Choose your reading route
eduKateSG · Why Mathematics? · Hájek Ratio Estimation: Normalized Weights and Stable Survey Means
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 8–14
Reproduce the example
Audit the numbers one step at a time.Route 3 · Chapters 15–21
Check assumptions
See where design and calibration matter.Route 4 · Chapters 22–28
Explore applications
Connect the method to a concrete investigation.Route 5 · Chapters 29–34
Study and verify
Practise, inspect failure cases and read further.
Full chapter index · Practice and checking · Mathematics Learning Hub
Why is mathematics important when survey weights do not add exactly to the population size? The Hájek ratio estimator divides a weighted outcome total by the weighted population count. This normalization often gives a stable estimate of a finite-population mean while making the scale of the weights cancel.
Did You Know? Multiplying every weight by the same positive constant leaves a Hájek mean unchanged. The numerator and denominator scale together.
Full chapter index
Chapters 1–7 · Understand the question
Chapters 8–14 · Reproduce the example
Chapters 15–21 · Check assumptions
Chapters 22–28 · Explore applications
Chapters 29–34 · Study and verify
Unequal-probability samples need weights, yet the sum of inverse-probability weights can fluctuate across samples. A weighted ratio estimates the mean on the sample’s represented scale.
For sampled values y_i and weights w_i=1/pi_i, the Hájek estimator is sum(w_i y_i)/sum(w_i). It is the Horvitz–Thompson outcome total divided by the Horvitz–Thompson population-size estimate.
Statistics Canada gives the Hájek mean as the estimated total divided by the estimated population count. Read the official direct-estimator section.
Suppose values are 10, 20 and 40 with weights 2, 1 and 3.
The weighted outcome total is 2×10+1×20+3×40=160.
| Value | Weight | Weighted outcome |
|---|---|---|
| 10 | 2 | 20 |
| 20 | 1 | 20 |
| 40 | 3 | 120 |
The sum of weights is 2+1+3=6. The Hájek mean is 160/6=26.6667.
| Numerator | Denominator | Weighted mean |
|---|---|---|
| 160 | 6 | 26.6667 |
The unweighted mean is (10+20+40)/3=23.3333. The difference occurs because the value 40 represents more population units under the design.
If all weights are divided by their mean 2, the normalized weights are 1, 0.5 and 1.5. The mean remains 80/3=26.6667.
Statistics Canada explains normalization as dividing weights by their subpopulation mean. Read the official weighting appendix. Normalization can help some analyses, but it does not recreate the survey design variance.
A ratio of two unbiased estimators is not generally exactly unbiased. The Hájek estimator trades a small ratio bias for normalization and often useful stability.
Under regularity conditions, numerator and denominator concentrate around their targets, so the ratio can be consistent and its bias becomes lower order.
For a total, multiply the Hájek mean by a known population size N. The result differs from the raw Horvitz–Thompson total when sum(w_i) differs from N.
For a subgroup, include only domain contributions in both numerator and denominator. Normalizing over the full sample rather than the analyzed domain can produce the wrong scale.
Base design weights are positive when inclusion probabilities are positive. Later calibration can sometimes produce nonpositive weights; then the ratio needs careful diagnostic review.
Treating normalized weights as frequency counts ignores stratification, clustering and finite-population selection. Use design-aware linearisation or replicate weights.
Approximate the ratio around its expected numerator and denominator. The resulting influence variable resembles w_i(y_i minus the estimated mean), divided by the estimated population count.
HT totals preserve exact design unbiasedness when probabilities are correct. Hájek ratios normalize by an estimated count and are commonly used for means and rates.
Hájek normalization imposes one count constraint. Calibration can impose many auxiliary constraints and uses an explicit distance from base weights.
Verify positive inclusion probabilities, inspect sum of weights, calculate both numerator and denominator, compare HT and Hájek estimates, and use the correct domain and replicate design.
Reproduce 160, 6 and 26.6667; rescale weights; enumerate tiny repeated samples; compare empirical bias and variance; and test the domain calculation.
Normalization does not fix frame errors, nonresponse bias, measurement error or unsupported population groups. A stable ratio can still target the wrong population.
Store y, pi, w, w×y and normalized w. Put numerator, denominator and ratio in separate labelled cells.
Unit-test scale invariance, calculate HT and Hájek side by side, and use replicate or linearised variance rather than ordinary weighted formulas.
Survey statisticians, epidemiologists, economists and policy analysts use normalized design-weighted means.
Build a six-unit population, assign unequal inclusion probabilities and compare HT and Hájek sampling distributions by enumeration.
