs sample

“30% of the sample” does not automatically mean “30% of the population.” Sampling and weighting determine whether generalisation is justified. Lesson No.033 owns evidence claims; this lesson protects the quantitative noun so a sample proportion is not silently rewritten as a population percentage.

62. Rate denominator must match opportunity | opportunity denominator

A defect rate per product, per batch and per inspection can differ because each denominator represents a different opportunity for defects to be observed. Choose the rate that matches the decision question. “Which factory produces more defects?” may need volume-adjusted rates, not raw counts.

63. Incidence and prevalence are specialised denominator systems | preview for Lesson 040

In health and epidemiology, incidence focuses on occurrence of new cases over a period, while prevalence concerns how common a condition is within a population at a point or period. These terms need domain-specific definitions and will be treated carefully in Lesson No.040 rather than folded into generic “rate” language.

64. Ratio can be dimensionless or have units | dimension analysis

A ratio of two quantities with the same unit may be dimensionless. A rate often retains units such as kilometres per hour. This technical distinction is useful but not always visible in everyday language. The practical learner question remains: what does the denominator mean?

65. Index is a transformed relative measure | index

An index often sets a baseline to a reference value such as 100 and expresses later values relative to it. “Index 120” does not mean 120 units of the original quantity. Lesson No.036 touched trend indices; here the owner is quantitative relation: an index is a derived relative measure with a defined base.

66. Normalisation makes unlike-sized systems more comparable | normalize

Per-capita, per-user and per-unit measures normalise totals by size or exposure. Normalisation can improve comparability but can also hide important totals. A small country may have high per-capita emissions but low total emissions. Good writing reports the metric suited to the question, sometimes both.

67. Standardisation is broader and technical | standardize

Standardisation may mean adjusting values to a common scale, applying standard definitions or using statistical procedures. It is not a decorative synonym for normalisation. Technical fields define it differently; use the term only when the method is known.

68. “Share of total” is often clearer than vague percentage | explicit whole

“The programme received 18% of total spending” is clearer than “the programme received 18%” when several denominators are possible. Explicitly naming the whole improves both comprehension and auditability.

69. Quantitative nouns should survive paraphrase | paraphrase boundary

If the source says “ratio,” do not automatically paraphrase “percentage.” If it says “rate,” preserve the denominator/time architecture. If it says “share,” preserve the whole. Paraphrase can change wording but must not change the mathematical relationship.

70. Part II checkpoint: the denominator family decides the noun | 第二部分检查点

Part-to-whole suggests proportion, percentage, fraction or share. Part-to-part strongly suggests ratio. Per-time/per-population/per-exposure relations suggest rates. Raw quantities suggest count, number or amount. Normalised quantities require the unit to stay visible. The first repair for a confusing sentence is often not “find a better synonym” but “recover the denominator.”

Part III — Compare quantitative relationships without changing the question | 第三部分：比较数量关系，但不要偷偷换问题

71. Absolute change keeps the original unit | absolute change

If enrolment rises from 100 to 120, the absolute change is +20 students. Absolute change is easy to understand because it preserves the original unit. It answers “how many more?” or “how much more?” It does not answer how large the change is relative to the starting level.

72. Relative change compares the change with the baseline | relative change

The same increase from 100 to 120 is a 20% relative increase. If another school rises from 1,000 to 1,020, both schools gain 20 students, but the relative increase is only 2% in the larger school. Absolute and relative changes answer different questions and can reverse which change appears larger.

73. Percentage-point change compares percentages directly | percentage points

If pass rate rises from 60% to 70%, the increase is ten percentage points. Relative to the original 60%, that is about a 16.7% increase. “Pass rate rose 10%” is ambiguous and often interpreted as a relative increase. Write ten percentage points when that is what the data show.

74. Base effect can make identical absolute changes look different | base effect

A +5 increase from 5 to 10 is a 100% increase; +5 from 100 to 105 is 5%. Relative language amplifies changes from small bases. This is not mathematically wrong, but the reader needs the baseline to judge practical significance.

75. Small denominator problems create unstable percentages | small denominator

If one event becomes two events, incidence doubles by 100%, but the absolute change is one event. In small groups, percentages can swing dramatically because each case represents a large share. Advanced reporting should often include the underlying count.

76. Large denominator can hide substantial counts | large population

A tiny percentage of a huge population can still represent many people. “Only 0.1%” sounds small, but 0.1% of ten million is ten thousand. Quantitative English should not use percentage adjectives to imply practical triviality without considering the underlying count.

77. Share can fall while count rises | denominator expansion

If one category grows more slowly than the total, its share falls despite absolute growth. This pattern appears in market share, budget share, population composition and vote share. A precise sentence may need both: “The number increased, but its share of the total declined.”

78. Share can rise while count falls | denominator contraction

If a whole shrinks faster than one category, the category’s share can rise even while its count falls. “Market share improved” does not necessarily mean the company sold more units. The denominator changed faster.

79. Ratio can improve while both components worsen | ratio dynamics

Suppose debt falls from 100 to 90 while income falls from 50 to 40. Debt-to-income ratio actually worsens from 2.0 to 2.25 despite lower debt. Ratio movement depends on both numerator and denominator. A trend in a ratio should not be explained by one component without checking the other.

80. Rate can rise because numerator rises | numerator-driven change

A defect rate can increase because defects increase while production stays constant. This is the intuitive case. But rates can also rise because the denominator falls, even if the numerator does not increase. Always inspect both components.

81. Rate can rise because denominator falls | denominator-driven change

If five errors occur in 1,000 transactions, the rate is 0.5%. If transactions fall to 500 while errors remain five, the rate doubles to 1%. The event count is unchanged. A sentence such as “errors doubled” would be false; only the rate doubled.

82. Rate can fall while event count rises | exposure growth

If errors rise from 5 to 6 while transactions rise from 1,000 to 2,000, the error count increases but the rate falls from 0.5% to 0.3%. This distinction matters in safety, operations and public reporting. Counts and rates can move in opposite directions.

83. Standardised rate may move differently from crude rate | standardisation

In some health, demographic and policy contexts, crude rates are adjusted or standardised to account for population composition. The technical method belongs to the domain. Linguistically, never treat a standardised rate as if it were the raw rate; retain the adjective because it tells readers the denominator/composition has been transformed.

84. Weighted percentage may not equal simple percentage | weighting

Survey results may be weighted to represent a target population. A weighted percentage gives observations unequal influence. If you compare weighted and unweighted results, name which one you are using. “30% of respondents” can be misleading if 30% is actually a weighted estimate rather than raw respondent share.

85. Weighted average is not average of subgroup averages unless weights match sizes | aggregation

If Class A has 10 students averaging 90 and Class B has 100 students averaging 70, the overall mean is not simply (90 + 70)/2. Group sizes matter. Lesson No.038 will own central tendency, but this example reinforces the denominator rule: aggregation depends on how much each group contributes.

86. Equal weighting is a decision, not a neutral default | equal weights

When a composite score gives every component 25%, that is an explicit modelling choice. If the weights change, the overall index may change even when component values do not. Quantitative English should expose weighting where it affects interpretation.

87. Composite score combines multiple quantities | composite metric

A composite score blends several measurements, often after scaling and weighting. It is not a raw count, proportion or rate. “The index rose” may reflect changes in one or more components. Do not explain a composite metric without knowing its construction.

88. Index number is relative to a base period | index base

If 2020 = 100 and the index is 120 in 2026, the indexed quantity is 20% above the base under the index definition. It does not mean 120 physical units. Changing the base year can change index numbers while preserving relative movement.

89. Share-of-change is different from share-of-level | contribution to growth

A sector may account for 10% of the economy but contribute 30% of annual growth if it expands rapidly. “Share of GDP” and “contribution to GDP growth” are different metrics. The first concerns level composition; the second concerns contribution to change.

90. Contribution percentage can have a different denominator | contribution

“X contributed 40% of total growth” divides X’s contribution by total growth, not by total output. If total growth is small or negative, contribution percentages can behave unexpectedly. Always identify what the percentage is a percentage of.

