Choose your reading route
eduKateSG · Why Mathematics? · Survey Domain Estimation: Subpopulations and Correct Standard Errors
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 a survey question concerns one subgroup but the sample design covers everyone? Domain estimation keeps the full survey design in the variance calculation and uses an indicator to define the subgroup. This preserves uncertainty in how many domain members each primary sampling unit would contain across repeated samples.
Did You Know? Deleting every out-of-domain row before running complex-survey software can produce the same subgroup point estimate yet a different, often too small, standard error.
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
Analysts frequently need a mean for adults over 65, a proportion for one region or a model for students in a programme. The subgroup membership count within each sampled PSU is itself random under the design.
A domain, subpopulation or subdomain is defined by an indicator D_i. Point estimates use D_i to include the target records while variance estimation retains the design information of all eligible sampled records.
CDC states that software requires all observations with nonzero analysis weight plus a subpopulation indicator; physically subsetting can treat domain sample size as fixed and underestimate variance. Read the NHANES module.
Estimate a domain total as the weighted sum of D_i y_i. Records outside the domain contribute zero to the total but still help preserve PSU and stratum structure for variance estimation.
Estimate the mean as sum(w_i D_i y_i) divided by sum(w_i D_i). Both numerator and denominator are random weighted totals.
Suppose weighted domain outcome total is 360 and weighted domain population total is 120. The domain mean is 360/120=3.
| Weighted outcome total | Weighted domain total | Mean |
|---|---|---|
| 360 | 120 | 3 |
For ratio X/Y at X=360 and Y=120, derivatives are 1/120≈0.008333 and −360/120²=−0.025. These combine numerator and domain-size uncertainty.
| dX | dY |
|---|---|
| 0.008333 | -0.025 |
A stratum may appear to have one domain PSU after subsetting even though the full sample has two PSUs. The deleted PSU's zero domain contribution is part of the repeated-sampling contrast, not missing design metadata.
Survey software usually offers a domain or subpopulation option. It keeps the design object intact while restricting the estimand, which differs from filtering the data first.
For each replicate, recompute the domain numerator and denominator using the domain indicator and replicate weights. Empty-domain replicates require a documented rule and stability warning.
A domain may have many respondent rows but few contributing PSUs. Precision and degrees of freedom depend on the design support, not the raw subgroup count alone.
If no sampled domain member appears in a stratum or replicate, that is a data and design fact. Adding arbitrary pseudo-observations changes the estimand and requires a formal model.
Combining age, region and programme indicators can create thin cells quickly. Pre-specify important domains and report multiplicity and reliability limits.
A model fitted only to a domain can still use full-design variance information through domain-aware software. The scientific target must clarify whether effects are descriptive within the domain or causal.
Domain membership and outcome availability are different indicators. Missing-data treatment can change weights, denominators and the target population.
Domain estimation asks about a subgroup under an existing design. Post-stratification or calibration changes weights to match control totals; it is a weighting adjustment, not merely a subgroup filter.
A domain estimate from one national survey shares the original design and weights. It is not equivalent to a new simple random survey conducted only inside the subgroup.
A domain's nominal n can overstate information when members cluster in few PSUs or carry unequal weights. Domain-specific design effects provide a more relevant precision summary.
Count domain respondents, contributing PSUs and strata; inspect empty and lonely domain cells; compare domain-command and naive-filter standard errors; verify the correct analysis weight.
Reproduce the domain mean 3, compare full-design indicator analysis with physical filtering, test Taylor and replicate methods and simulate random domain counts within PSUs.
Design-correct variance does not solve sparse-data bias, multiple testing, nonresponse, misclassification of domain membership or unsupported causal conclusions.
Keep all rows, create a zero-one domain indicator and calculate weighted domain numerator, denominator, mean, contributing PSUs and design-based standard error.
Build the full survey design first, call a domain estimator, unit-test 360/120=3 and compare with a deliberately filtered analysis to expose variance differences.
Health, education, labour and social-policy analysts use domain estimation for trustworthy subgroup reporting.
Simulate two PSUs per stratum with random domain membership. Compare standard errors from full-design indicators and row filtering over many samples.
Out-of-domain rows are not useless to variance estimation, a large subgroup row count need not mean many PSUs, and a domain result is not automatically causal.
CHAPTER 27 OF 34 · Explore applications
27. Connection to recent mathematics articles
The design-effect guide explains domain-specific information. The comparing-percentages guide helps keep subgroup denominators visible.
