Choose your reading route
eduKateSG · Why Mathematics? · Weighted Median, Cumulative Weights and the Population Midpoint
Start with the question, audit the worked example, or jump to the assumptions and practice. Every chapter stays open for uninterrupted reading.
Understand the question
Define the quantity and follow the mechanism.Route 2 · Chapters 5–9
Reproduce the example
Audit the numbers one step at a time.Route 3 · Chapters 10–15
Check assumptions
See where design and calibration matter.Route 4 · Chapters 16–20
Explore applications
Connect the method to a concrete investigation.Route 5 · Chapters 21–24
Study and verify
Practise, inspect failure cases and read further.
Full chapter index · Practice and checking · Mathematics Learning Hub
Why is mathematics important when one row can represent many people, households or events? A weighted median finds the point where cumulative weight reaches half of the total. It keeps the median's ordered, distribution-aware meaning while respecting unequal representation.
This matters in official statistics, survey analysis, education research and any summary where sample records stand for different amounts of a population. The calculation is short; the careful work is defining the population, choosing the right weights and stating the tie convention.
Did you know? A weighted median can be one observed value even when no single record carries half the weight. What matters is the cumulative weight on either side.
Full chapter index
Chapters 1–4 · Understand the question
Chapters 5–9 · Reproduce the example
Chapters 10–15 · Check assumptions
Chapters 16–20 · Explore applications
Chapters 21–24 · Study and verify
An ordinary median orders observations and finds the middle position. A weighted median orders values and finds where accumulated weight crosses 50% of total weight.
That distinction lets a small sample describe a larger target population when each record has a documented representation weight.
The Australian Bureau of Statistics explains that records are ranked in ascending order, their weights are accumulated, and the record at which half the population is accrued supplies the median value.
This definition makes the mechanism visible: order first, accumulate second, interpret the crossing point third. Read the ABS methodology.
Suppose travel times are 10, 20, 30 and 60 minutes with weights 1, 2, 5 and 2. Total weight is 10, so half is 5.
Cumulative weights are 1, 3, 8 and 10. The first cumulative total at least 5 occurs at 30 minutes, so 30 is the weighted median under this convention.
| Values | Weights | Cumulative weights |
|---|---|---|
| 10, 20, 30, 60 | 1, 2, 5, 2 | 1, 3, 8, 10 |
The ordinary median of 10, 20, 30 and 60 is 25 minutes because it averages the two middle observations. That treats the four rows equally.
The weighted median is 30 because the value 30 represents five units. The two answers address different populations, so neither should be substituted silently.
| Total weight | Half total | First crossing |
|---|---|---|
| 10 | 5 | 30 minutes |
Integer frequency weights can be understood by expanding the data. The example becomes 10, 20, 20, five copies of 30, and two copies of 60.
The middle positions lie among the 30s. Expansion is a useful classroom check, though cumulative sums are more efficient for large weights.
CHAPTER 6 OF 24 · Reproduce the example
6. Survey weights and represented populations
A survey weight may say how many population units a sampled record represents. It is not a score for how trustworthy or important that person is.
The US Census Bureau describes weights as a bridge from sample persons or households to population estimates. Read the Census weighting explanation.
CHAPTER 7 OF 24 · Reproduce the example
7. Person-weighted and household-weighted questions
One household record can support a household-weighted median or, after multiplying by household size, a person-weighted median. These describe different target units.
The ABS explicitly distinguishes these cases. Students should write the unit of analysis before calculating.
If cumulative weight equals exactly half at one value and the next positive weight lies at a higher value, software packages may use different conventions.
Some return the lower value, some the upper value and some interpolate. Declare the rule and test it with a tiny dataset.
Multiplying every positive weight by the same constant leaves the weighted median unchanged because the cumulative proportions stay the same.
This invariance is a powerful audit: weights 1, 2, 5, 2 and 10, 20, 50, 20 should give the same crossing value.
A zero-weight row contributes no population mass, though it may remain in the source file for traceability. Negative weights break the simple cumulative-population interpretation.
Advanced survey methods can create unusual weights, but a school explanation should not treat negative weights as ordinary frequencies.
A missing outcome with a positive weight cannot simply become zero. That would invent a real value and could move the median.
Use the survey's declared missing-data method, preserve the original flag and report how much total weight was excluded or adjusted.
Like the ordinary median, a weighted median is resistant to the magnitude of extreme values so long as those extremes do not carry enough weight to cross the halfway point.
It is not immune to influential weights. One extremely large weight can determine the answer.
One centre does not reveal spread, gaps, multiple clusters or subgroup differences. Two populations can share a weighted median while having very different distributions.