91. Ratio-to-total conversion changes denominator | ratio ↔ proportion

A ratio of 3:2 means the first group is three parts and the second two parts. The first group’s proportion of the combined total is 3/(3+2)=60%. This conversion is useful because learners often confuse 3:2 with 150%. The ratio compares parts; the proportion uses the sum as denominator.

92. “One in five” is a natural-language proportion | one in N

“One in five students” means 20% if the phrase is exact. News and public communication use one in N because it is intuitive. Do not convert approximate phrases such as “about one in five” into falsely precise percentages without preserving approximation.

93. “X out of Y” keeps the raw denominator visible | out of

“18 out of 50 participants” is often more transparent than “36%” when the sample is small. Counts show the real evidence base. Academic and public-health writing frequently benefit from reporting both count and percentage.

94. Percentage without count can hide sample size | n matters

“50% improved” could mean one of two people or five hundred of one thousand. Statistical confidence and generalisability differ dramatically. A percentage tells the proportion, not the evidence volume.

95. Count without rate can hide exposure | exposure matters

“Factory A had twice as many defects” may sound worse until you learn it produced ten times as many units. Counts need exposure denominators for fair comparison when opportunity differs.

96. Per-capita without total can hide system burden | total matters too

Per-capita numbers improve comparability across population size, but budgets, emissions or infrastructure still respond to totals. Sometimes the most honest account reports both: “Per-capita use is lower, but total use is higher because the population is larger.”

97. Rate ratio compares two rates | rate ratio

Some technical fields compare rates by division, producing a rate ratio. This is specialist statistical language. The resulting number is a ratio of two normalised metrics, not itself the original event rate. Follow the field’s formal interpretation.

98. Risk ratio and odds ratio are technical | technical comparisons

Health and statistics use terms such as risk ratio and odds ratio. They are not interchangeable and should not be interpreted by dictionary intuition alone. General learners should recognise them as formal comparative measures and consult domain definitions rather than paraphrasing them as “percentage difference.”

99. Rate per time and probability over time are different | hazard/probability caution

A rate describes occurrence relative to exposure/time; a probability describes chance over a defined horizon. In some models they are mathematically related, but not identical. Lesson No.035 owns probability; Lesson No.037 protects the rate noun from casual substitution.

100. Percentage can describe composition or change | two percentage jobs

“30% of revenue came from subscriptions” is compositional. “Revenue increased 30%” is change relative to a baseline. The same symbol % appears, but the denominator is conceptually different. In the first, total revenue is the whole. In the second, prior revenue is the base.

101. Percentage can describe probability too | probability format

“There is a 30% probability of delay” is neither composition nor growth. It expresses probability on a 0–100% scale. The numerical format is the same; the conceptual quantity is different. The noun following the percentage determines the relationship.

102. Percentage can describe confidence only informally | certainty framing

“I’m 90% sure” is common informal English but usually expresses subjective confidence rather than a formally calibrated probability. In analytical writing, distinguish confidence language from measured event probability.

103. Ratio can encode efficiency | output/input

Efficiency metrics often compare useful output with input. A ratio may improve because output rises, input falls, or both. “Efficiency improved” requires knowing the ratio definition and whether higher or lower is better. Not all ratios have the same desirable direction.

104. Lower ratio can be better or worse depending metric | direction of desirability

A lower student–teacher ratio may be desirable because fewer students share each teacher. A lower revenue-to-cost ratio may be undesirable. Do not attach “improvement” to ratio movement until you know the metric’s meaning.

105. “High rate” can mean common, fast or expensive depending noun | polysemy

High error rate = many errors relative to opportunity. High growth rate = rapid proportional change. High interest rate = cost/return percentage per period. High speed rate is usually unnatural because speed already is a rate. Learn collocations, not just the abstract definition.

106. “Rate of” and “ratio of” have different preposition patterns | grammar

We say rate of growth, rate of infection, ratio of students to teachers, proportion of students who passed, percentage of revenue from subscriptions, share of total spending. These constructions encode the metric’s logic. Store them as sentence frames.

107. “As a proportion of” explicitly supplies the denominator | as a proportion of

“Public spending as a proportion of GDP” tells the reader the denominator is GDP. This structure is especially useful when comparing across countries or years because it distinguishes absolute spending from spending relative to economic size.

108. “As a percentage of” is often clearer for non-specialists | as a percentage of

“Rent as a percentage of income” directly communicates part-to-whole/relative burden. “Rent-to-income ratio” may be more compact in technical contexts. The relationship can be similar while register differs.

109. “Share of” is useful for composition narratives | share of

“Renewables’ share of electricity generation rose.” This is more natural than “the percentage ratio of renewables rose.” Good quantitative English often uses conventional domain phrases rather than mathematical-sounding noun stacks.

110. “Rate per” should include the denominator scale | rate per

“The incident rate was 12 per 100,000 employee-hours” is complete. “The incident rate was 12” is not, unless the unit is established in the table/figure. Units can be omitted locally only when the context already owns them clearly.

111. Percentage-point language should survive headlines | headline compression

If a policy approval rate moves from 40% to 45%, “approval up 5 points” may be acceptable headline shorthand if the context makes percentage points clear. “Up 5%” would mean 42% under a relative calculation. Compression cannot change the metric.

112. “Double” can refer to count, rate, ratio or probability | double needs an object

“Cases doubled,” “the rate doubled,” “the odds doubled,” and “market share doubled” are different propositions. Never paraphrase one as another. If the metric is complex, repeat the noun rather than relying on vague “it doubled.”

113. “Half” can refer to level or reduction | half vs halve

“Half the students passed” describes a proportion. “The number of students halved” describes a 50% reduction in count. “The pass rate halved from 40% to 20%” describes a 50% relative reduction and a 20-percentage-point fall. These are different structures built around the same everyday word.

114. “One-third less” and “one-third as much” differ | fraction comparisons

If A is one-third less than B, A is two-thirds of B. If A is one-third as much as B, A is one-third of B. Natural-language fractional comparisons can be surprisingly error-prone; when stakes are high, numerical values are clearer.

115. Part III checkpoint: never compare metrics before naming their bases | 第三部分检查点

When two quantitative claims conflict, check whether they use the same numerator, denominator, unit, population, time period, weighting and baseline. Count can rise while rate falls. Share can fall while count rises. Ratio can worsen while one component improves. Percentage-point and relative percentage changes can both be correct. The apparent contradiction often disappears once the metric architecture is made explicit.

Part IV — Use quantity language across real domains | 第四部分：把数量、比例与比率词汇迁移到真实领域

116. Education: pass rate and average score answer different questions | 教育数据

A pass rate is the share of students meeting a threshold; an average score summarises the centre of the score distribution. A class can have a 90% pass rate with an average of 61 if most students cluster just above the pass mark. Another class can have a higher average but lower pass rate if performance is polarised. Good reporting names the metric rather than saying simply “results were better.”

117. Education: student–teacher ratio is not class size | ratio vs average

A school may have a student–teacher ratio of 15:1 but average class size of 25 because teachers also perform non-classroom roles, specialist teaching or support. The ratio is an institution-wide relationship; class size is a classroom-level count/average. Using one to imply the other can mislead parents and policymakers.

118. Education: attendance rate needs a defined opportunity denominator | attendance

Attendance rate may mean attended sessions divided by possible sessions, but institutional definitions vary. “95% attendance” is not automatically “the student attended 95 days.” If the reporting period contains 200 sessions, the count and percentage differ. Preserve the institution’s denominator definition.

119. Business: revenue share differs from unit share | market metrics

A premium brand can hold 20% of unit sales but 35% of market revenue if its products cost more. Saying “market share is 35%” without defining revenue versus units can create confusion. The noun phrase should reveal what is in numerator and denominator.

120. Business: gross margin is a percentage relationship, not gross profit | margin

Gross profit is an amount; gross margin usually expresses gross profit relative to revenue, often as a percentage. A company can increase gross profit while margin falls if revenue grows faster. Absolute amount and proportional margin can move differently.