Normalized weights are not automatically frequency weights, small ratio bias is not the same as invalidity, and a weighted mean does not prove causality.
CHAPTER 27 OF 34 · Explore applications
27. Connection to recent mathematics articles
The Horvitz–Thompson guide develops inverse-probability totals. The survey-domain guide explains subgroup estimation.
Students should show numerator and denominator before dividing. Parents can ask whether a published weighted average used the full sample or a defined domain.
- Reproduce 160/6.
- Normalize all weights.
- Prove scale invariance.
- Explain ratio bias.
- Choose a design-aware variance method.
### Is the Hájek mean exactly design-unbiased? Not generally.
### Does weight scale matter? Not for the point estimate.
### Does normalization fix nonresponse? No.
### Can ordinary standard errors be used? Usually not for a complex design.
Report target population, domain, inclusion probabilities, weight stages, numerator, denominator, normalization, variance method and limitations.
The weighted numerator estimates the outcome total while the weighted denominator estimates population size; their ratio estimates the mean.
Confirm the target and domain, positive weights, correct numerator and denominator, scale invariance, ratio-bias language, design-aware variance and no automatic bias-removal claim.
Continue through the Mathematics Learning Hub. Recalculate one weighted estimate, state its target population and explain the assumption that makes the adjustment useful.
Extended practice laboratory
Practice case 1: Define population
Setup: Start from the article’s Hajek-ratio mechanism for case 1. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the target population and sampling design, verify the weighting identity, state one diagnostic, record units and compare one defensible alternative with clear reasoning. Show the arithmetic, name the response and predictor units, and separate the numerical result from the claim it supports. Then vary one defensible assumption, calculate again and explain whether the conclusion changed because of the data, the fitted model or the validation purpose. Finish with one compact table or graph that could reveal a failure hidden by the summary.
Practice case 2: List values
Setup: Start from the article’s Hajek-ratio mechanism for case 2. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the target population and sampling design, verify the weighting identity, state one diagnostic, record units and compare one defensible alternative with clear reasoning. Show the arithmetic, name the response and predictor units, and separate the numerical result from the claim it supports. Then vary one defensible assumption, calculate again and explain whether the conclusion changed because of the data, the fitted model or the validation purpose. Finish with one compact table or graph that could reveal a failure hidden by the summary.
Practice case 3: List probabilities
Setup: Start from the article’s Hajek-ratio mechanism for case 3. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the target population and sampling design, verify the weighting identity, state one diagnostic, record units and compare one defensible alternative with clear reasoning. Show the arithmetic, name the response and predictor units, and separate the numerical result from the claim it supports. Then vary one defensible assumption, calculate again and explain whether the conclusion changed because of the data, the fitted model or the validation purpose. Finish with one compact table or graph that could reveal a failure hidden by the summary.
Practice case 4: Invert probabilities
Setup: Start from the article’s Hajek-ratio mechanism for case 4. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the target population and sampling design, verify the weighting identity, state one diagnostic, record units and compare one defensible alternative with clear reasoning. Show the arithmetic, name the response and predictor units, and separate the numerical result from the claim it supports. Then vary one defensible assumption, calculate again and explain whether the conclusion changed because of the data, the fitted model or the validation purpose. Finish with one compact table or graph that could reveal a failure hidden by the summary.
Practice case 5: Calculate weighted outcomes
Setup: Start from the article’s Hajek-ratio mechanism for case 5. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the target population and sampling design, verify the weighting identity, state one diagnostic, record units and compare one defensible alternative with clear reasoning. Show the arithmetic, name the response and predictor units, and separate the numerical result from the claim it supports. Then vary one defensible assumption, calculate again and explain whether the conclusion changed because of the data, the fitted model or the validation purpose. Finish with one compact table or graph that could reveal a failure hidden by the summary.
Practice case 6: Sum numerator
Setup: Start from the article’s Hajek-ratio mechanism for case 6. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the target population and sampling design, verify the weighting identity, state one diagnostic, record units and compare one defensible alternative with clear reasoning. Show the arithmetic, name the response and predictor units, and separate the numerical result from the claim it supports. Then vary one defensible assumption, calculate again and explain whether the conclusion changed because of the data, the fitted model or the validation purpose. Finish with one compact table or graph that could reveal a failure hidden by the summary.
Practice case 7: Sum denominator
Setup: Start from the article’s Hajek-ratio mechanism for case 7. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the target population and sampling design, verify the weighting identity, state one diagnostic, record units and compare one defensible alternative with clear reasoning. Show the arithmetic, name the response and predictor units, and separate the numerical result from the claim it supports. Then vary one defensible assumption, calculate again and explain whether the conclusion changed because of the data, the fitted model or the validation purpose. Finish with one compact table or graph that could reveal a failure hidden by the summary.
