To translate fractions, ratios and proportions accurately, first decide what is being compared. “Three out of five” describes a part-to-whole relationship. “Three to five” may describe a part-to-part ratio or another ordered comparison. “Three-fifths” is a fraction equal to 0.6, while a 3:5 ratio does not by itself say that three is three-fifths of the whole. These expressions are close enough in everyday language to invite errors, but mathematically they answer different questions.
This guide explains how to translate fractions, ratios, proportions, rates and part-to-whole comparisons without changing the denominator, reversing the order of a ratio or inventing a percentage that the source never stated. It is designed for educational, scientific, business and public-information texts where readers must understand how quantities relate. The central method is simple: name the numerator, name the denominator, identify whether the denominator includes the numerator, and preserve the order of comparison.
A reliable translation therefore treats the slash, colon, words such as “of,” “per,” “to,” “for every,” and expressions such as “one in ten” as mathematical relationships rather than punctuation choices. The target language can rearrange the sentence naturally, but the relationship must survive. A fluent translation that turns a 1:4 ratio into one quarter of the whole can be numerically plausible and still be wrong.
A fifty-second orientation
A fraction usually represents one quantity divided by another. A ratio compares two quantities in an ordered relationship. A proportion can mean a part of a whole, or an equation stating that two ratios are equal, depending on the source. In public-health teaching material, the CDC explicitly distinguishes ratios and proportions: in a proportion, the numerator is included within the denominator, while an ordinary ratio need not have that relationship.
Before translating, rewrite the source in a diagnostic sentence. For 3/8, write “three selected items out of eight total items” if that is the intended meaning. For 3:8, write “three A items for every eight B items” if the source compares two groups. This temporary expansion prevents a short target expression from hiding a changed denominator.
A useful external reference is the CDC’s Principles of Epidemiology section on ratios and proportions, which shows how the same arithmetic form can support different descriptive measures depending on how numerator and denominator are related. The translation lesson applies far beyond epidemiology: preserve the relationship before polishing the wording.
1. A numerator and denominator have different jobs
In the fraction 3/5, three is the numerator and five is the denominator. The denominator defines the reference whole under the ordinary part-to-whole reading. The numerator identifies how many of those equal parts are being considered. Swapping them gives 5/3, a value greater than one and a different claim.
Translators sometimes focus on the nouns and miss the structural words. “Three of five students completed the task” is not equivalent to “five of three students.” The sentence order may change dramatically in another language, but the group of five must remain the reference set and the group of three must remain the selected subset.
A useful QA note writes both roles explicitly: numerator = completed students; denominator = all five students in the stated group. If the target sentence no longer makes those roles recoverable, it needs revision. Mathematical fidelity is not achieved merely because both numbers still appear somewhere in the paragraph.
2. “Out of” usually signals a whole, but the source still controls the meaning
“Seven out of ten responses were correct” usually means seven correct responses among ten total responses. The denominator contains both correct and incorrect responses. The implied fraction is 7/10, equal to 70% if a percentage conversion is authorised.
Do not replace “out of” with a part-to-part expression. “Seven correct for every ten incorrect” describes a completely different dataset. Likewise, “seven to ten” without a clear noun pair can become ambiguous in translation because the reader no longer knows whether ten is the whole or a second category.
When the target language does not have a direct compact equivalent, make the whole explicit. “Seven correct responses among ten total responses” may be longer, but it protects the denominator. Clarity is preferable to a shorter phrase that could be interpreted as two separate groups.
3. A ratio can compare quantities that do not form one whole
Suppose an invented event has 30 tutors and 90 learners. The tutor-to-learner ratio is 30:90, which simplifies to 1:3. That means one tutor for every three learners. The denominator-like second term is not a total that includes the first term; it is another category.
If someone converts 1:3 directly into “one quarter of participants are tutors,” that conclusion requires an additional step: one tutor plus three learners gives four participants in each simplified group, so tutors are one quarter of the combined total under the stated two-category assumption. The ratio itself and the derived share are related but not identical statements.
A translation-only task should usually preserve the ratio the source chose. A private calculation can verify what the ratio implies, but it does not automatically authorise adding a derived percentage. This separation protects source fidelity while still allowing rigorous QA.
4. Ratio order matters
A ratio of adults to children of 2:5 is not the same as a ratio of children to adults of 2:5. Reversing the nouns requires reversing the numerical order if the same group is being described. The correct reciprocal relationship would be children to adults = 5:2.