121. Business: conversion rate depends on funnel stage | funnel denominator

Purchase conversion may be purchases divided by all website visits, unique visitors, leads, qualified leads or checkout starts. All can be called conversion rates in different teams. A number without denominator definition is not portable across dashboards.

122. Business: retention rate and churn rate are related but not always simple complements | retention/churn

In simple one-period models, retention and churn may sum to 100%. In real products, definitions can differ by cohort, activity threshold, reactivation and time window. Never assume complementarity without reading the metric specification.

123. Surveys: response rate is not support rate | survey metrics

A survey response rate concerns how many invited/eligible people responded. Support rate concerns how many respondents selected a position. “60% support” among a 20% response rate does not mean 60% of the invited population supports the position. Two different denominators sit inside the same study.

124. Surveys: percentages can sum above 100% when multiple responses are allowed | multi-select

If respondents can select several reasons, each reason’s percentage may use respondents as denominator. The percentages can sum above 100% because categories overlap. The phrase “respondents could select more than one option” protects interpretation.

125. Surveys: weighted percentage is a population estimate | survey weighting

A weighted survey percentage may estimate population composition after adjusting unequal sampling or response. It should not be described casually as “X% of respondents” unless that is literally the raw respondent fraction. Better: “The weighted estimate was X%.”

126. Public policy: budget share differs from spending per capita | policy denominators

Education spending can rise as a share of total government expenditure while falling per student, or vice versa, depending population and budget growth. Policy arguments often appear contradictory because they choose different denominators. A careful writer names both when relevant.

127. Public policy: tax rate is not tax amount | tax

A person can pay more tax even when the tax rate falls if taxable income rises enough. Conversely, tax amount can fall while the statutory rate stays unchanged. Rate, base and amount are separate components.

128. Public policy: coverage rate differs from service quality | coverage

Coverage tells us what share of a target population receives or can access a service. It does not tell us how good the service is. “Coverage reached 95%” should not be paraphrased “the service is highly effective.” Quantity and quality are different dimensions.

129. Operations: defect count and defect rate can move opposite ways | operations

A factory that doubles output may record more defects but lower defects per thousand units. Management decisions may need both count and rate: total rework burden depends on count, process quality depends more directly on the normalised rate.

130. Operations: utilisation rate needs a capacity denominator | utilisation

Utilisation commonly relates actual use to available capacity, but capacity can be defined theoretically, practically or scheduled. “90% utilisation” means little without knowing which capacity baseline is used. Different denominators can make the same operation appear more or less loaded.

131. Operations: throughput is a rate-like flow quantity | throughput

Throughput often means units processed per unit time. It is not the same as utilisation or efficiency. A system can have high utilisation but low throughput if bottlenecks or rework consume capacity. Technical terms should remain in their own metric architecture.

132. Finance: return can be amount or rate | return

Investment return may mean a monetary gain or a rate of return expressed relative to invested capital. “The investment returned $10,000” and “the return was 10%” are different forms. Financial writing should specify absolute return, percentage return, annualised return or another conventional metric.

133. Finance: yield and rate are domain-specific | yield

Yield can represent income relative to price or other domain-specific relationships. It is not a sophisticated synonym for percentage. Technical finance uses exact formulas; learners should follow the field rather than infer meaning from the surface number.

134. Finance: leverage ratios need precise numerator/denominator definitions | leverage

Debt-to-equity, debt-to-assets and other leverage ratios answer different questions. Saying “the leverage ratio is 2” without naming the ratio can be ambiguous. A ratio label is part of the measurement, not optional metadata.

135. Science: concentration is amount relative to volume or another base | concentration

Concentration relates a substance amount to a volume/mass or another defined base. Units such as mg/L carry the denominator. The everyday word “amount” cannot replace concentration without losing normalisation.

136. Science: density and concentration are different | density vs concentration

Density often relates mass to volume; concentration relates the amount of a component to a mixture/solution base. Technical definitions vary by field. Both are normalised quantities, but the numerator and denominator differ.

137. Scientific reporting: molar ratio / mixing ratio are ratios, not percentages automatically | technical ratios

Scientific disciplines use specialised ratios and fractions. The numerical value may be converted to percentage, but the original noun encodes a domain relationship. Do not paraphrase technical ratio names unless you understand the transformation.

138. Computing: error rate and latency are different metric families | technical systems

Error rate is typically a proportion/rate of failed operations; latency is a time duration per event; throughput is events per time. Saying “system performance improved by 20%” hides which metric improved and whether higher/lower is desirable.

139. Computing: cache hit rate is a proportion-like rate | hit rate

Cache hit rate often means hits divided by total requests/queries under a system definition. A higher hit rate can coexist with slower overall performance if other bottlenecks worsen. The metric name does not contain the entire system story.

140. Machine learning: accuracy is a proportion, but class imbalance matters | accuracy

Accuracy often represents correct predictions divided by all predictions. A high percentage can be misleading when one class dominates. Precision, recall and other metrics use different denominators. Learners should treat each as a defined metric, not as synonyms for “percentage correct.”

141. Language learning: vocabulary coverage is a proportion | coverage

Vocabulary coverage can describe what proportion of running words in a text are known/covered by a word list or learner. It is not the same as vocabulary size. A learner might know many words but still have low coverage in a specialised text because the denominator is the text’s token distribution.

142. Language learning: error rate differs from error count | error metrics

A longer essay can contain more errors but a lower error rate per 100 words. If the writing task doubles in length, raw errors alone may not fairly represent accuracy. Normalisation should match the pedagogical question.

143. Language learning: lexical diversity metrics are not simple percentages | specialised metrics

Measures of lexical diversity use different formulas and denominator structures. Do not reduce them to “percentage of advanced words.” Lesson No.026 already separated lexical density from sophistication; this lesson adds the quantitative caution: metric labels must be interpreted by definition.

144. Journalism: “X% more” should reveal the base | reporting ethics

A headline that says “risk rises 200%” may describe a movement from 1 in 10,000 to 3 in 10,000. The relative increase is large; the absolute risk remains small. Good public communication includes baseline where omission would distort perception.

145. Journalism: “one in three” can be clearer than 33.3% | audience design

Different numerical formats communicate differently. “One in three” may be easier for general audiences; 33.3% may look more precise. The format should match source precision and audience needs, not merely aesthetic preference.

146. Journalism: raw count matters for rare events | counts plus rates

For rare events, report both: “The rate rose from 1 to 2 per 100,000, corresponding to 5 additional cases.” This prevents a relative doubling from sounding like a mass event while still acknowledging the change.

147. Comparative reporting: same denominator is required | apples-to-apples

Country A’s cases per 100,000 cannot be directly compared with Country B’s raw count without converting to a common basis. “Apples-to-apples comparison” is informal business language for using equivalent denominators, periods and definitions.

148. Cross-country rates may still differ in definition | harmonisation

Even when both sources report a rate per 100,000, case definitions, age structure, reporting systems or time periods can differ. A common numerical format does not guarantee conceptual comparability.

149. Cross-time comparisons can fail when denominator definition changes | series breaks

If an organisation redefines “active user,” a later conversion or retention rate may not be directly comparable with earlier years. Quantitative vocabulary should carry caveats about metric-definition changes.

150. Targets can be count targets or rate targets | target architecture

“Reduce complaints to 100 per month” is a count target. “Reduce complaints to fewer than 2 per 1,000 transactions” is a rate target. If transaction volume grows, the two targets may imply different operational priorities.

151. Threshold can apply to any metric family | threshold + metric

Thresholds may be defined for counts, percentages, rates or ratios: “more than 50 cases,” “above 10%,” “over 5 per 1,000,” “ratio greater than 2.” Lesson No.034 owns thresholds; Lesson No.037 ensures the threshold’s metric is correctly named.

152. Benchmark can be a reference value, rate or ratio | benchmark

A benchmark is a comparison reference. Industry average conversion rate, target cost per student or standard debt ratio can all be benchmarks. A benchmark does not tell you whether a metric is count/rate/ratio; it tells you what value you compare against.