Students should write the indicator inside both numerator and denominator. Parents can ask how many schools, clinics or neighbourhoods actually support a subgroup estimate.
- Calculate the domain mean 3.
- Derive two ratio gradients.
- Explain zero contributions.
- Compare filtering and domain commands.
- Count contributing PSUs.
### Should out-of-domain rows be deleted? Usually not before design-based variance estimation.
### Is domain n fixed? Not under repeated sampling.
### Can a replicate domain be empty? Yes.
### Does correct variance remove bias? No.
Report domain definition, universe, weight, nominal n, contributing PSUs and strata, estimator, variance method, degrees of freedom, empty replicates and reliability rule.
An out-of-domain record contributes zero to the domain total while its PSU membership preserves the structure needed to measure how domain composition varies across samples.
Confirm full-design retention, indicator-based estimation, numerator and denominator recomputation, contributing PSU counts, empty-cell handling and no causal overreach.
Continue through the Mathematics Learning Hub. Recalculate one variance or effective-sample-size example, then explain which design feature changes the answer.
Extended practice laboratory
Practice case 1: Define universe
Setup: Start from the article’s survey-domain-estimation mechanism for case 1. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the sampling design, use the documented variance rule, 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: Define indicator
Setup: Start from the article’s survey-domain-estimation mechanism for case 2. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the sampling design, use the documented variance rule, 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: Keep full design
Setup: Start from the article’s survey-domain-estimation mechanism for case 3. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the sampling design, use the documented variance rule, 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: Calculate domain numerator
Setup: Start from the article’s survey-domain-estimation mechanism for case 4. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the sampling design, use the documented variance rule, 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 denominator
Setup: Start from the article’s survey-domain-estimation mechanism for case 5. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the sampling design, use the documented variance rule, 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: Calculate mean
Setup: Start from the article’s survey-domain-estimation mechanism for case 6. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the sampling design, use the documented variance rule, 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: Derive gradients
Setup: Start from the article’s survey-domain-estimation mechanism for case 7. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the sampling design, use the documented variance rule, 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: Count PSUs
Setup: Start from the article’s survey-domain-estimation mechanism for case 8. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the sampling design, use the documented variance rule, 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: Count strata
Setup: Start from the article’s survey-domain-estimation mechanism for case 9. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the sampling design, use the documented variance rule, 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: Use Taylor method
Setup: Start from the article’s survey-domain-estimation mechanism for case 10. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the sampling design, use the documented variance rule, 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: Use replicate weights
Setup: Start from the article’s survey-domain-estimation mechanism for case 11. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the sampling design, use the documented variance rule, 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: Check empty replicate
Setup: Start from the article’s survey-domain-estimation mechanism for case 12. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the sampling design, use the documented variance rule, 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: Use percentage
Setup: Start from the article’s survey-domain-estimation mechanism for case 13. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the sampling design, use the documented variance rule, 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: Use ratio
Setup: Start from the article’s survey-domain-estimation mechanism for case 14. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the sampling design, use the documented variance rule, 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: Use regression
Setup: Start from the article’s survey-domain-estimation mechanism for case 15. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the sampling design, use the documented variance rule, 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 missing outcomes
Setup: Start from the article’s survey-domain-estimation mechanism for case 16. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the sampling design, use the documented variance rule, 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: Cross-classify domain
Setup: Start from the article’s survey-domain-estimation mechanism for case 17. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the sampling design, use the documented variance rule, 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 filtering
Setup: Start from the article’s survey-domain-estimation mechanism for case 18. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the sampling design, use the documented variance rule, 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: Compare separate survey
Setup: Start from the article’s survey-domain-estimation mechanism for case 19. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the sampling design, use the documented variance rule, 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: Calculate domain DEFF
Setup: Start from the article’s survey-domain-estimation mechanism for case 20. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the sampling design, use the documented variance rule, 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 weight
Setup: Start from the article’s survey-domain-estimation mechanism for case 21. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the sampling design, use the documented variance rule, 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 support
Setup: Start from the article’s survey-domain-estimation mechanism for case 22. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the sampling design, use the documented variance rule, 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 failures
Setup: Start from the article’s survey-domain-estimation mechanism for case 23. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the sampling design, use the documented variance rule, 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 survey-domain-estimation mechanism for case 24. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the sampling design, use the documented variance rule, 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 survey-domain-estimation mechanism for case 25. Task: Change the named quantity, recompute the dependent result and compare with the original.
Check: Preserve the sampling design, use the documented variance rule, 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.