METHODS & STANDARDS

How do you know?

For GIRI Institute, methodology is not a technical appendix added after a conclusion. It is the documented path from a research question to a claim that another qualified observer can understand, examine, and challenge.

GIRI RESEARCH STANDARD — GRS v1.0

Defined QuestionDefined PopulationDefined EvidenceDefined MethodDocumented MeasurementQuality AssuranceTransparent Reporting

POPULATION & SAMPLE

Define what the study can actually speak about.

Target Population

The full population about which the research question is framed.

Sampling Frame

The operational list or source from which study units can be selected.

Study Sample

The units selected for intended observation.

Observed Sample

The units for which usable evidence is actually collected.

EVIDENCE CLASSES

Evidence is classified by relationship to the research object.

Direct Primary Evidence

Direct observation, measurement, or first-hand evidence collected for the study.

Authoritative Primary Sources

Official records, first-party documentation, regulatory records, or authoritative source material.

Scholarly & Institutional Research

Peer-reviewed or institutionally produced research relevant to the question.

Professional & Industry Evidence

Industry records, professional documentation, and credible practitioner evidence.

Open & Contextual Evidence

Contextual sources that may inform interpretation but require appropriate caution.

Evidence class is not a universal quality score. Reliability still depends on source, context, measurement, and use.

RESEARCH REASONING

Observation is not interpretation. Interpretation is not recommendation.

OBSERVATION

What the evidence directly shows under the defined measurement rules.

INTERPRETATION

What the observed evidence may mean within the research question and limitations.

RECOMMENDATION

What action may be appropriate based on evidence, objectives, and judgment.

MEASUREMENT HIERARCHY

Scores should be built from observable evidence.

ObservationVariableIndicatorDimensionIndex / Score

For GLVS™, observations feed assessment items and indicators, which roll into eight defined dimensions and then into the overall score.

MEASUREMENT PRINCIPLES

Six properties guide instrument quality.

Validity

Does the measure represent the concept it claims to measure?

Reliability

Can the process produce sufficiently consistent results?

Consistency

Are rules applied uniformly across comparable observations?

Transparency

Can a qualified reader understand how the measure was produced?

Comparability

Can results be compared responsibly across units, places, or time?

Reproducibility

Can another qualified researcher understand and repeat the method where practical?

INSTRUMENT LIFECYCLE

Measurement instruments should mature before they scale.

QuestionConcept DefinitionVariablesIndicatorsMeasurement RulesInstrumentPilot / CalibrationRevisionOperational VersionValidationVersion Control

CALIBRATION

Analysts should learn the same rules before independent production.

Training
Shared test observations
Independent scoring
Comparison
Disagreement analysis
Rule clarification
Reassessment
Operational approval

INTER-RATER RELIABILITY

Provisional operational guidance.

GREEN≥85% agreement

CAUTION75–84.9%

STOP / REVIEW<75%

These are provisional management bands for GIRI operations, not universal scientific thresholds. Instrument-specific reliability metrics may differ.

QUALITY ASSURANCE

Research quality should have multiple checkpoints.

QA-1

Analyst Self-Verification

The original analyst checks completeness, evidence, and rule application.

QA-2

Independent Review

A second qualified reviewer examines selected or required observations.

QA-3

Research Quality Audit

System-level review of data quality, rules, patterns, and material error.

QA-4

External Methodological Review

Independent methodological review for important studies where warranted.

DATA QUALITY MANAGEMENT

Operational quality targets make problems visible.

≥98%Target completeness

95–100%Provenance target, with critical fields near complete

<1%Duplicate-rate target

These are institutional management targets, not claims that every dataset has already achieved them.

DATA ACCESS

Five access classes.

DA-1 — Open
DA-2 — Public Summary
DA-3 — Research Access
DA-4 — Restricted
DA-5 — Internal

VERSION CONTROL

Methods and research objects change visibly.

v1.0 — original operational/released version.

v1.1 — minor non-material revision.

v2.0 — substantial revision that may affect method, comparability, or interpretation.

REPLICATION

Replication can test different dimensions of reliability.

Analyst Replication

Different analysts apply the same rules to comparable evidence.

Vertical Replication

The method is applied to another industry.

Geographic Replication

The method is applied in another Research Market.

Temporal Replication

The same or comparable units are reassessed over time.

External Replication

Independent researchers attempt to reproduce or test the work.

CORRECTIONS

Correct the research record; do not erase it.

Minor errors should be corrected and documented. Material analytical errors should receive formal correction. Invalidated findings should be amended. Fundamentally unreliable work may need withdrawal or retraction while preserving the record of what changed and why.

NEGATIVE FINDINGS

Research does not exist to produce favorable outcomes.

Negative, null, contradictory, or commercially inconvenient findings remain research findings when produced by a defensible process.

AI IN RESEARCH

Assistive, not self-validating.

AI may assist research discovery, extraction, classification, coding, quality checks, drafting, and administrative work. AI-generated output is not automatically verified evidence.

AI may assist the research process. It does not replace verification.

STOP CONDITIONS

Quality problems can stop expansion.

Research should pause when disagreement, missing evidence, sampling problems, instrument ambiguity, data-quality failures, ethics concerns, or resource constraints make the resulting claims unreliable.

Correction takes priority over growth.

RESEARCH GATES

Scale follows readiness.

GATE A

Methodology Readiness

Is the instrument sufficiently defined and calibrated for operational use?

GATE B

Field Research Reliability

Can trained analysts apply it consistently to real observations?

GATE C

Market Replication

Can the research system produce comparable work in another market or vertical?

Define carefully. Observe consistently. Measure transparently. Verify independently. Correct openly.

GIRI research should disclose enough methodology for a qualified independent researcher to understand how the findings were produced.