153. Target, benchmark and actual are different roles | roles

Actual pass rate may be 72%, target 80%, benchmark 75%. Mixing these roles can make a report unreadable. Quantitative nouns should be paired with role labels when multiple reference values appear.

154. Estimate and observed value are different | estimate

A survey percentage may be an estimate of a population proportion; a census percentage may be directly calculated from a near-complete count. “Estimated proportion” and “observed sample proportion” should not be collapsed when uncertainty matters.

155. Confidence interval belongs to the estimate, not the numerator itself | estimate uncertainty

When a proportion/rate is estimated from a sample, uncertainty surrounds the estimate. Technical interpretation depends on statistical framework. The linguistic lesson is to keep estimate, interval, sample, population roles clear rather than speaking as if the sample percentage were a perfectly known population fact.

156. Rounding can change displayed percentages | rounding

Percentages rounded independently may sum to 99% or 101%. That does not necessarily indicate an error. “Figures may not sum to 100% because of rounding” is standard reporting language.

157. Suppression and small-cell rules change what can be shown | reporting constraints

Some datasets suppress small counts for privacy or reliability. A missing percentage may therefore mean “not reported,” not zero. Quantitative vocabulary should distinguish zero, unavailable, not applicable and suppressed values.

158. Zero is not missing | zero vs missing

0 cases means the recorded count is zero under the definition. Missing data means no value is available. A dash in a table might represent zero, missing or not applicable depending on legend. Never infer one from formatting alone.

159. Not applicable is not zero | N/A

A ratio may be undefined or not applicable when the denominator is zero or the metric does not make conceptual sense. Writing “0%” can create a false quantity where no valid calculation exists.

160. Part IV checkpoint: metric names are compressed methods | 第四部分检查点

In real domains, rate, share, ratio, margin, yield, coverage, accuracy, utilisation are compressed descriptions of how a number is constructed. Before interpreting the value, unpack the method: numerator, denominator, unit, period, eligibility, weighting and adjustment. A metric name is not decoration; it is a miniature methodology.

Part V — Mandarin-to-English quantitative control | 第五部分：中文母语学习者的数量与比例词汇转换

Mandarin quantitative vocabulary is flexible. 比例, 比率, 占比, 率 and 百分比 can overlap in ordinary explanation, while English reporting often prefers conventional metric names. The safest route is relationship-first translation: identify count, part-to-whole, part-to-part, per-unit, per-time, share-of-total or relative change before choosing the English noun.

161. 数量 can be quantity, number or amount | 数量不是 always quantity

“学生数量” is usually number of students. “用水量” may be amount/volume of water used. “产品数量” can be number/quantity of products. Quantity is broader/formal; number is natural for countable units; amount suits uncountable measures.

162. 数目 normally maps to number / count | 数目 = countable quantity

“申请人数数目” is awkward in English if translated literally. Use number of applicants or applicant count in data labels. Count sounds more technical/data-oriented; number is broad.

163. 金额 = amount / sum / value, depending context | 金额

“付款金额” → amount paid/payment amount. “总金额” → total amount/sum. “交易价值” may be transaction value. Do not use number merely because the figure is numeric.

164. 比例 often maps to proportion, ratio or percentage | 比例是高风险翻译点

“女生占全班的比例” → proportion/percentage of girls in the class. “男女比例 3:2” → male-to-female ratio of 3:2. “比例提高到 30%” may be the proportion/share rose to 30%. Identify whether the comparison is part-to-whole or part-to-part.

165. 比率 can be ratio or rate | 比率需要看分母

“师生比率” → student–teacher ratio. “失业率” → unemployment rate. “错误比率” could be error rate if failures are divided by opportunities. Chinese 比率 does not tell you whether English prefers ratio or rate; the denominator architecture does.

166. 百分比 = percentage, but the metric may still be a rate/share | percentage is format

“失业率的百分比” is often redundant because unemployment rate is already expressed as a percentage. “市场占比为 20%” → market share was 20%, not “market percentage was 20%.” English often uses the metric noun plus a percentage value.

167. 百分点 = percentage point | 百分点 ≠ percent

“从 40% 上升到 45%，增加 5 个百分点” → rose by five percentage points. “增加 5%” would mean 42% if interpreted as a relative increase from 40. This distinction should be automatic in advanced quantitative English.

168. 占比 = share / proportion / percentage of total | 占比强调在整体中的位置

“线上销售占比 40%” → online sales accounted for 40% of total sales / online sales’ share was 40%. The whole should be explicit if ambiguity exists. Ratio is usually not the default because 占比 typically means part-to-whole.

169. 份额 = share, often market/budget/vote | share

“市场份额” → market share. “预算份额” → share of the budget. “投票份额” → vote share. Share is a strong collocational owner; replacing it with percentage or ratio can sound unnatural even when the number is convertible.

170. 率 often maps to rate, but definition still matters | 率 = rate often

通过率 = pass rate; 转化率 = conversion rate; 失业率 = unemployment rate; 出生率 = birth rate. English conventional names often match. But the formula and denominator differ across metrics, so shared suffix “rate” does not imply identical mathematical construction.

171. 人均 = per capita / per person | 人均

“人均收入” → income per capita / per-capita income. In everyday contexts, per person may be clearer. “人均每户” would mix denominators and needs careful reconstruction. Name the unit exactly.

172. 户均 = per household / average per household | 户均

“户均用电量” → electricity use per household / average household electricity use. The first foregrounds normalisation; the second may imply an average. They can coincide numerically but package the metric differently.

173. 每单位 = per unit | 每单位

“每单位成本” → cost per unit/unit cost. “每用户收入” → revenue per user. “每小时请求数” → requests per hour. The preposition per makes the denominator explicit and should not disappear in translation.

174. 总量 = total / aggregate amount | total

“总排放量” → total emissions. “总销售额” → total sales/revenue. Aggregate is more technical and may mean combined across units. Per-capita and total values should not be mixed.

175. 增量 = increment / increase / additional amount | 增量有多条路线

“新增用户” → new/additional users, not necessarily “incremental users.” “收入增量” in analytical contexts may be incremental revenue. Increment sounds technical and stepwise; ordinary increase is often safer.

176. 倍数 = multiple / times / fold | multiplicative language

“是原来的三倍” → three times the original level / three times as high. “增长三倍” is ambiguous even in Chinese: it may mean become three times as large or increase by three times. In English, avoid ambiguity by giving start/end values when possible.

177. 翻倍 = double | doubling

“人数翻倍” → the number doubled. “比率翻倍” → the rate doubled. Keep the metric noun because doubling a count is not the same as doubling a rate or share.

178. 减半 = halve / fall by half | halve

“错误率减半” → the error rate halved / fell by half. If it moves from 10% to 5%, that is a five-percentage-point fall and 50% relative reduction. Different descriptions can coexist.

179. 占 = account for / make up / constitute | part-to-whole verbs

“线上销售占总销售的 40%” → Online sales accounted for/made up 40% of total sales. Constituted is more formal. These verbs express part-to-whole relationship naturally and often avoid clumsy noun stacks.

180. 达到 X% can be reach / stand at / rise to | 达到的状态

“占比达到 40%” → the share reached 40%. “目前为 40%” → stands at 40%. “从 30% 升到 40%” → rose to 40%. The verb adds time/change information beyond the percentage itself.

181. 提高 X% may mean relative increase or percentage-point increase | 提高的歧义

Chinese business speech sometimes leaves the base unclear. Before translating, determine whether 40% → 44% (10% relative increase) or 40% → 50% (10 percentage-point increase) is intended. English should resolve the ambiguity, not reproduce it.

182. 下降 X% has the same ambiguity | relative fall vs points

“率下降 5%” can mean relative decrease by 5% or fall by five percentage points depending source. If the underlying values are available, translate explicitly.

183. 比重 often maps to share / proportion / weight | 比重

“服务业比重” → the service sector’s share/proportion. In a weighted index, 比重 may be weight. The context decides whether it is composition or modelling importance.

184. 权重 = weight | weight is not share automatically

In an index, a component can receive a 30% weight, meaning it contributes according to a calculation rule. This is not necessarily its real-world share. Weight is a modelling coefficient; share is a part of an observed/defined whole.