Pair it with weighted quantiles, counts, uncertainty estimates and a graph when decisions depend on more than the middle.
Give groups the same four values but different representation weights. Ask them to predict the direction of movement before accumulating.
Then ask which weight story is defensible: frequency, population representation, exposure or an arbitrary preference. The story matters as much as the arithmetic.
Sort value and weight columns together, compute cumulative weight, calculate half the total, and locate the first row whose cumulative total meets or exceeds the threshold.
Add checks for unsorted values, missing pairs, negative weights and a zero total.
Keep each value paired with its weight through sorting. Sum with adequate numerical precision, then apply one documented crossing and tie rule.
Unit tests should cover odd and even expanded counts, exact-half boundaries, duplicate values, zero weights and rescaled weights.
A weighted median can improve representation when weights correctly describe the target population. It can also mislead when weights are opaque, outdated or attached to the wrong unit.
Fairness is therefore a design question, not a property guaranteed by the word weighted.
A point estimate is not a certainty statement. Complex survey weights, clustering and stratification affect the sampling uncertainty.
Use design-aware variance or replication methods supplied by the survey rather than attaching an ordinary textbook standard error automatically.
Official statisticians use weighted medians for income, wealth and other skewed variables. Policy analysts, survey researchers and data scientists use the same cumulative-weight idea for quantiles and thresholds.
The transferable skill is connecting every weight to a represented unit and every summary to a decision question.
CHAPTER 20 OF 24 · Explore applications
20. Connection to recent mathematics articles
The weighted-mean guide explains product-and-sum weighting. The quartile coefficient guide shows how ordered positions also support robust spread.
Students should write the target population, sort values, show cumulative weights and mark the halfway threshold. Parents can ask what one record represents and whether the chosen weighting scheme matches the claim.
Confidence grows when a learner can explain why the crossing point is meaningful, not merely obtain it from software.
- Calculate the weighted median of values 5, 8, 12 with weights 2, 5, 3.
- Rescale every weight by 100 and explain why the result stays fixed.
- Construct two datasets with the same weighted median but different spread.
### Is a weighted median the same as a weighted mean? No. The median follows ordered cumulative weight; the mean uses weighted products and is sensitive to every value's magnitude.
### Must weights sum to one? No. Only their relative amounts matter, provided the total is positive and the convention is clear.
### Can the result fall between observed values? It depends on the declared tie and interpolation convention. Many official-statistics procedures select the record where cumulative weight reaches half.
Start with the Mathematics Learning Hub, then read the ABS official methodology linked above and compare the weighted median with the weighted mean on the same data.
Extended practice laboratory
Practice case 1: Sort value-weight pairs
Setup: Use the article’s weighted-median mechanism and change one declared input for case 1. Task: Recalculate from raw values, preserve units and state which mathematical or design condition changed.
Check: Check the relevant invariant, separate the number from its claim, and name the next diagnostic or source of evidence. Rebuild the relevant counts before interpreting the coefficient, preserve the stated design, and write one sentence separating the numerical result from the claim it can support. Then vary one defensible assumption, calculate again, and explain whether the conclusion changes because of new data, a new reference model, or a new decision purpose. Finish by naming the unit, tolerance, and evidence needed for transfer. Record one graph or table that would expose failure hidden by the summary.
Practice case 2: Sum total weight
Setup: Use the article’s weighted-median mechanism and change one declared input for case 2. Task: Recalculate from raw values, preserve units and state which mathematical or design condition changed.
Check: Check the relevant invariant, separate the number from its claim, and name the next diagnostic or source of evidence. Rebuild the relevant counts before interpreting the coefficient, preserve the stated design, and write one sentence separating the numerical result from the claim it can support. Then vary one defensible assumption, calculate again, and explain whether the conclusion changes because of new data, a new reference model, or a new decision purpose. Finish by naming the unit, tolerance, and evidence needed for transfer. Record one graph or table that would expose failure hidden by the summary.
Practice case 3: Mark the halfway threshold
Setup: Use the article’s weighted-median mechanism and change one declared input for case 3. Task: Recalculate from raw values, preserve units and state which mathematical or design condition changed.
Check: Check the relevant invariant, separate the number from its claim, and name the next diagnostic or source of evidence. Rebuild the relevant counts before interpreting the coefficient, preserve the stated design, and write one sentence separating the numerical result from the claim it can support. Then vary one defensible assumption, calculate again, and explain whether the conclusion changes because of new data, a new reference model, or a new decision purpose. Finish by naming the unit, tolerance, and evidence needed for transfer. Record one graph or table that would expose failure hidden by the summary.
