Population and coverage
Check ages, regions, sectors, exclusions and collection quality. A national rate may not describe a local area.
Record to retain · Population and scope
GUIDE / INDICATORS
Compare public indicators without false equivalence
Two figures with the same name may measure different populations, periods or methods. Comparison starts with metadata, not a chart.
01 / GUIDE / INDICATORS
Frame the comparison before opening data
State the question and intended decision: trend, regional gap or operational choice. Set statistical unit, period and denominator. Then read each producer’s definitions and any series breaks.
GuidesFrom observation to decision
The ranking reverses when the denominator changes
Two territories publish event counts using the same period and definition. A has 200 events for 10 000 residents; B has 300 for 30 000.
Fictional learning case: counts and dates illustrate the method and do not describe a real assignment.
Scale from 0 to 25 events per 1 000 residents; unadjusted crude rates.
- Territory A20 / 1 000
- Territory B10 / 1 000
| Observation | Useful check | Next action |
|---|---|---|
| Raw count | B: 300; A: 200. | B has more events within this scope. |
| Rate per 1 000 residents | A: 200 / 10 000 × 1 000 = 20; B: 10. | A has the higher crude per-resident rate. |
| Interpreting the comparison | Check age, exposure and population composition. | Do not turn this crude rate into an adjusted individual risk measure. |
Your decision — Is it possible to conclude that every A resident has twice the risk?
Explained answer
No. A crude territorial rate adjusts neither exposure nor population structure. It describes the observed ratio in scope, not every person’s situation.
Further reading: Eurostat statistical learning resources. The case figures and calculations are original fictional examples, not Eurostat data. Eurostat — Statistics for beginners
02 / GUIDE / INDICATORS
Five comparability conditions
If one cannot be aligned, present the series side by side without a misleading ranking.
Concept definition
Read exact concepts and changes: reported employment, estimated employment and participation differ.
Record to retain · Definition and methodUnit and denominator
Convert only with an explicit rule. A percentage, count and per-capita value are not directly comparable.
Record to retain · Unit and formulaTime and version
Separate observation year, publication, extraction and revision. Retain the series used.
Record to retain · Code, date and versionUncertainty and interpretation
Find margins, missing values, modelled estimates and breaks. Correlation does not establish cause.
Record to retain · Limits and assumptions03 / GUIDE / INDICATORS
Four steps to a traceable comparison
The method values explainable differences over dramatic rankings. If populations, periods or definitions differ, do not immediately calculate a difference. Show both values with their scope, then explain what must be harmonised. A missing value remains missing: replacing it with zero changes the meaning of the comparison.
- 01
Fix the question
Record populations, years and intended decision before choosing portals.
- 02
Export series
Keep identifiers, units, metadata, files and extraction dates.
- 03
Normalize carefully
Align series only after checking definitions and document every calculation.
- 04
Present the gap
Show value, source, limitation and alternative interpretations together.
04 / GUIDE / INDICATORS
Three distinct uses
Different decisions require different checks.
Compare two countries
Check population and method behind international indicators before stating a gap.
Read the dossierTrack a trend
Identify rebasing, revisions and missing periods before announcing acceleration.
Read the dossierCite a figure in a dossier
Connect the claim to the exact series, download and evidentiary limit.
Read the dossierA useful reference supports another check
Cite the precise page and version read. A portal title, search engine or access date does not replace the document date. Missing information stays missing; retain the uncertainty in your record.
Compose a public-source reference05 / GUIDE / INDICATORS
Method references
06 / FAQ
Comparison questions
Why do portals show different figures?
They may use different years, units, revisions or methods, or republish one producer’s data.
Can a missing year be filled?
Interpolation is an estimate. Label and explain it; never present it as an observation.
Is a ranking always useful?
Only when metric, scope and uncertainty permit an honest comparison for the decision.
Can percentages based on different population sizes be aggregated?
Recover numerators, denominators, populations and periods. A simple average of percentages can hide very different group sizes. Aggregate only compatible scopes and explain weighting, exclusions and remaining unknown data.
07 / Discuss a context
Compare what can be compared
A metadata table and a few counterchecks support stronger conclusions than isolated figures.