185. 构成 = composition / make-up | composition

“收入构成” → revenue composition / breakdown of revenue. “人员构成” → workforce composition/make-up. Breakdown often introduces category shares in tables/charts and is more accessible than composition.

186. 分布 ≠ proportion; distribution belongs to Lesson 039 | 分布

“收入分布” describes how values are arranged across people, not simply each category’s share. Lesson No.039 will own distribution, spread and concentration. This distinction prevents “distribution” from becoming a vague synonym for “percentage breakdown.”

187. 频率 ≠ rate automatically | frequency preview

“发生频率” describes how often something occurs. It may be expressed as events per unit time, but English frequency and rate are not universally interchangeable. Lesson No.040 will own frequency, prevalence and incidence.

188. 比例失衡 can be imbalance / disproportion | imbalance

“男女比例失衡” may be gender imbalance / imbalanced sex ratio. “资源配置比例失衡” may be disproportionate allocation. English often prefers an evaluative noun/adjective rather than literally “ratio imbalance.”

189. 成比例 = proportional / in proportion to | proportionality

“A 与 B 成比例” → A is proportional to B. In technical mathematics, direct/inverse proportionality has precise meanings. Do not use proportional merely because both variables tend to rise together.

190. 不成比例 = disproportionate / out of proportion | disproportionate

“成本增长与收益不成比例” may be cost growth is disproportionate to the gains. Disproportionate often carries evaluative meaning: too large/small relative to something else. It is not simply “the ratio changed.”

191. 绝对值 / 相对值 = absolute / relative value | absolute vs relative

In general data communication, “absolute number/count” often contrasts with “relative percentage/rate.” In mathematics, absolute value is a specific technical concept. Avoid using the mathematical phrase when you merely mean raw total.

192. 原始值 = raw value | raw

“原始数据/原始值” → raw data/raw values. Normalised, adjusted or indexed values are derived transformations. Carry the adjective so readers know which form they are seeing.

193. 标准化 = standardised / normalized, but not always same | 标准化

Statistical, laboratory and data-science contexts use standardise and normalise differently. Do not translate 标准化 with whichever word sounds more technical. Follow the method’s own terminology.

194. 基数 = base / base value / denominator, depending context | 基数

“基数较小导致百分比增长很大” → The percentage increase appears large because the base value is small. In population statistics, 基数 can refer to baseline population; in ratio explanation, denominator may be the more exact term.

195. 分母效应 = denominator effect | denominator effect

When a ratio/rate changes largely because the denominator changes, denominator effect is understandable analytical language. The exact term depends on field, but the explanation should name the denominator rather than attributing the whole movement to the numerator.

196. 基期 = base period | base period

“以 2020 年为基期” → using 2020 as the base period/year. Index numbers and relative changes depend on the chosen base. Baseline may be better when the comparison is experimental or pre-intervention rather than an index base.

197. 同比 / 环比 require explicit comparison language | year-on-year / period-on-period

“同比增长” commonly maps to year-on-year growth when comparing the same period with a year earlier. “环比” may be month-on-month / quarter-on-quarter / period-on-period depending frequency. Do not translate both as generic “compared with last period.”

198. 占总量的 X% = account for / make up X% of total | explicit denominator

This is one of the safest Mandarin-to-English patterns because the denominator is explicit. “A 占总支出的 30%” → A accounted for 30% of total spending.

199. 每百 / 每千 / 每十万 = per hundred / thousand / 100,000 | scaling

Keep the scale exactly. “每十万人 50 例” → 50 cases per 100,000 people. This is not 50%. The denominator scale is part of the metric.

200. 每年每十万人 = per 100,000 people per year | two denominators/time

Rates can contain both population and time. Translate all denominator components. Dropping the time period can change an incidence rate into a cumulative count-like statement.

201. 样本占比 ≠ population share automatically | sample proportion

“样本中 40%” → 40% of the sample/respondents. To write “40% of the population,” the study must justify generalisation. Sample percentage and population proportion are related through inference, not translation.

202. 加权占比 = weighted share/percentage | weighted

“加权后占比为 35%” → the weighted percentage/share was 35%. Do not write “35% of respondents” if weighting means the value no longer equals the raw respondent count proportion.

203. 贡献率 may be contribution share/rate, depending formula | 贡献率

Chinese economic/business terms such as 贡献率 can have specific formulas. English may use contribution to growth, contribution share, contribution rate depending source. Do not invent a literal translation without verifying the formula.

204. Part V bilingual audit | 双语审计

• Is the Chinese term about count, amount, part-to-whole or part-to-part?
• Does English have a conventional metric name?
• Is the denominator population, time, exposure, revenue, total or another unit?
• Is the number a level, share, change or probability?
• Does a percentage-point distinction need to be made explicit?
• Is the value raw, weighted, adjusted or indexed?

205. Part V checkpoint: translate the formula hidden inside the word | 第五部分检查点

When Mandarin gives you 比例, 比率, 占比, 率, 百分比, 比重 or 人均, do not begin with an English synonym list. Reconstruct the hidden formula: what is numerator, what is denominator, what unit, what time period, what whole? Then select the English metric that conventionally describes that relationship.

Part VI — Denominator failure laboratory | 第六部分：分母失误实验室

The fastest way to master quantitative vocabulary is to watch correct arithmetic produce a wrong sentence. In each case below, the numbers may be valid, yet the noun, denominator, comparison base or time window changes the meaning. The aim is not merely to calculate correctly. It is to make the English sentence reveal exactly what was calculated.

206. Case 1 — count rises while share falls | 数量上升，占比下降

Department A grows from 40 to 50 employees while the company grows from 100 to 160. A’s headcount rises by 25%, but its workforce share falls from 40% to 31.25%. “Department A shrank” is false. “Department A’s share shrank” is correct. The noun share prevents a compositional decline from being mistaken for an absolute decline.

207. Case 2 — share rises while count falls | 数量下降，占比上升

A product sells 900 units this year instead of 1,000 last year, but the total market falls from 10,000 units to 7,000. Its unit sales decline, yet market share rises from 10% to about 12.9%. “The product grew” is too vague. Better: “Unit sales fell 10%, but market share increased because the total market contracted more sharply.”

208. Case 3 — pass rate rises while average falls | 教育指标方向相反

Last year, half the class scored 95 and half scored 45. This year, everyone scores 62. If the pass mark is 50, pass rate rises from 50% to 100% while the mean falls from 70 to 62. “Results improved” depends on which outcome matters. Quantitative vocabulary should state the metric before the evaluation.

209. Case 4 — student–teacher ratio is not class size | 师生比 ≠ 班级人数

A school has 600 students and 40 teachers, producing a student–teacher ratio of 15:1. Classes may still contain 25 students because teachers have planning time, specialist roles and non-classroom duties. “Classes average 15 students” cannot be inferred from the institutional ratio. The ratio’s denominator is teachers, not classes.

210. Case 5 — 40% to 50% is not “up 10%” | 百分点与相对百分比

A completion rate rises from 40% to 50%. The absolute change in the rate is ten percentage points. Relative to the initial rate, completion is 25% higher. Writing “completion rose 10%” is ambiguous and usually mathematically different. When percentages themselves change, decide whether you are reporting percentage points or relative percentage change.

211. Case 6 — one event becoming two can sound enormous | small-base amplification

A rare event occurs once last year and twice this year. The count doubles and the relative increase is 100%, but the absolute increase is one event. A headline saying “events surge 100%” is numerically defensible but may be practically misleading. Better reporting pairs relative and absolute change when the base is tiny.

212. Case 7 — 0.1% can still mean many people | small percentage, large population

If 0.1% of a ten-million-person population is affected, the count is 10,000. “Only 0.1%” adds an evaluative cue that may minimise a substantial absolute burden. A proportion answers “what share?”; a count answers “how many?” Responsible language often reports both.

213. Case 8 — errors rise while error rate falls | count vs exposure

A service grows from 10,000 to 30,000 monthly transactions. Errors rise from 100 to 150. The error count increases 50%, but the error rate falls from 1% to 0.5%. “Quality worsened because errors increased” ignores exposure. “Operational burden increased, while errors per transaction fell” preserves both relevant truths.

