Did you know that a school data table can show 30 per 1,000 and 3% for the same proportion? One is not necessarily a mistake. The problem begins when someone sees the number 30, adds a percent sign, and announces a tenfold larger result. That tiny change in the unit can turn a careful comparison into a misleading headline.
A per-1,000 figure scales a count relative to a specified denominator to a reference group of 1,000. A percentage scales the same proportion to 100. If the numerator represents the same people and period, 30 per 1,000 equals 3%, not 30%. The practical rule is to identify the numerator, the denominator, the time period and the unit before comparing or quoting the number.
Start with a diagnostic question, not a formula
Suppose an illustrative school’s library records that 18 out of 600 eligible pupils borrowed a particular optional reference kit during one week. A student writes, ‘30% borrowed it.’ Before correcting the sentence, ask: What does 18 count? What does 600 count? Does anyone appear more than once? What does ‘per 1,000’ mean here? The learner’s answers show whether the error is calculation, denominator choice or a misunderstanding of the label.
In this invented example each pupil is counted at most once, and 600 is the complete eligible group for that week. Therefore the observed share is 18 ÷ 600 = 0.03. Expressed as a percentage it is 0.03 × 100 = 3%. Expressed per 1,000 it is 0.03 × 1,000 = 30 per 1,000 eligible pupils. We are describing a single observation, not predicting a pupil’s future behaviour.
Hand-check the units and meaning
- State the counting event: 18 distinct pupils borrowed the kit during the stated week.
- State the denominator: 600 eligible pupils, not 600 loans or 600 school days.
- Calculate the unscaled fraction: 18/600 = 3/100 = 0.03.
- Multiply by 100 for a percentage: 0.03 × 100 = 3%.
- Multiply by 1,000 for a per-1,000 figure: 0.03 × 1,000 = 30 per 1,000.
- Reverse-check: 30 ÷ 1,000 × 600 = 18 pupils, agreeing with the original count.
A dimensional label matters. ‘Thirty’ alone is incomplete. ‘Thirty borrowers’ is an absolute count; ’30 borrowers per 1,000 eligible pupils during the week’ is a scaled share. A class of 100 and a school of 1,000 may have the same percentage while involving very different numbers of people. Conversely, the same count can represent very different proportions when denominators differ.
Why the headline can sound ten times too large
A percentage means ‘per hundred.’ A per-thousand expression means ‘per thousand.’ The conversion is therefore mechanical: per-1,000 figure ÷ 10 = percentage, provided both describe the same proportion with identical counting rules and period. A data table showing 42 per 1,000 corresponds to 4.2%; a table showing 7 per 1,000 corresponds to 0.7%. Neither calculation involves rounding first or inserting the original raw count in place of the numerator.
Watch for a second common error: an apparent ‘rate’ may count repeated events, not distinct people. If a school recorded 42 incidents among 700 pupils over a term, 60 incidents per 1,000 pupils is an event-to-population ratio; it does not mean 6% of pupils each experienced exactly one incident. One pupil could have several incidents. A true incidence rate may instead have person-time in the denominator. Always copy the source definition rather than assuming every item marked ‘rate’ is an individual risk.
Comparisons fail when the two denominators differ
Consider two imaginary reading clubs. Club A has 12 kit borrowers from 300 eligible pupils. Club B has 20 from 800. Club A’s figure is 12 ÷ 300 × 1,000 = 40 per 1,000, or 4%. Club B’s is 20 ÷ 800 × 1,000 = 25 per 1,000, or 2.5%. Club B has more borrowers in absolute number but a lower eligible-pupil share. Which quantity you report depends on the question: total demand for kits or participation relative to the available population.
Notice that we have not concluded why one club has a different participation share. Access hours, borrowing rules, student interests and the week chosen might matter. Arithmetic tells us what the observations are; a causal explanation requires separate evidence. The related Core Aim guide to statistical footnotes shows why exclusions and notes should be checked before a number is cited.
A deliberate error-repair exercise
Show a learner the sentence: ‘There were 24 cases per 1,000 pupils, meaning 24% of pupils were affected.’ Ask for two corrections rather than one. First, 24 per 1,000 converts to 2.4% if it is a distinct-pupil proportion. Second, the sentence must not automatically say ‘pupils affected’ if cases may repeat for an individual. A safer statement without further information is ’24 recorded cases per 1,000 pupils during the defined period.’ This distinction between a numerical conversion and a truthful population claim is precisely what a mature student should learn.
Progressive practice that reveals genuine understanding
Guided: translate the unit
Give the student ‘5 out of 250 pupils’. Draw a bar divided into 250 equally sized positions and shade five. Simplify 5/250 to 1/50 = 0.02. The learner writes 2% and 20 per 1,000 pupils. Ask them to explain aloud why these are the same proportion, not two different observations.
Supported: compare groups with different sizes
Use the two clubs above but hide the worked answer. Have the student compute each group’s per-1,000 result, show its units and decide which club has more borrowers in total and which has a larger participation share. A correct explanation must state both comparisons and why the answers point to different clubs.
Independent: inspect the note before calculating
Invent a new table: 9 recorded incidents, 450 pupils, term 1; 11 recorded incidents, 550 pupils, term 2. Both yield 20 incidents per 1,000 pupils (9/450 × 1,000 = 20; 11/550 × 1,000 = 20), but ask whether the figure describes unique affected pupils. The correct independent response flags the missing repeat-event information rather than quietly turning incidents into people.
How parents and teachers can track improvement
A useful three-question check is: (1) Can the learner calculate from the underlying counts? (2) Can the learner attach the complete denominator and period to the answer? (3) Can the learner explain what the value does not prove? Record the error type on each attempt: scale conversion, wrong denominator, missing time, count-versus-person confusion, or unsupported explanation. Repeat a comparable exercise later with different values. Progress is the student independently giving the right mathematical conversion and a measured, limited interpretation.
Parents need not know specialist statistics to help. Ask, ‘Per thousand of what?’ and ‘Did the source count people or events?’ If the learner cannot say, return to the table heading or footnote. The Core Aim lesson on changing club membership adds another useful case where a denominator must be chosen carefully. For broader teaching and self-checking paths, use the eduKate education learning hub and the diagnosis-first Clementi mathematics lesson reference.
Limitations: correct arithmetic cannot fix undefined data
Without a source’s precise population, period, case definition and treatment of repeats, a per-thousand value can be numerically consistent but scientifically unclear. A rounded rate may not reconstruct the original whole count exactly, and a low count may vary considerably from one observation period to another. Avoid assuming a reported difference is important or caused by a particular programme simply because the numbers have been standardised.
Frequently asked questions
Is 1 per 1,000 equal to 1%?
No. One per 1,000 is 0.001 of the denominator, or 0.1%. One percent is ten per 1,000.
Can a per-thousand figure exceed 1,000?
For a proportion counting each person at most once, it cannot exceed 1,000 per 1,000. A ratio of repeated events to a population can exceed 1,000 if individuals have multiple events; that is another reason to check the definition before interpreting it as a probability.
Should I always convert per 1,000 to a percentage?
No. Preserve the source’s familiar unit when it supports a fair comparison, and provide a percentage as a clearly labelled explanation if helpful. Never change a published unit silently. Public-health training materials from the US Centers for Disease Control and Prevention show how ratios, proportions and rates require care with their denominators and multipliers.
The useful habit to keep
Whenever a data point looks dramatic, temporarily remove the headline and reconstruct the sentence from its components: what happened, to whom, over what period, out of how many, multiplied by what scale? Education succeeds here when the student can check a claim independently instead of memorising a conversion trick.