DOCUMENTED REFERENCE
World Bank GDP: six codes with different price bases
The six series names were checked in the World Bank API on 22 October 2026. Current dollars, constant dollars, local currency and international dollars are not interchangeable ways to state one amount. Choose the question before the series.
| Reference | Value or form | What it identifies |
|---|---|---|
| NY.GDP.MKTP.CD | USD | GDP in current US dollars. Change may combine output, prices and currency conversion rather than volume growth alone. |
| NY.GDP.MKTP.KD | USD · 2015 | GDP in constant 2015 US dollars in the checked metadata. Preserve the reference year instead of treating values as current prices. |
| NY.GDP.MKTP.CN | LCU | GDP in current local currency. LCU does not make different countries’ currencies identical; retain currency name and scale. |
| NY.GDP.MKTP.KN | LCU · constant | GDP in constant local currency. Check national reference and method; LCU values from different countries are not directly comparable common units. |
| NY.GDP.MKTP.PP.CD | int$ | GDP in current international dollars at purchasing power parity. PPP adjusts spatial price differences, not a payment exchange rate. |
| NY.GDP.MKTP.PP.KD | int$ · 2021 | GDP in constant 2021 international dollars. The checked PPP base year differs from 2015 in the US-dollar KD series. |
Implication for the record
Retain identifier, unit, price base, period and extraction vintage. Base or PPP-program changes may revise history. Do not mix totals and per-capita series or read total GDP as income distribution.
Must the largest-GDP country have the wealthiest inhabitants?
No. Population also affects totals. Per-capita measures need a compatible denominator and remain averages, not household situations or income distributions.
Primary documentation
- World Bank · GDP metadata, constant 2015 US$
- World Bank · GDP, PPP, constant 2021 international dollars
- World Bank · GDP metadata, constant LCU
Documentation consulted on .
DOCUMENTED REFERENCE
NEET: check age, education and subsampling
The acronym alone does not define a population. Metadata for the 2024 EU-LFS youth module, certified 5 December 2025, specify scope and differences from regular series. This describes that module, not a universal 15–34 youth definition.
| Reference | Value or form | What it identifies |
|---|---|---|
| NEET | Not employed + no education/training | Both conditions apply: not employed and not in education or training. People may be unemployed or outside the labour force. |
| Window | 4 weeks | No formal or non-formal education or training during the four weeks ending with the reference week. Annual school registration alone cannot reconstruct the variable. |
| Age | 15–34 · lfso_24 | The 2024 module targets ages 15–34. Individual tables may further select subgroups; record actual ages. |
| Denominator | Population in scope | The rate uses the same-age, same-sex population selected by the table, not just young labour-force participants. |
| Non-response | Participation question | Metadata exclude non-response to the regular formal education participation question from the denominator. A raw population count is not a substitute. |
| Sampling | Module ≠ regular series | The module uses subsampling. Eurostat notes its NEET rates can differ from regular series edat_lfse_23; differences are not automatically calculation errors. |
Implication for the record
Keep table code, age, sex, education window and missing-response treatment. NEET, youth unemployment and school dropout have different conditions and denominators. Establish matching scope before comparing publications.
Can a 15–29 NEET rate be directly compared with the 15–34 module?
No. Align ages and survey design, then check education definitions and subsampling. A rate difference under changed coverage does not establish social change.
Primary documentation
Documentation consulted on .
DOCUMENTED REFERENCE
Weights and missing values: preserve value-weight pairs
State whether weights represent frequency, population or another convention. A formula alone does not make data representative. R weighted.mean() documentation clarifies missing-value behavior for that function, not every library.
| Reference | Value or form | What it identifies |
|---|---|---|
| Weighted mean | Σ(wᵢxᵢ) / Σwᵢ | Pair values and weights before operations. Sorting one column alone can yield plausible but misassigned results. |
| Equal weights | Arithmetic mean | Equal weights with nonzero sum recover the arithmetic mean. Group means usually require appropriate group sizes or weights. |
| Scaling | w → c × w | A common positive multiplier leaves the ratio unchanged. Weight totals alone do not establish observed headcount. |
| na.rm | Missing x values | In documented R behavior na.rm removes missing x before calculation; matching weights must follow the same rows. |
| Missing w | Missing result in R default | Missing weights are not specially handled by R’s default method and produce missing results. na.rm is not general weight repair. |
| Zero weight | Omitted from weighted sum | Zero-weight values are omitted from the weighted sum. This does not mean the observation was measured as zero. |
Implication for the record
For non-negative population weights require a finite positive sum. Zero or infinite totals can be undefined. Retain initial count, missing values, missing weights and retained total. Exclusion does not automatically solve nonresponse bias.