214. Case 9 — rate rises while count stays constant | denominator contraction

Five incidents occur in both periods, but operating hours fall from 10,000 to 5,000. Incidents per 10,000 hours double even though incident count does not change. “Incidents doubled” is false; “the incident rate doubled” is accurate. The denominator is doing the movement.

215. Case 10 — total emissions vs per-capita emissions | 总量与人均量

Country A emits more in total because it has a much larger population, while Country B emits more per person. “Country B emits more” is incomplete. The speaker must choose whether the decision concerns total atmospheric burden, individual-level normalisation or another denominator. Both metrics can be relevant without being interchangeable.

216. Case 11 — revenue share vs unit share | 收入份额与销量份额

A premium product accounts for 10% of units but 25% of revenue. Both are market shares under different numerators. “It has 25% market share” is underspecified. “It accounts for 25% of category revenue but 10% of units sold” identifies the metric behind each percentage.

217. Case 12 — response rate vs support among respondents | 两层分母

A survey invites 10,000 people, 2,000 respond, and 1,200 respondents support a proposal. Response rate is 20%; support among respondents is 60%; support among all invitees is 12% if non-response is treated simply as non-support—which is usually not a justified inference. The safe statement keeps the respondent denominator explicit.

218. Case 13 — weighted percentage vs raw respondent share | 加权比例与原始比例

In a sample, 45% of actual respondents support X. After survey weights adjust age and region, the estimated population support is 51%. Saying “51% of respondents support X” is factually wrong. The correct phrase is “The weighted estimate of support is 51%.” Weighting changes contribution to the estimate, not the raw respondent count.

219. Case 14 — 3:2 ratio is not 150% share | ratio-to-proportion conversion

If red to blue items are in a 3:2 ratio, red items are 1.5 times as numerous as blue. Red does not make up 150% of the combined group. Red’s proportion of the total is 3/(3+2)=60%. The denominator changes when a part-to-part ratio is converted to a part-to-whole percentage.

220. Case 15 — one in five and 20% can differ in tone | format and audience

“One in five” and 20% are mathematically equivalent when exact, but they feel different to readers. One-in-N language can be more concrete and memorable. A percentage can look more formal. Choice should match audience and source precision, while approximation must remain visible: “about one in five” should not become an exact 20.0%.

221. Case 16 — category percentages exceed 100% legitimately | overlapping categories

Respondents can choose multiple study strategies. 70% choose retrieval, 55% choose rereading and 40% choose discussion. The percentages sum to 165% because categories overlap. “The data are wrong because totals exceed 100%” is incorrect. The denominator is respondents for each item, and individuals can enter multiple numerators.

222. Case 17 — subgroup percentages can hide unequal group sizes | same percentage, different count

Ten percent of a 50-person group is five people; ten percent of a 5,000-person group is 500. Saying “both groups have the same problem level at 10%” may be valid proportionally but ignores absolute burden. Policy planning may care about both prevalence and count.

223. Case 18 — total rate can reverse subgroup pattern | aggregation warning

When subgroup composition differs strongly, an overall rate can move differently from each subgroup’s rate. This family of phenomena includes Simpson-like reversals. The technical statistics are beyond this vocabulary lesson; the linguistic repair is to ask whether an aggregate percentage hides subgroup composition before claiming that one group performs better overall.

224. Case 19 — average of rates needs weighting | combining denominators

Branch A has 10 transactions with a 90% success rate; Branch B has 1,000 transactions with a 50% rate. The overall success rate is not the simple mean of 90% and 50%. Rates must be recombined using their underlying denominators or appropriate weights. “Average rate” can be mathematically ambiguous without method.

225. Case 20 — ratio direction can reverse interpretation | ratio order

A student–teacher ratio of 20:1 means 20 students per teacher. The teacher–student ratio is 1:20. Saying “the ratio improved from 15 to 20” is impossible to evaluate until order and desired direction are known. In some contexts lower student–teacher ratios are preferred; the reciprocal ratio moves upward.

226. Case 21 — ratio can worsen even when numerator falls | two-component movement

Debt falls from 100 to 90, but income falls from 50 to 30. Debt-to-income ratio rises from 2 to 3. “Debt improved, so leverage improved” is not supported. A ratio is a relationship between two moving quantities; describing only one component cannot explain the ratio.

227. Case 22 — contribution to growth vs share of output | different wholes

A sector accounts for 15% of total output but 50% of this year’s increase in output. Saying “the sector represents half the economy” is wrong. Its share of the change is 50%; its share of the level is 15%. Both percentages use different denominators.

228. Case 23 — per-user revenue vs total revenue | normalisation trade-off

Total revenue rises because user count grows, while revenue per user declines. A company can be larger but monetise each user less. “Revenue performance improved” is incomplete. The two metrics answer different questions about scale and intensity.

229. Case 24 — utilisation rate vs throughput | capacity vs flow

A machine can be utilised 95% of the time but process fewer units per hour because rework consumes capacity. Utilisation is a share of capacity/time used; throughput is flow per time. High utilisation is not automatically high productivity.

230. Case 25 — accuracy vs precision/recall in classification | denominator changes by metric

A model can achieve 99% accuracy on a dataset where 99% of cases are negative by predicting “negative” every time. Accuracy uses all cases as denominator. Recall for the positive class uses actual positives. Precision uses predicted positives. The different denominators answer different performance questions.

231. Case 26 — vocabulary coverage vs vocabulary size | language-learning denominator

A learner may know 8,000 word families but still have incomplete coverage of a specialised legal text. Vocabulary size counts known lexical units; coverage asks what proportion of tokens/types in a target text are accounted for. “Large vocabulary” and “high coverage” are related but not identical.

232. Case 27 — essay error count vs error rate | text length matters

Essay A has 10 errors in 200 words; Essay B has 12 errors in 500 words. B has more errors but a lower error rate per 100 words. A tutor deciding correction workload may care about count; a tutor assessing accuracy may prefer the normalised rate. Metric choice should follow the learning job.

233. Case 28 — total cost vs cost per student | education finance

A programme’s total cost rises 20% because enrolment rises 40%, causing cost per student to fall. “The programme became more expensive” is true in total budget terms but false per student. The denominator determines whether the sentence describes scale or efficiency.

234. Case 29 — 80% completion among starters vs all enrolled | cohort denominator

If 1,000 enrol, 800 start, and 640 finish, completion among starters is 80% but completion among enrolled participants is 64%. Neither percentage is “the” completion rate until the cohort definition is specified. Course providers should state the denominator.

235. Case 30 — metric redefinition creates artificial improvement | definition change

A platform used to define “active user” as one session per month; it changes the definition to one session per week. Retention rates before and after may not be directly comparable. A numerical trend can be partly a classification change. Write “under the new definition” or note the series break.

236. Case 31 — denominator exclusion creates survivorship-looking success | exclusion

If a programme reports improvement only among students who completed all sessions, dropouts disappear from the denominator. “90% improved” may accurately describe completers while overstating programme-wide outcomes. A precise sentence says “Among completers, 90% improved.”

237. Case 32 — missing data can inflate or deflate proportions | missing denominator

Suppose 100 people are sampled, 80 answer a question and 40 answer “yes.” “40% said yes” uses all sampled people as denominator; “50% of respondents to the question said yes” uses 80. Which is appropriate depends on the reporting convention. Missingness must not be silently converted into a response category.

238. Case 33 — rounding produces 101% | display arithmetic

Three categories have precise shares 33.6%, 33.6% and 32.8%. Rounded to whole numbers they display as 34%, 34% and 33% = 101%. The correct note is “Percentages may not sum to 100% because of rounding,” not “one category must be wrong.”

239. Case 34 — N/A is not 0% | undefined metric

If a branch had no eligible cases, a completion rate may be not applicable because the denominator is zero. Reporting 0% suggests eligible cases existed but none completed. Undefined/not applicable and zero are different quantitative states.

