Research Methodology & Data Standards
Purpose of This Methodology
This page describes the current StatRanker standard for building, reviewing, and updating ranking datasets. The goal is reproducibility: a reader should be able to understand what the metric means, where each value came from, which year it represents, and whether StatRanker calculated or projected anything.
1. Source Roles
Sources are classified by what they contribute to the research:
- numeric_source — supplies the value used in the ranking.
- validation_source — independently checks or corroborates a value, ranking order, date, or scope.
- methodology_source — defines the metric, denominator, unit, coverage, or calculation method.
- context_source — supports interpretation but is not used as the numeric basis of the ranking.
Different URLs from the same publishing organization are treated as one source organization rather than as multiple independent sources.
2. Source Hierarchy
When sources conflict and the same metric, unit, scope, and year are being compared, StatRanker generally gives priority in this order:
- Official publisher of the metric or official statistical source.
- International organization.
- Government agency, regulator, or official administrative database.
- Academic or established industry database.
- Commercial research provider or other secondary publisher.
Conflicting values are not averaged by default. A conflict should be resolved by source quality, metric definition, year, scope, and publication methodology.
3. Value Statuses
official_value
A value published directly by an official source for the stated metric and year.
industry_observed_value
A published observed value from an established industry or commercial data provider. It is not described as an official government statistic.
calculated_value
A StatRanker calculation derived from confirmed observed inputs. It is not a forecast. The page should disclose the inputs, years, sources, formula, assumptions, and a method note.
official_forecast
A forecast or projection published directly by the authoritative source. The forecast year and publishing source must be explicit.
modeled_projection
A StatRanker projection based on a confirmed base value plus a published growth rate, trend, parameter, or scenario. The page must disclose the base year, base value, base source, projection parameter and source, formula, assumptions, and method note.
4. Ranking Construction
- Rank is generated from the raw numeric value, not from a rounded display value.
- Higher-is-better or lower-is-better direction must be defined before ranking.
- Top N must equal the number of confirmed comparable rows shown.
- A Top 100 ranking must contain 100 supported rows.
- Rows without a source, year, method, or status should not be included.
- Different units or scopes are not mixed without an explicit note and a valid normalization method.
5. Comparability Rules
Values are compared only when the metric is materially comparable. We do not silently mix:
- different measurement years;
- nominal and purchasing-power-adjusted values;
- current and constant currency values;
- counts and rates;
- absolute and per-capita measures;
- city, metro, country, state, or other geographic scopes.
When mixed years are unavoidable because the source publishes the latest available value by entity, the page should say so and disclose the year at row level where practical.
6. Snapshots, Publication Years, and Forecasts
The year in a page title can describe the page snapshot rather than the measurement year only when that distinction is explicit. For example, “2026 snapshot based on 2024 official data” means the page was prepared for 2026 using the latest available 2024 observations.
A calculated value based on observed inputs is not called an official forecast. A StatRanker projection is not presented as an official value.
7. Calculated and Derived Metrics
Derived statistics such as averages, medians, gaps, concentration ratios, rates, or per-capita values must be reproducible from disclosed inputs. A derived metric should identify its numerator, denominator, scope, formula, and relevant assumptions.
StatRanker does not claim a “global share” unless a valid global denominator from the same scope and compatible period is available.
8. Missing, Stale, and Outlier Values
Missing values are not automatically filled with proxies. Stale values may be included only when the page explains that they are the latest available values and the comparison remains meaningful. Material outliers should be checked against the source and, where useful, validated with another authoritative source.
9. Research Added Value
A ranking should provide more than a sorted copy of a source table. When the data supports it, StatRanker adds reproducible analysis such as gaps between ranks, median or average, concentration, regional or category distribution, freshness checks, historical comparison, cross-source validation, or conflict and outlier analysis.
Interpretive claims about why a pattern exists require an authoritative context source. If that support is not available, we describe the observed pattern without presenting an unsupported causal explanation.
10. Corrections and Legacy Content
Older StatRanker pages may be reviewed against this current standard. If a page contains unsupported proxy rows, unclear formulas, missing source years, or a ranking length that exceeds the confirmed dataset, the preferred correction is to rebuild the page from confirmed data. When that cannot be done immediately, the page may be reduced to the confirmed rows or temporarily marked noindex while the dataset is rebuilt.
URLs, slugs, permalinks, and canonicals are not changed as part of a content correction unless there is a separate editorial reason to change the page address.
11. What a Ranking Does Not Prove
A ranking orders entities by the stated metric. It does not automatically prove overall quality, causation, desirability, safety, policy effectiveness, or future performance. Those interpretations require separate evidence.
12. Transparency on Each Ranking Page
Where the current template is used, a ranking page should identify the metric, unit, direction, snapshot or target year, source type, coverage, value statuses, retrieval or review date, methodology, limitations, and source links. Compiled datasets should disclose that they combine multiple publishing organizations and provide row-level source or method notes where applicable.
About StatRanker
For the publication mission, editorial independence, and update policy, see About StatRanker.