May missing weights be replaced by one?
Only with methodological justification. It imposes a new convention and can change represented population; document and compare valid-pair results.
Primary documentation
Documentation consulted on .
DOCUMENTED REFERENCE
Living standards: consumption units are not headcount
Insee and Eurostat describe the modified OECD scale for differing household composition. It divides disposable income by an equivalent size, accounting for conventional scale economies. This differs from survey weighting and per-person division.
| Reference | Value or form | What it identifies |
|---|---|---|
| First adult | 1.0 UC | First adult counts one unit, not a share of ownership or a wage-based coefficient. |
| Other person ≥ 14 | 0.5 UC | Each additional person aged fourteen or over counts half a unit. Exactly fourteen is included. |
| Child < 14 | 0.3 UC | Each child under fourteen counts three tenths. A comparison convention, not every individual’s actual needs. |
| Example composition | 1 + 0.5 + 0.3 = 1.8 UC | Composition example, not a household observation: two adults and a child under fourteen total 1.8 units for three people. |
| Equivalised income | Disposable household income / UC | Numerator is household disposable income under the system, not automatically an individual gross wage. |
| Allocation | Same value for each household member | Assign the equivalised value to each person. Statistical assignment does not prove equal access to household resources. |
Implication for the record
Distinguish person, household and consumption unit. Across countries check income period, coverage, currency and price adjustment. Household and person distributions can differ even from identical household incomes.
Can living standards times headcount recover household income?
Not with this scale. Multiply by household consumption units under the same definition. Three people do not automatically mean three units.
Primary documentation
Documentation consulted on .
DOCUMENTED REFERENCE
Connectivity: distinguish people, fixed subscriptions and mobile lines
The extract covers 2021–2023 for France, Finland, Brazil and India. It contains 36 telecommunications observations and 12 reference populations. These four countries illustrate definitions and methods without forming a ranking or a representative global sample.
| Reference | Value or form | What it identifies |
|---|---|---|
| IT.NET.USER.ZS | % · people | The ITU series distributed by the World Bank describes people who used the Internet from any location or device in the last three months. It does not count accounts or use of an AI service. |
| IT.NET.BBND.P2 | subscriptions / 100 people | The ratio counts fixed subscriptions per 100 people and includes residential and organisational subscriptions. The published definition uses a downstream threshold of 256 kbit/s; the category does not measure delivered speed. |
| IT.CEL.SETS.P2 | subscriptions / 100 people | The series counts cellular subscriptions offering voice, including active prepaid accounts under its definition. Data-only cards are among the exclusions: more than 100 subscriptions per 100 people does not mean more than 100% of people are equipped. |
| SP.POP.TOTL | midyear population | WDI total population is a midyear estimate. It is retained separately to document weights; its presence does not guarantee the same survey population as the Internet-use indicator. |
Explore three series and their units
Local selections and calculations; no World Bank request during use.
Frozen example displayed; tables remain readable without JavaScript.
The chart shows two decimals; the table preserves received digits. Mean calculations are rounded to six decimals after calculation. These decimals do not measure uncertainty.
| Country | Year | Individuals using the Internet · % | Fixed broadband subscriptions · /100 | Cellular subscriptions offering voice · /100 | WDI population |
|---|---|---|---|---|---|
| France · FRA | 2021 | 86.09548575 | 47.6442336298287 | 113.952719366702 | 67842811 |
| France · FRA | 2022 | 85.33331734 | 48.250226559098 | 115.950519429629 | 68184457 |
| France · FRA | 2023 | 86.83635789 | 48.6688340139444 | 116.687800394775 | 68372286 |
| Finland · FIN | 2021 | 92.80806082 | 33.6396818306917 | 129.036333202492 | 5541017 |
| Finland · FIN | 2022 | 92.98879471 | 34.2951599474189 | 128.023293416281 | 5556106 |
| Finland · FIN | 2023 | 93.51385075 | 35.3318092510781 | 127.47302579722 | 5583911 |
| Brazil · BRA | 2021 | 80.68989314 | 19.8363162401481 | 104.824727184587 | 209550294 |
| Brazil · BRA | 2022 | 80.52775094 | 20.9534469027015 | 101.245468903076 | 210306415 |
| Brazil · BRA | 2023 | 84.15060238 | 22.4192301618889 | 101.020256020808 | 211140729 |
| India · IND | 2021 | 49.2580986 | 1.9487996093033 | 81.603974311212 | 1414203896 |
| India · IND | 2022 | 55.9 | 2.35231555917724 | 80.1818007717416 | 1425423212 |
| India · IND | 2023 | 60.25230026 | 2.74534696441771 | 80.5631998772888 | 1438069596 |
Telecommunications indicators originate with the International Telecommunication Union (ITU). Distribution: World Bank, World Development Indicators. Population: WDI sources detailed in the documentation. Extract and presentation calculations prepared for this site.