240. Case 35 — suppression is not zero | privacy/reliability

A table may suppress cells with fewer than five cases. A dash then means “not reported” rather than zero. Summary writing must preserve the reporting code. Turning suppression into absence can create false claims.

241. Case 36 — rate target and count target produce different behaviour | management target

A service targets fewer than 100 complaints per month and fewer than two complaints per 1,000 transactions. If volume doubles, the count target becomes harder relative to scale while the rate target remains comparable. The organisation must decide whether it is controlling total burden or quality per transaction.

242. Case 37 — benchmark and actual use the same metric but different roles | benchmark roles

Actual conversion rate is 5%; target is 7%; industry benchmark is 6%. A sentence such as “conversion is 7%” may accidentally report a target as an observed value. Label actual, target, benchmark and forecast explicitly.

243. Case 38 — estimate vs observed sample proportion | inference

In a sample, 52% select option A. After weighting and modelling, the estimated population proportion is 49%. Both numbers have roles. “52% of the population chose A” improperly moves a sample observation into a population claim. Lesson No.033’s evidence stage and Lesson No.037’s denominator stage must work together.

244. Case 39 — percentage probability vs population percentage | same symbol, different quantity

“30% of users will churn” can mean a forecasted population share in a future period. “Each user has a 30% churn probability” describes individual probabilistic modelling. Aggregation may connect them under assumptions, but the sentences do not automatically mean the same thing.

245. Case 40 — “half as much” vs “50% less” | equivalent when defined clearly

If B is 100 and A is 50, A is half as much as B and 50% less than B. But “two times less” is ambiguous/nonstandard in precise English. Use clear multiplicative or percentage language rather than phrases whose arithmetic interpretation varies across speakers.

246. Denominator-failure diagnostic 1 — what is the whole? | 诊断一

If you cannot finish the phrase “X% of ___,” the percentage is not ready to report. Supply the whole: all students, respondents with valid answers, eligible applicants, total revenue, total transactions, total electricity generation. The missing noun after of is often the hidden source of ambiguity.

247. Denominator-failure diagnostic 2 — what is the opportunity? | 诊断二

For an event rate, ask what gives the event a chance to occur: people, device-hours, trips, transactions, patient-days, words written. A count divided by the wrong opportunity base can compare unlike exposure.

248. Denominator-failure diagnostic 3 — did the denominator change? | 诊断三

When a rate/share changes, decompose numerator and denominator. If numerator remains stable but denominator falls, the rising metric has a denominator-driven explanation. If both move, describe both before attributing cause.

249. Denominator-failure diagnostic 4 — are categories mutually exclusive? | 诊断四

If category percentages sum above 100%, ask whether respondents/items can belong to multiple categories. If yes, the sum is not expected to equal 100%. If categories are intended to partition the whole, an excess may indicate rounding or classification error.

250. Part VI checkpoint: a numerically correct sentence can still be conceptually wrong | 第六部分检查点

The most dangerous quantitative errors are often not arithmetic errors. They are relationship errors: correct numerator, wrong denominator; correct percentage, wrong noun; correct ratio, reversed order; correct relative change, missing base; correct rate, wrong population; correct sample statistic, overextended population claim. Advanced vocabulary makes the formula visible enough for readers to audit the sentence.

Part VII — FENCE, retrieval and mastery | 第七部分：FENCE、提取与真正掌握

251. Build the quantity FENCE | 数量关系 FENCE

Fence	Question	Language job
F0 Metric	What exactly is measured?	count, amount, share, rate, ratio
F1 Numerator	What is counted?	name the event/category/value
F2 Denominator	Out of what / per what?	whole, population, exposure, time
F3 Unit	What unit and scale?	students, dollars, per 1,000, per hour
F4 Form	How is it expressed?	count, %, ratio, per-unit rate
F5 Base	Compared with what starting value?	absolute vs relative change
F6 Adjustment	Raw, weighted, indexed or standardised?	preserve method label
F7 Comparison	Are definitions/time/populations equivalent?	apples-to-apples check
F8 Transfer	Can I paraphrase/translate without changing formula?	meaning-preserving output

252. Build a metric-control card | 指标控制卡

Field	Example
Metric name	completion rate
Numerator	students completing
Denominator	students who started
Unit	people
Time period	2026 cohort
Form	percentage
Raw/adjusted	raw
Exclusions	students who enrolled but never started
Comparison base	2025 cohort

253. Practice A — choose the metric noun | 练习 A

1. 30 students out of 120 students.
2. 30 students compared with 20 teachers.
3. 5 defects per 1,000 units.
4. Company revenue divided by total market revenue.
5. Total litres of water used.
6. Requests processed each second.
7. Income divided by population.
8. 40% now compared with 30% last year.

Choose among number, amount, proportion, percentage, ratio, rate, share, per-capita measure, then explain the denominator in one sentence.

254. Practice B — recover the hidden denominator | 练习 B

9. “Conversion is 8%.”
10. “Support is 55%.”
11. “The rate is 12.”
12. “Market share is 20%.”
13. “Incidents doubled.”
14. “One in four are affected.”

For each, write the question you would ask before interpreting the number: 8% of which stage? 55% of whom? 12 per what? 20% of what market measure? Doubled from what baseline? One in four of which population?

255. Practice C — percentage or percentage points? | 练习 C

15. 20% → 25%
16. 40% → 60%
17. 5% → 10%
18. 80% → 72%

For each, state the percentage-point change and the relative percentage change. Then write a sentence that cannot be misread.

256. Practice D — count vs rate | 练习 D

19. Errors 100 → 120; transactions 10,000 → 20,000.
20. Cases 5 → 5; exposure hours 1,000 → 500.
21. Complaints 50 → 60; customers 1,000 → 2,000.
22. Defects 20 → 15; products 100 → 50.

Write one sentence about the count and one about the rate. Do not use “improved” or “worsened” until you have decided which metric the decision should prioritise.

257. Practice E — share vs count | 练习 E

23. Category A: 40/100 → 50/160.
24. Product sales: 1,000/10,000 → 900/7,000.
25. Women employees: 30/60 → 40/100.
26. Online revenue: 20/100 → 30/200.

Describe both absolute count/amount and part-to-whole share. Identify cases where one rises while the other falls.

258. Practice F — ratio to proportion | 练习 F

27. A:B = 1:1
28. A:B = 2:1
29. A:B = 3:2
30. A:B = 4:1

Convert A’s share of the combined total into a percentage. Then explain why the ratio number itself cannot simply be read as that percentage.

259. Practice G — Mandarin repair | 练习 G

31. 线上销售占总收入的比例为 35%。
32. 师生比例为 15 比 1。
33. 错误率从 10% 降到 5%，下降了五个百分点。
34. 总排放上升，但人均排放下降。
35. 样本中 42% 支持该方案。
36. 加权后，估计总体支持率为 48%。
37. 每十万人的发生率为 12。
38. 市场份额上升，但销量下降。

260. Practice H — paraphrase without denominator drift | 练习 H

Original: “Among the 800 students who started the programme, 640 completed it, giving a completion rate of 80%.” Unsafe paraphrase: “80% of enrolled students completed the programme.” The denominator changed from starters to enrolled students. Write three alternative paraphrases that keep the denominator as starters.

261. Practice I — headline repair | 练习 I

1. “Risk doubles!” from 1 in 10,000 to 2 in 10,000.
2. “Pass rate up 10%!” from 60% to 70%.
3. “Market share grows!” while unit sales fall.
4. “Errors increase 50%!” while transactions triple.

Rewrite each headline into a neutral one-line quantitative statement that includes the relevant base or denominator.

262. Practice J — metric-definition audit | 练习 J

Choose a dashboard you use. For five metrics, write: formal name, numerator, denominator, time window, exclusions, weighting/adjustment, desired direction. If you cannot fill the card, you know the label but not the metric.

263. A 15-minute quantitative-vocabulary session | 15 分钟训练

Time	Task
0–3 min	Choose one percentage/rate from a real report.
3–6 min	Recover numerator and denominator.
6–9 min	Rewrite as count + proportion + plain-English share where possible.
9–12 min	Check absolute vs relative change.
12–15 min	Explain the metric aloud without notes.