Practice case 4: Build cumulative weights
Setup: Use the article’s weighted-median mechanism and change one declared input for case 4. Task: Recalculate from raw values, preserve units and state which mathematical or design condition changed.
Check: Check the relevant invariant, separate the number from its claim, and name the next diagnostic or source of evidence. Rebuild the relevant counts before interpreting the coefficient, preserve the stated design, and write one sentence separating the numerical result from the claim it can support. Then vary one defensible assumption, calculate again, and explain whether the conclusion changes because of new data, a new reference model, or a new decision purpose. Finish by naming the unit, tolerance, and evidence needed for transfer. Record one graph or table that would expose failure hidden by the summary.
Practice case 5: Find the crossing record
Setup: Use the article’s weighted-median mechanism and change one declared input for case 5. Task: Recalculate from raw values, preserve units and state which mathematical or design condition changed.
Check: Check the relevant invariant, separate the number from its claim, and name the next diagnostic or source of evidence. Rebuild the relevant counts before interpreting the coefficient, preserve the stated design, and write one sentence separating the numerical result from the claim it can support. Then vary one defensible assumption, calculate again, and explain whether the conclusion changes because of new data, a new reference model, or a new decision purpose. Finish by naming the unit, tolerance, and evidence needed for transfer. Record one graph or table that would expose failure hidden by the summary.
Practice case 6: Expand frequency weights
Setup: Use the article’s weighted-median mechanism and change one declared input for case 6. Task: Recalculate from raw values, preserve units and state which mathematical or design condition changed.
Check: Check the relevant invariant, separate the number from its claim, and name the next diagnostic or source of evidence. Rebuild the relevant counts before interpreting the coefficient, preserve the stated design, and write one sentence separating the numerical result from the claim it can support. Then vary one defensible assumption, calculate again, and explain whether the conclusion changes because of new data, a new reference model, or a new decision purpose. Finish by naming the unit, tolerance, and evidence needed for transfer. Record one graph or table that would expose failure hidden by the summary.
Practice case 7: Rescale all weights
Setup: Use the article’s weighted-median mechanism and change one declared input for case 7. Task: Recalculate from raw values, preserve units and state which mathematical or design condition changed.
Check: Check the relevant invariant, separate the number from its claim, and name the next diagnostic or source of evidence. Rebuild the relevant counts before interpreting the coefficient, preserve the stated design, and write one sentence separating the numerical result from the claim it can support. Then vary one defensible assumption, calculate again, and explain whether the conclusion changes because of new data, a new reference model, or a new decision purpose. Finish by naming the unit, tolerance, and evidence needed for transfer. Record one graph or table that would expose failure hidden by the summary.
Practice case 8: Test an exact-half tie
Setup: Use the article’s weighted-median mechanism and change one declared input for case 8. Task: Recalculate from raw values, preserve units and state which mathematical or design condition changed.
Check: Check the relevant invariant, separate the number from its claim, and name the next diagnostic or source of evidence. Rebuild the relevant counts before interpreting the coefficient, preserve the stated design, and write one sentence separating the numerical result from the claim it can support. Then vary one defensible assumption, calculate again, and explain whether the conclusion changes because of new data, a new reference model, or a new decision purpose. Finish by naming the unit, tolerance, and evidence needed for transfer. Record one graph or table that would expose failure hidden by the summary.
Practice case 9: Compare ordinary median
Setup: Use the article’s weighted-median mechanism and change one declared input for case 9. Task: Recalculate from raw values, preserve units and state which mathematical or design condition changed.
Check: Check the relevant invariant, separate the number from its claim, and name the next diagnostic or source of evidence. Rebuild the relevant counts before interpreting the coefficient, preserve the stated design, and write one sentence separating the numerical result from the claim it can support. Then vary one defensible assumption, calculate again, and explain whether the conclusion changes because of new data, a new reference model, or a new decision purpose. Finish by naming the unit, tolerance, and evidence needed for transfer. Record one graph or table that would expose failure hidden by the summary.
Practice case 10: Compare weighted mean
Setup: Use the article’s weighted-median mechanism and change one declared input for case 10. Task: Recalculate from raw values, preserve units and state which mathematical or design condition changed.
Check: Check the relevant invariant, separate the number from its claim, and name the next diagnostic or source of evidence. Rebuild the relevant counts before interpreting the coefficient, preserve the stated design, and write one sentence separating the numerical result from the claim it can support. Then vary one defensible assumption, calculate again, and explain whether the conclusion changes because of new data, a new reference model, or a new decision purpose. Finish by naming the unit, tolerance, and evidence needed for transfer. Record one graph or table that would expose failure hidden by the summary.