This is a common translation risk because some languages naturally place descriptive nouns in a different order. The translator may improve the sentence by moving “children” earlier and accidentally leave the numbers unchanged. The result looks smooth but reverses the comparison.
Write a verbal ratio before translating: “two adults for every five children.” After the target sentence is complete, translate it back into that verbal form. If the same statement returns, the ratio order survived. This simple reverse test is more reliable than checking punctuation alone.
5. Equivalent fractions can look different while representing the same proportion
The fractions 1/2, 2/4 and 50/100 are numerically equal. Multiplying or dividing numerator and denominator by the same nonzero factor preserves the fraction’s value. A source may choose one form because it reflects actual counts, while another form is easier for teaching or comparison.
Do not simplify automatically when the original counts matter. “18 of 36 participants” tells the reader the sample size. Rewriting it as “one half of participants” preserves the share but removes evidence about how many people were observed. In research, assessment and survey texts, that loss may matter.
Conversely, a teaching text may intentionally simplify 18/36 to 1/2 to show equivalence. Preserve the pedagogical purpose. Translation quality includes knowing which information belongs to the number’s value and which belongs to the number’s chosen representation.
6. Mixed numbers and improper fractions should not be mistaken for separate counts
The mixed number 2 1/2 means two and one half, equal to 2.5 or 5/2. It does not mean “two items and one item out of two” unless the surrounding context explicitly decomposes the quantity that way. Spacing and typography carry structure.
An improper fraction such as 7/4 equals 1.75. Converting it to 1 3/4 can help a general reader, but that is a representation change rather than ordinary word translation. In a mathematical proof or formula, preserving the fraction form may be more appropriate because later operations depend on it.
Keep source form, target audience and calculation role separate. A recipe, school worksheet and algebraic derivation may all benefit from different displays of the same quantity. The numerical value must remain constant while the presentation decision changes.
7. Fractions of quantities require the unit to travel with the operation
“Three quarters of a litre” is 0.75 litre. “Three quarters of the containers” describes a share of a count, not a volume. The fraction acts on whatever noun phrase follows it. Translating the number correctly but attaching it to the wrong noun changes the measured object.
For an invented instruction, “use one third of the prepared solution” refers to the solution amount. If the target says “prepare one third strength solution,” the fraction now modifies concentration rather than quantity. The same words one third appear, but the operation is different.
A strong review asks, “One third of what?” Write the answer explicitly. Fractions rarely stand alone in real technical prose; they operate on a defined whole. The noun phrase that defines that whole deserves the same protection as the fraction itself.
8. A proportion can mean a part of a whole
In descriptive statistics and epidemiology, a proportion often has a numerator that is included in the denominator. If 20 of 100 items are defective, the defective proportion is 20/100 = 0.20. The corresponding percentage is 20% if conversion is appropriate.
The wording must preserve category inclusion. If the denominator becomes “non-defective items” rather than “all items,” the measure changes from a proportion to a part-to-part ratio of defective to non-defective. For the same dataset, that ratio would be 20:80, or 1:4.
This is why the word proportion should not be translated merely as a generic comparison word. It can carry a specific statistical structure. When the source definition is technical, keep that structure visible through the target terminology or an explanatory phrase.
9. A proportion can also mean an equality between ratios
In school mathematics, a proportion may be an equation such as 2/3 = 8/12. The statement says the two ratios are equal. It does not merely say that each fraction describes some share independently.
A translator should preserve the equality relation. In explanatory prose, “two is to three as eight is to twelve” can express the same relationship. But if the target language uses a phrase that sounds like approximate similarity rather than equality, the mathematical force may weaken.
When a text moves between the statistical and school-mathematics senses of proportion, do not assume one target term will work equally well in every sentence. The Vocabulary Learning Hub is relevant here because technical vocabulary often requires sense-level rather than word-level matching.
10. Cross-multiplication checks equality; it does not create meaning by itself
For 2/3 = 8/12, the cross-products are 2 × 12 and 3 × 8, both equal to 24. That verifies equality. But cross-multiplication cannot tell you whether the source quantities should have been compared in that order. The semantic setup must be correct first.
Suppose a source compares cost to quantity, but the translation reverses one fraction to quantity per cost. A cross-product calculation may still produce some valid equation, but it would answer a different real-world question. Mathematical technique does not rescue a mistranslated denominator.