Update reported by the API : · Extract accessed on :
Data reused under CC BY 4.0.
Implication for the record
These figures are a frozen extract, not the latest available values or a measure of this site’s traffic. Retain code, year, documented unit and extraction date. The API has an empty unit field and an empty obs_status field here: this does not replace reading definitions or certify national comparability.
Primary documentation
- World Bank / ITU · Individuals using the Internet
- World Bank / ITU · Fixed broadband subscriptions
- World Bank / ITU · Mobile cellular subscriptions
- World Bank · Population, total
Documentation consulted on .
DOCUMENTED REFERENCE
Four countries: see what population weighting changes
For the four selected countries in 2023, the simple mean of Internet-use shares is 81.188278%. The scenario using WDI populations as weights gives 64.343169%. Both are calculated here from published values, not copied from an official ITU or World Bank aggregate.
| Reference | Value or form | What it identifies |
|---|---|---|
| Equal weights | 81.188278 % | The simple mean gives equal weight to France, Finland, Brazil and India. It describes the mean of four national rates. |
| WDI population weights | 64.343169 % | The weighted scenario multiplies each rate by the WDI population of the same country and year, then divides by the sum of selected populations. |
| Σ population · 2023 | 1723166522 | Total weights are 1,723,166,522 in this 2023 scenario. They come from population estimates, not the combined size of Internet-use surveys. |
| Difference | 16.845109 pp | The difference between calculations is 16.845109 percentage points. It reflects the weighting choice here, not an error margin or a measure of individual digital exclusion. |
Change the group and its weights
Illustrative scenario using published rates and same-year WDI populations; this is not an official aggregate. The four countries do not represent the world. No missing value is replaced by zero.
Local selections and calculations; no World Bank request during use.
Frozen example displayed; tables remain readable without JavaScript.
- Number of countries
- 4
- Sum of WDI population weights
- 1723166522
- Simple mean of rates · %
- 81.188278
- Weighted-scenario mean · %
- 64.343169
- Simple − weighted · percentage points
- 16.845109
- Lowest group rate · %
- 60.252300
- Highest group rate · %
- 93.513851
| Country | Individuals using the Internet · % | WDI population | Share of total weight · % |
|---|---|---|---|
| France | 86.83635789 | 68372286 | 3.967828 |
| Finland | 93.51385075 | 5583911 | 0.324049 |
| Brazil | 84.15060238 | 211140729 | 12.253066 |
| India | 60.25230026 | 1438069596 | 83.455057 |
simple = Σrᵢ / n weighted = Σ(Pᵢ × rᵢ) / ΣPᵢ gap = simple − weighted
The chart shows two decimals; the table preserves received digits. Mean calculations are rounded to six decimals after calculation. These decimals do not measure uncertainty.
Telecommunications indicators originate with the International Telecommunication Union (ITU). Distribution: World Bank, World Development Indicators. Population: WDI sources detailed in the documentation. Extract and presentation calculations prepared for this site.
Update reported by the API : · Extract accessed on :
Data reused under CC BY 4.0.
Implication for the record
This grouping is illustrative: usage rates and WDI populations may have different national bases and methods. It does not yield an official world rate or a verified count of Internet users. Retain the same country group and document weights when comparing years; missing data block this calculation rather than silently removing a country.
Primary documentation
Documentation consulted on .
Related checks
- Statistical adjustment: SA, CA and SCA are not interchangeable
- HICP 2026: distinguish reference-period and classification changes
- Real CPI data: rebase without inventing living-standard comparisons
- Six labour ratios: changing the denominator changes the question
- Macro, micro and support weighting: five report cases
- Population density: check the area in the denominator
- Choose a digital measure for the question being asked
- A corrected version and a new period are different relationships
DOCUMENTED REFERENCE
Read a real series: Internet use, 2020–2024
These observations compare percentage-point and relative changes in a published series. The three countries are selected for reading, not as a digital-performance ranking.