264. A 30-minute denominator workshop | 30 分钟分母实验室

Time	Task
0–5 min	Collect five metrics from different domains.
5–10 min	Classify part-to-whole / part-to-part / per-unit / raw count.
10–15 min	Reverse numerator/denominator to see how interpretation changes.
15–20 min	Convert counts ↔ percentages where legitimate.
20–25 min	Write a misleading version, then repair it.
25–30 min	Translate one Mandarin metric paragraph into precise English.

265. Seven-day quantitative-language challenge | 七天训练

Day	Focus
1	number / amount / quantity / count
2	proportion / percentage / share
3	ratio / rate / per-unit measures
4	absolute vs relative vs percentage points
5	denominator change and composition effects
6	Mandarin false-equivalence repair
7	full FENCE audit on a real report

266. Thirty-day denominator-control challenge | 30 天系统升级

• Build 30 metric-control cards.
• Repair 20 percentage-vs-percentage-point errors.
• Find 20 examples where count and rate move differently.
• Find 10 examples where share and count move differently.
• Convert 15 ratios into part-to-whole proportions correctly.
• Translate 20 Chinese quantitative terms by formula, not word-for-word.
• Audit 10 dashboard metrics for denominator drift.
• Rewrite 10 sensational relative-change headlines with absolute baselines.
• Explain 10 metrics orally in under 60 seconds each.

267. Twelve-week progression | 12 周路线

Weeks	Focus	Evidence of mastery
1–2	count / amount / total / unit	correct countability and units
3–4	proportion / percentage / share	explicit whole/denominator
5–6	ratio / rate / per-unit	correct denominator family
7–8	absolute / relative / percentage points	no baseline ambiguity
9–10	weighting / normalisation / index	method labels preserved
11–12	cross-domain independent use	meaning survives paraphrase/translation

268. First weak-link diagnostic | 第一个弱点诊断

Symptom	Likely weak link	Repair
I call every % a rate.	metric-family confusion	part-to-whole vs per-unit drills
I mix ratio and proportion.	denominator architecture	part-to-part conversion
I say +10% when 40→50%.	percentage-point control	points vs relative practice
I compare raw counts across different exposure.	normalisation	per-unit rate practice
I say share fell, so count fell.	composition effect	numerator/denominator decomposition
I drop weighted/standardised labels.	method preservation	raw vs adjusted pairs
I translate 比例 as ratio every time.	bilingual mapping	formula-first translation
I cannot explain what “of” refers to.	hidden denominator	finish “X% of ___”

269. How a strong tutor teaches quantity vocabulary | 高水平导师怎样教？

A weak lesson teaches definitions: percentage = 百分比, ratio = 比率, rate = 率. A strong lesson gives the same numerator different denominators and asks learners to watch the metric change. Thirty students can be 30%, 60%, 3:2, 30 per 1,000 or simply a count of 30 depending on the comparison base. The learner should feel the denominator changing the meaning before memorising terminology.

270. The learner should become their own denominator editor | 自我编辑

Before publishing any percentage, ratio or rate, ask: What exactly is counted? Out of what? Per what? Which period? Is the value raw or adjusted? Is the starting value visible? If I say “higher,” am I comparing equivalent metrics? If I paraphrase the source, did the denominator survive?

271. Reading harvest: underline the denominator phrase | 阅读提取

When reading reports, underline phrases such as of respondents, per 100,000 people, as a share of revenue, for every teacher, per transaction, among completers. These small phrases often carry more interpretive meaning than the headline number.

272. Listening harvest: hear “of”, “per” and “among” | 听力提取

In spoken presentations, denominators are often unstressed: “60 percent of respondents,” “12 per thousand,” “among users who completed onboarding.” Missing that tail phrase can make the remembered statistic wrong. Replay and shadow denominator phrases, not only numbers.

273. Writing activation: metric sentence frame | 写作激活

Use the frame: Metric + value + denominator/base + period + adjustment if relevant. Example: “The weighted completion estimate was 72% among students who started the course in 2026.” Only then add comparison: “up five percentage points from 2025.”

274. Speaking activation: 45-second metric explanation | 口语激活

Choose one KPI and explain it without numbers first: “Conversion rate is the share of eligible visits that end in purchase.” Then add numerator, denominator and time period. Finally give the value. If you cannot explain the formula in plain English, the metric is not yet active vocabulary.

275. Paraphrase test: preserve the denominator | 改述测试

Original: “Among students who completed at least eight lessons, 75% passed.” Unsafe: “75% of students passed.” Safe: “Three quarters of students who completed at least eight lessons passed.” A successful paraphrase changes wording but not membership of the denominator.

276. Summary test: keep the comparison base | 摘要测试

Original: “The response rate rose from 20% to 30%, a ten-percentage-point increase and a 50% relative rise.” A concise summary can say “Response rose ten percentage points to 30%.” It must not say “Response rose 10%” unless that is explicitly meant as relative change.

277. Translation test: preserve conventional metric names | 翻译测试

Do not translate a formula and then invent an English noun. Check conventional labels: market share, conversion rate, debt-to-income ratio, per-capita income, percentage-point increase. Domain collocation is part of quantitative accuracy.

278. Canonical ownership boundary | 本课所有权边界

Lesson 037 owns quantity relationships and denominators: count, amount, proportion, percentage, share, ratio, rate, per-unit/per-capita language, absolute/relative change and percentage points. It coordinates with Lesson 036 for trends, Lesson 035 for probability, Lesson 034 for thresholds, Lesson 038 for central values, Lesson 039 for distributions and Lesson 040 for occurrence/frequency.

279. Recommended reference floor | 推荐参考资源

• Cambridge Dictionary — ratio — relationship between two groups or amounts.
• Cambridge Dictionary — proportion — part compared with whole/another amount and usage patterns.
• Cambridge Dictionary — percentage — amount expressed out of 100.
• Cambridge Dictionary — percent — percent grammar and examples.
• University of Manchester Academic Phrasebank — Describing Trends — percentage and quantitative change language.

280. SEO language map | 本课关键词范围

This lesson naturally serves percentage vs proportion, ratio vs rate, percentage vs percentage points, quantity vocabulary English, share vs percentage, rate denominator, numerator denominator English, per capita English, ratio proportion difference, absolute vs relative change, count vs rate, market share vs sales, student teacher ratio, weighted percentage, denominator effect, advanced quantitative English, C1 C2 data vocabulary, 英语比例词汇, 百分比百分点区别, 比例比率英语区别, 占比英语, 人均英语, 分母英语, 数量英语 and 中文母语高级英语.

281. Your assignment | 本课作业

Choose a real report containing at least five quantitative metrics. Build one metric-control card for each. Then write a 250-word explanation for a general reader that contains no unexplained denominator shifts.

1. Identify each metric name.
2. Recover numerator and denominator.
3. Name unit and time period.
4. Mark raw/weighted/adjusted/indexed status.
5. Identify count, proportion, percentage, ratio, rate or share.
6. Check whether comparisons use the same definition.
7. Write absolute and relative changes where relevant.
8. Use percentage-point language correctly.
9. Translate the paragraph into Mandarin.
10. Translate it back to English and audit denominator survival.

282. Why this owner earns longform depth | 为什么这一课需要长文深度？

The learning job is not arithmetic tutoring. It is linguistic control of quantitative relationships across reports, exams, research, business, science and bilingual transfer. A short glossary cannot teach why a count can rise while a share falls, why a rate can fall while events increase, why 40%→50% is +10 points but +25% relatively, why a ratio’s order matters, or why “50% of respondents” is not automatically “50% of the population.” These are distinct reasoning failures with distinct repairs.

283. Final rule | 最后一条规则

The number tells you how much. The denominator tells you what the number means.

数字告诉你有多少；分母告诉你这个数字到底是什么意思。

Never compare percentages until you know what each one is a percentage of.

在不知道每个百分比的分母之前，不要比较它们。

Series: EDKS-ADV-VOC-ZH · Lesson 037.