Use cross-multiplication as a QA tool after identifying the quantities. First label each term. Then verify equality. The order prevents a mechanically correct calculation from validating a semantically wrong translation.
11. “Per” creates a rate, which is a specialised ratio
“Five items per box” compares item count with box count. “Five boxes per item” is the reciprocal rate and describes a radically different arrangement. The word per therefore carries the direction of division.
Rates may compare quantities with different units: kilometres per hour, currency units per kilogram, cases per 10,000 people. Translation must preserve both numerator and denominator units. A target that says “10 kilometres per litre” instead of “10 litres per 100 kilometres” is not a word-for-word variant; it uses a reciprocal efficiency measure.
When a target culture commonly uses a reciprocal convention, conversion is an editorial or localisation decision requiring calculation. Do not reverse the rate while keeping the number. The source quantity and target quantity may move in opposite numerical directions even when they describe the same physical efficiency.
12. “One in N” needs careful interpretation
“One in five” usually expresses one occurrence among five total cases, corresponding to a proportion of 1/5 or 20%. But the phrase can be used loosely in conversational sources, so verify the intended denominator before converting it into formal notation.
Do not confuse “one in five people” with “one person for every five others.” The first usually implies one selected person within a group of five; the second can imply a ratio of one to five between two categories, producing groups of six under a complete two-category interpretation.
If the distinction matters, expand the target: “one out of every five people in the stated population.” The extra words protect the denominator and reduce ambiguity. Translation can be concise later, but only after the mathematical relationship is secure.
13. “One to four” is not automatically one quarter
A ratio of 1:4 between group A and group B means one A for every four B. Under an assumption that A and B exhaust the population, A forms one part out of five combined parts, or 20%. It does not form 25% of the whole.
The 25% value comes from 1/4, which answers a different question: one divided by four. This is one of the most common traps when a colon ratio is casually rewritten as a slash fraction. The symbols may appear to invite substitution, but the real-world denominator changes.
A translator should therefore resist automatically replacing colon notation with a fraction word. Translate the comparison first. If the source wants a share of the combined whole, derive it explicitly and only when the assignment permits that transformation.
14. Ratios can be scaled without changing their relationship
The ratio 2:3 is equivalent to 4:6 and 20:30. Multiplying both terms by the same positive factor preserves the relationship. This is useful when translating recipes, maps, mixtures or staffing descriptions where quantities may be scaled.
However, the source may care about actual counts. A class with two tutors and three learners is not operationally identical to one with twenty tutors and thirty learners even though the simplified ratio matches. Scale equivalence preserves the ratio, not every real-world consequence.
Do not replace observed counts with a simplified ratio when sample size, capacity or quantity matters. Store both where necessary: actual counts 20:30; simplified ratio 2:3. That keeps the evidence and the comparison available without confusing one for the other.
15. Mixture ratios need a declared convention
An invented mixture described as 1:4 could mean one part concentrate plus four parts water. The final combined mixture then contains five total parts under that interpretation. But some industries or informal instructions may use ratio language differently. The source definition controls the translation.
A flawed target might interpret 1:4 as one part concentrate in four total parts, which would mean one part concentrate plus three parts water. The mixture becomes more concentrated even though the visible ratio looks unchanged.
Translate the relationship in words when safety or quality depends on it: “one part concentrate to four parts water.” Do not derive a final concentration percentage unless the source or brief requires it. A clear ratio statement is often safer than an unexplained percentage conversion.
16. Scale ratios and map ratios describe correspondence
A scale of 1:100 means one unit on the representation corresponds to one hundred of the same units in reality when the convention is stated that way. One centimetre on the drawing corresponds to one hundred centimetres, or one metre, in the represented object.
Do not treat the second term as a percentage. A 1:100 scale is not one percent in the ordinary sense of a measured share, even though 1/100 equals 1%. The ratio communicates a correspondence between lengths.
Keep the direction visible: drawing to reality, model to original or image to object. A translation that reverses the nouns reverses the scale relationship. The existing map and cartographic translation guides provide specialist context; this article focuses on the ratio logic beneath them.
17. Fractions and percentages can be equivalent but rhetorically different
Three quarters equals 75%. One eighth equals 12.5%. Converting between the forms can help readers, but the source may choose a fraction because it is concrete and countable, or a percentage because it supports comparison across groups of different sizes.