Practice case 11: Add a zero weight
Setup: Use the article’s weighted-median mechanism and change one declared input for case 11. Task: Recalculate from raw values, preserve units and state which mathematical or design condition changed.
Check: Check the relevant invariant, separate the number from its claim, and name the next diagnostic or source of evidence. Rebuild the relevant counts before interpreting the coefficient, preserve the stated design, and write one sentence separating the numerical result from the claim it can support. Then vary one defensible assumption, calculate again, and explain whether the conclusion changes because of new data, a new reference model, or a new decision purpose. Finish by naming the unit, tolerance, and evidence needed for transfer. Record one graph or table that would expose failure hidden by the summary.
Practice case 12: Move one large weight
Setup: Use the article’s weighted-median mechanism and change one declared input for case 12. Task: Recalculate from raw values, preserve units and state which mathematical or design condition changed.
Check: Check the relevant invariant, separate the number from its claim, and name the next diagnostic or source of evidence. Rebuild the relevant counts before interpreting the coefficient, preserve the stated design, and write one sentence separating the numerical result from the claim it can support. Then vary one defensible assumption, calculate again, and explain whether the conclusion changes because of new data, a new reference model, or a new decision purpose. Finish by naming the unit, tolerance, and evidence needed for transfer. Record one graph or table that would expose failure hidden by the summary.
Practice case 13: Duplicate a value
Setup: Use the article’s weighted-median mechanism and change one declared input for case 13. Task: Recalculate from raw values, preserve units and state which mathematical or design condition changed.
Check: Check the relevant invariant, separate the number from its claim, and name the next diagnostic or source of evidence. Rebuild the relevant counts before interpreting the coefficient, preserve the stated design, and write one sentence separating the numerical result from the claim it can support. Then vary one defensible assumption, calculate again, and explain whether the conclusion changes because of new data, a new reference model, or a new decision purpose. Finish by naming the unit, tolerance, and evidence needed for transfer. Record one graph or table that would expose failure hidden by the summary.
Practice case 14: Change the unit of analysis
Setup: Use the article’s weighted-median mechanism and change one declared input for case 14. Task: Recalculate from raw values, preserve units and state which mathematical or design condition changed.
Check: Check the relevant invariant, separate the number from its claim, and name the next diagnostic or source of evidence. Rebuild the relevant counts before interpreting the coefficient, preserve the stated design, and write one sentence separating the numerical result from the claim it can support. Then vary one defensible assumption, calculate again, and explain whether the conclusion changes because of new data, a new reference model, or a new decision purpose. Finish by naming the unit, tolerance, and evidence needed for transfer. Record one graph or table that would expose failure hidden by the summary.
Practice case 15: Use household weights
Setup: Use the article’s weighted-median mechanism and change one declared input for case 15. Task: Recalculate from raw values, preserve units and state which mathematical or design condition changed.
Check: Check the relevant invariant, separate the number from its claim, and name the next diagnostic or source of evidence. Rebuild the relevant counts before interpreting the coefficient, preserve the stated design, and write one sentence separating the numerical result from the claim it can support. Then vary one defensible assumption, calculate again, and explain whether the conclusion changes because of new data, a new reference model, or a new decision purpose. Finish by naming the unit, tolerance, and evidence needed for transfer. Record one graph or table that would expose failure hidden by the summary.
Practice case 16: Use person weights
Setup: Use the article’s weighted-median mechanism and change one declared input for case 16. Task: Recalculate from raw values, preserve units and state which mathematical or design condition changed.
Check: Check the relevant invariant, separate the number from its claim, and name the next diagnostic or source of evidence. Rebuild the relevant counts before interpreting the coefficient, preserve the stated design, and write one sentence separating the numerical result from the claim it can support. Then vary one defensible assumption, calculate again, and explain whether the conclusion changes because of new data, a new reference model, or a new decision purpose. Finish by naming the unit, tolerance, and evidence needed for transfer. Record one graph or table that would expose failure hidden by the summary.
Practice case 17: Handle a missing value
Setup: Use the article’s weighted-median mechanism and change one declared input for case 17. Task: Recalculate from raw values, preserve units and state which mathematical or design condition changed.
Check: Check the relevant invariant, separate the number from its claim, and name the next diagnostic or source of evidence. Rebuild the relevant counts before interpreting the coefficient, preserve the stated design, and write one sentence separating the numerical result from the claim it can support. Then vary one defensible assumption, calculate again, and explain whether the conclusion changes because of new data, a new reference model, or a new decision purpose. Finish by naming the unit, tolerance, and evidence needed for transfer. Record one graph or table that would expose failure hidden by the summary.