Practice case 8: Calculate ratio
Setup: Start from the article’s Hajek-ratio mechanism for case 8. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the target population and sampling design, verify the weighting identity, state one diagnostic, record units and compare one defensible alternative with clear reasoning. Show the arithmetic, name the response and predictor units, and separate the numerical result from the claim it supports. Then vary one defensible assumption, calculate again and explain whether the conclusion changed because of the data, the fitted model or the validation purpose. Finish with one compact table or graph that could reveal a failure hidden by the summary.
Practice case 9: Normalize weights
Setup: Start from the article’s Hajek-ratio mechanism for case 9. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the target population and sampling design, verify the weighting identity, state one diagnostic, record units and compare one defensible alternative with clear reasoning. Show the arithmetic, name the response and predictor units, and separate the numerical result from the claim it supports. Then vary one defensible assumption, calculate again and explain whether the conclusion changed because of the data, the fitted model or the validation purpose. Finish with one compact table or graph that could reveal a failure hidden by the summary.
Practice case 10: Verify invariance
Setup: Start from the article’s Hajek-ratio mechanism for case 10. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the target population and sampling design, verify the weighting identity, state one diagnostic, record units and compare one defensible alternative with clear reasoning. Show the arithmetic, name the response and predictor units, and separate the numerical result from the claim it supports. Then vary one defensible assumption, calculate again and explain whether the conclusion changed because of the data, the fitted model or the validation purpose. Finish with one compact table or graph that could reveal a failure hidden by the summary.
Practice case 11: Calculate HT total
Setup: Start from the article’s Hajek-ratio mechanism for case 11. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the target population and sampling design, verify the weighting identity, state one diagnostic, record units and compare one defensible alternative with clear reasoning. Show the arithmetic, name the response and predictor units, and separate the numerical result from the claim it supports. Then vary one defensible assumption, calculate again and explain whether the conclusion changed because of the data, the fitted model or the validation purpose. Finish with one compact table or graph that could reveal a failure hidden by the summary.
Practice case 12: Use known N
Setup: Start from the article’s Hajek-ratio mechanism for case 12. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the target population and sampling design, verify the weighting identity, state one diagnostic, record units and compare one defensible alternative with clear reasoning. Show the arithmetic, name the response and predictor units, and separate the numerical result from the claim it supports. Then vary one defensible assumption, calculate again and explain whether the conclusion changed because of the data, the fitted model or the validation purpose. Finish with one compact table or graph that could reveal a failure hidden by the summary.
Practice case 13: Create domain
Setup: Start from the article’s Hajek-ratio mechanism for case 13. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the target population and sampling design, verify the weighting identity, state one diagnostic, record units and compare one defensible alternative with clear reasoning. Show the arithmetic, name the response and predictor units, and separate the numerical result from the claim it supports. Then vary one defensible assumption, calculate again and explain whether the conclusion changed because of the data, the fitted model or the validation purpose. Finish with one compact table or graph that could reveal a failure hidden by the summary.
Practice case 14: Linearise ratio
Setup: Start from the article’s Hajek-ratio mechanism for case 14. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the target population and sampling design, verify the weighting identity, state one diagnostic, record units and compare one defensible alternative with clear reasoning. Show the arithmetic, name the response and predictor units, and separate the numerical result from the claim it supports. Then vary one defensible assumption, calculate again and explain whether the conclusion changed because of the data, the fitted model or the validation purpose. Finish with one compact table or graph that could reveal a failure hidden by the summary.
Practice case 15: Add clustering
Setup: Start from the article’s Hajek-ratio mechanism for case 15. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the target population and sampling design, verify the weighting identity, state one diagnostic, record units and compare one defensible alternative with clear reasoning. Show the arithmetic, name the response and predictor units, and separate the numerical result from the claim it supports. Then vary one defensible assumption, calculate again and explain whether the conclusion changed because of the data, the fitted model or the validation purpose. Finish with one compact table or graph that could reveal a failure hidden by the summary.
Practice case 16: Add strata
Setup: Start from the article’s Hajek-ratio mechanism for case 16. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the target population and sampling design, verify the weighting identity, state one diagnostic, record units and compare one defensible alternative with clear reasoning. Show the arithmetic, name the response and predictor units, and separate the numerical result from the claim it supports. Then vary one defensible assumption, calculate again and explain whether the conclusion changed because of the data, the fitted model or the validation purpose. Finish with one compact table or graph that could reveal a failure hidden by the summary.