“Three of four schools responded” tells the reader the count and total. “Seventy-five percent of schools responded” hides the small sample size unless the count appears elsewhere. Both can be mathematically correct, but they communicate different evidence.
The previous Translate guide to percentages, percentage points and basis points explains percentage-specific risks. Here, the key rule is to preserve the source representation when its form carries information beyond the numerical value.
18. Percentages above 100 can be valid while part-to-whole proportions cannot exceed one
A part-to-whole proportion lies between zero and one when the numerator is genuinely a subset of the denominator. The corresponding percentage lies between 0% and 100%. If a source claims that 130% of a fixed population belongs to one mutually exclusive subset, something needs clarification.
By contrast, a relative increase can exceed 100%. A quantity that grows from 10 to 25 has increased by 150% relative to its starting value. That percentage is not a population proportion. Translation must identify which mathematical type the percentage belongs to.
Use impossible-range checks during QA. If a translated part-to-whole share exceeds the whole, inspect the denominator, categories and conversion. A simple mathematical boundary can reveal a semantic error that ordinary proofreading misses.
19. Missing categories can make a ratio look like a complete population
Suppose a fictional survey reports 40 respondents choosing A and 60 choosing B, but another 20 chose C. The A:B ratio is still 40:60, or 2:3. However, A is not 40% of all respondents; it is 40/120, or one third.
A translator who assumes the ratio’s two categories exhaust the whole may derive a false percentage. The source might have omitted other categories because the paragraph focuses only on A and B. Do not create a whole from the visible ratio unless the text establishes that the categories are complete.
This is an important stop rule: before turning a part-to-part ratio into a part-to-whole share, verify that all relevant categories are included. If not, preserve the ratio and avoid the derived percentage.
20. Ratios with units should not always be simplified numerically
A source might report 50 kilometres per 2 hours. Numerically, that ratio simplifies to 25 kilometres per hour. The simplified form is often useful, but the original may describe an actual journey segment lasting two hours.
Similarly, “12 cases in 3 days” can be transformed to an average of four cases per day, but only if an average rate is the intended interpretation. The source might instead be reporting a total observed over a specific three-day window.
Do not replace an observed total with an average rate simply because simplification is mathematically possible. Translation should preserve the author’s measurement frame. Derived rates can support checking or an authorised explanatory note, not automatic rewriting.
21. Worked case: survey results
Consider this fictional source: “Of 240 valid responses, 144 selected option A and 96 selected option B. The A:B ratio was 3:2, and option A represented 60% of valid responses.” The counts, ratio and proportion are internally consistent.
A flawed target might say: “Three fifths selected A for every two fifths selecting B, a ratio of 3:2 among all invited participants.” The first clause can be mathematically reconstructed, but the final phrase changes valid responses into all invited participants, introducing a denominator the source never supplied.
A correct translation keeps the valid-response denominator attached to the 60% and preserves A:B order. The ratio can remain 3:2 without implying anything about non-respondents. The private check recomputes 144/240 = 0.6 and 144:96 = 3:2.
22. Worked case: a recipe ratio
Use this invented source: “Combine one part concentrate with four parts water. For a five-litre batch under this simplified exercise, use one litre concentrate and four litres water.” The 1:4 ratio refers to component amounts, and the stated batch size confirms five total parts.
A mistranslation saying “one quarter concentrate and three quarters water” would describe a 1:3 component ratio, not 1:4. The fraction one quarter sounds close to “one to four,” but the denominators are different.
The correct QA method expands the ratio to total parts: one plus four equals five, so concentrate is one fifth of the final simplified volume. That derived fraction verifies the source relationship but need not replace the source wording. The example is arithmetic instruction only and not a recommendation for any real chemical or food preparation.
23. Worked case: a staffing ratio
A fictional programme reports “one instructor for every six learners.” If there are exactly two categories and the ratio holds uniformly, the simplified instructor-to-learner ratio is 1:6. In each seven-person grouping, instructors form one seventh of the combined count.
A translation that says “one instructor for every six participants” may be wrong if instructors are themselves participants in the broader programme. The word learner defines the denominator category. Replacing it with a broader noun can collapse the distinction between the two groups.
Terminology and mathematics therefore interact. Preserve the group labels as carefully as the numbers. A ratio is only meaningful when its two terms still refer to the same categories after translation.
