Methodology

What makes a comparison fair?

A fair comparison preserves what each number measures. Shared units do not erase differences in water flow, time period, geography, system boundary, data origin, uncertainty, or local consequence.

A comparison should answer a bounded question. If the sources cannot support that question, BasisMark narrows the finding, changes the language, or does not publish a conclusion.

Finding and twelve checks

A finding is a bounded statement supported by evidence. It is not the same as a topic, headline, verdict, or ratio.

BasisMark evaluates twelve checks:

1. Metric alignment
2. Unit alignment
3. Time alignment
4. Geographic alignment
5. Source quality
6. Value origin
7. System boundary
8. Double-counting risk
9. Uncertainty treatment
10. Local-context relevance
11. Source currency
12. Ratio eligibility

Each check is recorded as Pass, Partial, Fail, or Not Applicable, with a short explanation. A failed check can block a ratio or require a narrower finding.

Separate judgments

• Public label: short reader-facing summary of comparison fairness.
• Technical verdict: controlled methodological classification.
• Confidence: strength of evidence for the exact bounded finding.
• Source quality: strength and relevance of individual evidence.
• Review status: what internal or external review has occurred.

These fields must remain separate. Internal review does not increase confidence automatically, and external review does not certify a result. Review the separate label, verdict, and confidence definitions.

Water-flow taxonomy

BasisMark distinguishes withdrawals, consumption, applied water, discharged water, recycled or reclaimed water, direct operational water, and indirect or supply-chain water. A value is not moved from one category to another simply because both can be converted into liters or gallons.

Direct and indirect values remain separate unless an approved lifecycle boundary explicitly supports combining them. Applied water is not automatically consumption. Reused water is still labeled and is not treated as zero impact.

Sources and local context

BasisMark prefers original reports, official datasets, standards, and direct technical documentation. News reports, summaries, and search results may help locate evidence but do not replace it.

National findings do not establish local impact. Local interpretation requires evidence about water source, watershed stress, seasonality, facility design, infrastructure, competing demand, community effects, and other place-specific conditions.

Publication and external review

Methodology v1.1 remains unchanged. It requires scoped external review before the first flagship finding is publicly released. External review means a qualified person outside the project examines a defined technical or editorial scope and discloses relevant conflicts. It does not mean peer review, certification, endorsement, or control over publication.

The first flagship has not completed scoped external review. Its page therefore remains an explicitly unfinished research project with no final public finding.