The indicator measures the share of the population using the Internet in the previous three months. It measures neither connection time, access quality nor AI-assistant use.
IT.NET.USER.ZS · Extracted · Latest update reported by the API .
| Country | 2020 | 2021 | 2022 | 2023 | 2024 |
|---|---|---|---|---|---|
| France | 84.71 | 86.10 | 85.33 | 86.84 | 88.65 |
| Germany | 89.81 | 91.43 | 91.63 | 92.48 | 93.50 |
| Spain | 93.21 | 93.90 | 94.49 | 95.45 | 95.76 |
France: from 84.71% in 2020 to 88.65% in 2024; 3.95 percentage points, or 4.66% relative to the starting value.
Table values are rounded to two decimals; calculations use the received values before that rounding. Examine revisions and collection differences before making a causal explanation.
Source: ITU, distributed by the World Bank. Extract and chart by Alexis Roux ZG. CC BY 4.0
DEEPER READING / PRACTICE
Align definitions before comparing figures.
Two indicators have neighboring names but are based on different populations or methods. The comparison begins with definitions, periods and units.
Updated on .
Priority checks
Check the numerator, denominator, territory, period and method. Identify the revisions and provisional series, and then decide whether a direct comparison is defensive.
Interpretation pitfalls
The same text can cover several measures. Converting the units does not fix a difference in definition or coverage.
Working record to retain
The comparability matrix documents indicators, definitions, conversions and reserves. It explains why some numbers can be closer and others must remain separate.
PRACTICAL CASE / WORKED EXAMPLE
Two rates relate to different denominators
The pedagogical example concerns the preparation or interpretation of data. It shows why a technically possible calculation should remain linked to verified definitions, versions and transformation rules.
Example context
Two publications use the same word for an indicator. One reports the cases to the total population, the other to a subset selected according to its method.
Suggested approach
Release numerator, denominator, period and exclusions. Compare definitions and versions. To determine whether a transformation is possible or whether the values must be presented separately.
Conclusion and next step
The matrix explains the difference in definition. Apparent ranking does not become a conclusion on territories or organizations without a truly comparable basis.
Questions for deeper practice
Does the same unit guarantee comparability?
Not to. Two percent can rely on different populations and rules. The unit must be examined with the definition, coverage, period and calculation conditions.
DETAILED WORKSHOP / APPLICATION
Building a comparability matrix before classification
01 / Prepare the necessary information
Assemble values, definitions, sources and periods. Identify numerators, denominators, units, territories and versions required for comparison.
02 / Carry out the approach
Align the dimensions and raise the differences. Examine possible conversions and their documentation. Separate directly comparable indicators from those that require a separate reading.
03 / Examine an alternative situation
If a unit conversion leaves a population difference, keep the reserve. Digital harmonization does not automatically make measures equivalent.
04 / Decide on the next step
The comparison is defensive when its conditions are explicit. Non-comparable indicators are not integrated into a ranking that would hide their definition difference.
05 / Hand over the verification records
Transmit matrix, original values and rules. New numbers can be added with the same control rather than according to the only similarity of the titles.
ANNOTATED WORKING DOCUMENT / REVIEW GRID
A matrix that exposes the conditions of comparison
The matrix combines value, definition, period, population and method. It shows what conversions are possible and what differences prevent direct comparison.
On a small screen, scroll the grid horizontally to read all three columns.
| Observation | Supported interpretation | Next step |
|---|---|---|
| Numerator and denominator describe the same scope. | The comparison can be done according to the other documented conditions. | Keep sources, versions and method reserves. |
| Two percent are based on different populations. | Common unity is not enough to make the results equal. | Separate measurements or search for a compatible base. |
| The ranking ignores a series review. | Values can mix different documentary states. | Align the versions or expressly announce the difference. |
Reviewing the working document
The final verification examines the definitions before the digital differences. Non-comparable indicators should not be incorporated into a synthesis that would make their difference in population or method disappear.
REPORTING / EXAMPLE WORDING
Presentation of results: Align definitions before comparing figures
Point to clarify
A common unity does not guarantee that populations and definitions are equal.
Before handing over
Check the numerator, denominator, period and version before the ranking.
Example of a qualified conclusion
The indicators are approached according to the dimensions aligned in the matrix. Non-comparable values remain presented with their own definition.