24. Worked case: a rate per population
Suppose a fictional report states “12 events per 10,000 observations.” This is a rate representation chosen to avoid an awkward small decimal. It corresponds to 0.0012 events per observation under a simple arithmetic interpretation, or 0.12% if each observation can contribute at most one event and the source supports a proportion reading.
Do not automatically translate the rate as 0.12%. The “per 10,000” form may be standard for the field, and an observation could have a structure that makes the percentage interpretation inappropriate. Preserve the published measure unless adaptation is explicitly required.
The CDC’s ratio examples show why multiplying a small per-person quantity by a power such as 10,000 can make a rate easier to understand. Translation should keep the chosen reference population visible, because changing “per 10,000” to “per 1,000” requires a corresponding numerical change.
25. Practice clinic with explained answers
Practice one. Seven out of twenty items pass inspection. The fraction is 7/20 and the proportion is 0.35. A percentage representation is 35% if authorised. The denominator is all twenty items, not the thirteen failures.
Practice two. The ratio of red to blue objects is 2:3. Under a two-category whole, red objects form 2/5 of the combined total, or 40%. They do not form 2/3 of the total.
Practice three. The ratio of teachers to students is 1:10. Reversing the noun order requires 10:1. Keeping 1:10 while saying students to teachers reverses the relationship.
Practice four. Simplify 18:24 to 3:4 by dividing both terms by six. Keep the actual counts if sample size matters; the simplified ratio alone does not tell the reader there were forty-two observed items.
Practice five. Convert three quarters to a decimal. 3/4 = 0.75. In percentage form it is 75%. Preserve the source form when its representation is part of the teaching purpose.
Practice six. A mixture is defined as one part A to nine parts B. Under that explicit component convention, A is one tenth of the final two-component mixture. It is not one ninth.
Practice seven. Five kilometres in twenty-five minutes is not “five kilometres per twenty-five kilometres.” The denominator is time. Unit labels must travel with the ratio terms.
Practice eight. A scale is 1:50. Under the stated drawing-to-reality convention, one centimetre on the drawing represents fifty centimetres in reality. Reversing the noun order without reversing the ratio changes the scale.
Practice nine. A source reports 30 successes and 70 failures. Success proportion is 30/100 = 30%. Success-to-failure ratio is 30:70 = 3:7. Those are related but different measures.
Practice ten. The source says one in eight. Under a part-to-whole interpretation, the proportion is 1/8 = 12.5%. Do not rewrite it as a 1:8 part-to-part ratio without checking the intended meaning.
Practice eleven. A ratio 4:5 becomes 8:10 after multiplying both terms by two. The ratio is equivalent, but the actual count has doubled. Preserve whichever level of information the source communicates.
Practice twelve. The source says “two thirds of the sample,” but the target draft says “two to three in the sample.” Restore the part-to-whole relationship. A colon-like reading would introduce a second group rather than a whole.
26. A release checklist for fractions, ratios and proportions
Identify every numerator and denominator. Mark whether the numerator is included in the denominator. For each colon ratio, label the first and second terms in words. For every rate, record the numerator and denominator units. For every derived percentage, verify that conversion was authorised and that the whole is complete.
Check whether simplified forms remove useful sample-size information. Check whether a target noun is broader or narrower than the source category. Check whether one in N, A:B and A/B have been treated as distinct structures rather than visual variants. Finally, reverse-read the target: can you reconstruct the original comparison without guessing?
These checks take less time than repairing a published numerical misunderstanding. They also improve prose because once the mathematical roles are explicit, the translator can choose clearer words instead of clinging to source syntax out of fear of changing the numbers.
27. Where this article sits in the eduKate translation system
This specialist article extends Translate | Names, Numbers, Dates and Units and complements the separate guide to Percentages, Percentage Points and Basis Points. The canonical architecture remains Master Art of Translation, with the Vocabulary Learning Hub and How English Works supporting precise terminology and grammatical attachment.
The governing rule is simple: before translating the words around a fraction or ratio, identify the whole, the parts and the direction of comparison. Once those are stable, the sentence can be reorganised naturally. Without that step, a target can be perfectly grammatical while dividing the wrong quantities.
Fractions, ratios and proportions are compact because they compress relationships. Good translation reverses that compression privately, checks the relationship, and then rebuilds an equally compact target expression only when the meaning is secure. That is how numerical language stays readable without becoming numerically different.