Practice case 17: Use replicates
Setup: Start from the article’s Hajek-ratio mechanism for case 17. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the target population and sampling design, verify the weighting identity, state one diagnostic, record units and compare one defensible alternative with clear reasoning. Show the arithmetic, name the response and predictor units, and separate the numerical result from the claim it supports. Then vary one defensible assumption, calculate again and explain whether the conclusion changed because of the data, the fitted model or the validation purpose. Finish with one compact table or graph that could reveal a failure hidden by the summary.
Practice case 18: Compare calibration
Setup: Start from the article’s Hajek-ratio mechanism for case 18. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the target population and sampling design, verify the weighting identity, state one diagnostic, record units and compare one defensible alternative with clear reasoning. Show the arithmetic, name the response and predictor units, and separate the numerical result from the claim it supports. Then vary one defensible assumption, calculate again and explain whether the conclusion changed because of the data, the fitted model or the validation purpose. Finish with one compact table or graph that could reveal a failure hidden by the summary.
Practice case 19: Change weight scale
Setup: Start from the article’s Hajek-ratio mechanism for case 19. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the target population and sampling design, verify the weighting identity, state one diagnostic, record units and compare one defensible alternative with clear reasoning. Show the arithmetic, name the response and predictor units, and separate the numerical result from the claim it supports. Then vary one defensible assumption, calculate again and explain whether the conclusion changed because of the data, the fitted model or the validation purpose. Finish with one compact table or graph that could reveal a failure hidden by the summary.
Practice case 20: Add tiny probability
Setup: Start from the article’s Hajek-ratio mechanism for case 20. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the target population and sampling design, verify the weighting identity, state one diagnostic, record units and compare one defensible alternative with clear reasoning. Show the arithmetic, name the response and predictor units, and separate the numerical result from the claim it supports. Then vary one defensible assumption, calculate again and explain whether the conclusion changed because of the data, the fitted model or the validation purpose. Finish with one compact table or graph that could reveal a failure hidden by the summary.
Practice case 21: Audit target
Setup: Start from the article’s Hajek-ratio mechanism for case 21. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the target population and sampling design, verify the weighting identity, state one diagnostic, record units and compare one defensible alternative with clear reasoning. Show the arithmetic, name the response and predictor units, and separate the numerical result from the claim it supports. Then vary one defensible assumption, calculate again and explain whether the conclusion changed because of the data, the fitted model or the validation purpose. Finish with one compact table or graph that could reveal a failure hidden by the summary.
Practice case 22: Audit domain
Setup: Start from the article’s Hajek-ratio mechanism for case 22. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the target population and sampling design, verify the weighting identity, state one diagnostic, record units and compare one defensible alternative with clear reasoning. Show the arithmetic, name the response and predictor units, and separate the numerical result from the claim it supports. Then vary one defensible assumption, calculate again and explain whether the conclusion changed because of the data, the fitted model or the validation purpose. Finish with one compact table or graph that could reveal a failure hidden by the summary.
Practice case 23: Audit variance
Setup: Start from the article’s Hajek-ratio mechanism for case 23. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the target population and sampling design, verify the weighting identity, state one diagnostic, record units and compare one defensible alternative with clear reasoning. Show the arithmetic, name the response and predictor units, and separate the numerical result from the claim it supports. Then vary one defensible assumption, calculate again and explain whether the conclusion changed because of the data, the fitted model or the validation purpose. Finish with one compact table or graph that could reveal a failure hidden by the summary.
Practice case 24: Name limits
Setup: Start from the article’s Hajek-ratio mechanism for case 24. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the target population and sampling design, verify the weighting identity, state one diagnostic, record units and compare one defensible alternative with clear reasoning. Show the arithmetic, name the response and predictor units, and separate the numerical result from the claim it supports. Then vary one defensible assumption, calculate again and explain whether the conclusion changed because of the data, the fitted model or the validation purpose. Finish with one compact table or graph that could reveal a failure hidden by the summary.
Practice case 25: Write conclusion
Setup: Start from the article’s Hajek-ratio mechanism for case 25. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the target population and sampling design, verify the weighting identity, state one diagnostic, record units and compare one defensible alternative with clear reasoning. Show the arithmetic, name the response and predictor units, and separate the numerical result from the claim it supports. Then vary one defensible assumption, calculate again and explain whether the conclusion changed because of the data, the fitted model or the validation purpose. Finish with one compact table or graph that could reveal a failure hidden by the summary.