Practice case 18: Reject a negative weight
Setup: Use the article’s weighted-median mechanism and change one declared input for case 18. Task: Recalculate from raw values, preserve units and state which mathematical or design condition changed.
Check: Check the relevant invariant, separate the number from its claim, and name the next diagnostic or source of evidence. Rebuild the relevant counts before interpreting the coefficient, preserve the stated design, and write one sentence separating the numerical result from the claim it can support. Then vary one defensible assumption, calculate again, and explain whether the conclusion changes because of new data, a new reference model, or a new decision purpose. Finish by naming the unit, tolerance, and evidence needed for transfer. Record one graph or table that would expose failure hidden by the summary.
Practice case 19: Plot cumulative proportion
Setup: Use the article’s weighted-median mechanism and change one declared input for case 19. Task: Recalculate from raw values, preserve units and state which mathematical or design condition changed.
Check: Check the relevant invariant, separate the number from its claim, and name the next diagnostic or source of evidence. Rebuild the relevant counts before interpreting the coefficient, preserve the stated design, and write one sentence separating the numerical result from the claim it can support. Then vary one defensible assumption, calculate again, and explain whether the conclusion changes because of new data, a new reference model, or a new decision purpose. Finish by naming the unit, tolerance, and evidence needed for transfer. Record one graph or table that would expose failure hidden by the summary.
Practice case 20: Compare two populations
Setup: Use the article’s weighted-median mechanism and change one declared input for case 20. Task: Recalculate from raw values, preserve units and state which mathematical or design condition changed.
Check: Check the relevant invariant, separate the number from its claim, and name the next diagnostic or source of evidence. Rebuild the relevant counts before interpreting the coefficient, preserve the stated design, and write one sentence separating the numerical result from the claim it can support. Then vary one defensible assumption, calculate again, and explain whether the conclusion changes because of new data, a new reference model, or a new decision purpose. Finish by naming the unit, tolerance, and evidence needed for transfer. Record one graph or table that would expose failure hidden by the summary.
Practice case 21: Inspect uncertainty
Setup: Use the article’s weighted-median mechanism and change one declared input for case 21. Task: Recalculate from raw values, preserve units and state which mathematical or design condition changed.
Check: Check the relevant invariant, separate the number from its claim, and name the next diagnostic or source of evidence. Rebuild the relevant counts before interpreting the coefficient, preserve the stated design, and write one sentence separating the numerical result from the claim it can support. Then vary one defensible assumption, calculate again, and explain whether the conclusion changes because of new data, a new reference model, or a new decision purpose. Finish by naming the unit, tolerance, and evidence needed for transfer. Record one graph or table that would expose failure hidden by the summary.
Practice case 22: Check representation
Setup: Use the article’s weighted-median mechanism and change one declared input for case 22. Task: Recalculate from raw values, preserve units and state which mathematical or design condition changed.
Check: Check the relevant invariant, separate the number from its claim, and name the next diagnostic or source of evidence. Rebuild the relevant counts before interpreting the coefficient, preserve the stated design, and write one sentence separating the numerical result from the claim it can support. Then vary one defensible assumption, calculate again, and explain whether the conclusion changes because of new data, a new reference model, or a new decision purpose. Finish by naming the unit, tolerance, and evidence needed for transfer. Record one graph or table that would expose failure hidden by the summary.
Practice case 23: Avoid fairness overclaim
Setup: Use the article’s weighted-median mechanism and change one declared input for case 23. Task: Recalculate from raw values, preserve units and state which mathematical or design condition changed.
Check: Check the relevant invariant, separate the number from its claim, and name the next diagnostic or source of evidence. Rebuild the relevant counts before interpreting the coefficient, preserve the stated design, and write one sentence separating the numerical result from the claim it can support. Then vary one defensible assumption, calculate again, and explain whether the conclusion changes because of new data, a new reference model, or a new decision purpose. Finish by naming the unit, tolerance, and evidence needed for transfer. Record one graph or table that would expose failure hidden by the summary.
Practice case 24: Write a weighted-median conclusion
Setup: Use the article’s weighted-median mechanism and change one declared input for case 24. Task: Recalculate from raw values, preserve units and state which mathematical or design condition changed.
Check: Check the relevant invariant, separate the number from its claim, and name the next diagnostic or source of evidence. Rebuild the relevant counts before interpreting the coefficient, preserve the stated design, and write one sentence separating the numerical result from the claim it can support. Then vary one defensible assumption, calculate again, and explain whether the conclusion changes because of new data, a new reference model, or a new decision purpose. Finish by naming the unit, tolerance, and evidence needed for transfer. Record one graph or table that would expose failure hidden by the summary.