6.9 Bias Controls
The methodology incorporates explicit bias controls at the research design, recruitment, data collection, and analysis stages:
-
Cardano Insider Optimism: Primary research will actively seek non-Cardano stakeholders, including those who evaluated Cardano and chose alternatives, as validation counterweights.
-
Regional Overgeneralisation: The RARF framework operates at country-cluster level, not continental level, to prevent broad geographic labels from masking intra-regional variation.
-
English-Language Bias: PRISMA’s hub network in sub-Saharan Africa includes access to participants in non-English operating environments. Where local-language primary research is required, this will be documented and the limitation noted where full coverage is not possible.
-
Elite and Capital-City Bias: The ALJ methodology deliberately surfaces grassroots and implementation-level actors who are typically absent from capital-city-focused research. This is one of the distinctive advantages of PRISMA’s approach.
-
Access Overclaiming: All counterparty access will be classified as Interviewed, Warm Access, Referred, Desk-Identified, or Hypothetical. Findings based on Hypothetical or Desk-Identified sources will be clearly marked with lower confidence levels.
-
Treating Intent as Adoption: The CASS framework is the operational control for this risk. No signal will be classified above its evidence-supported tier.
19. Risk and Bias Mitigation Plan
| Risk / Bias Type | Description | Control Measure |
|---|---|---|
| Cardano insider optimism | Over-weighting evidence from Cardano community members who have a reputational or financial interest in positive findings | Minimum of two non-Cardano respondents per priority region; negative-case evidence actively sought; CASS classification applied regardless of source affiliation |
| Regional overgeneralisation | Treating continental-level labels as analytically coherent markets | RARF operates at country-cluster level; findings documented at sub-national specificity where evidence supports it |
| English-language bias | Under-representing actors who operate in non-English environments | PRISMA hub network includes non-English operating contexts; limitations documented; local facilitators engaged where language access constraints exist |
| Elite and capital-city bias | Over-representing urban, capital-city, or institutionally visible actors at the expense of implementation-level practitioners | ALJ methodology explicitly surfaces grassroots and implementation-level actors; evidence from different actor levels cross-checked |
| Government access overclaiming | Representing desk-identified or referral-based government contacts as confirmed access | All counterparty access classified using the five-tier access status system; findings based on unconfirmed access marked with lower confidence |
| Partner self-promotion | Local partners over-reporting the strength of their institutional relationships or the depth of adoption signals | Partner claims corroborated with independent evidence; CASS classification applied to evidence regardless of source |
| Regulatory overclaiming | Presenting uncertain or evolving regulatory situations with greater certainty than the evidence supports | Three-tier regulatory certainty classification applied; uncertain findings clearly labelled; local counsel recommended where clarity is insufficient |
| Treating intent as adoption | Classifying letters of intent, expressed interest, or informal engagement as adoption signal | CASS framework is the structural control; no finding above Tier 3 can be based on expressed interest alone |
| Scope creep | Absorbing adjacent RFP scopes (stablecoin liquidity, enterprise readiness, delivery partner design) into this proposal | Cross-RFP handoff memo is a mandatory deliverable; scope discipline reviewed at each checkpoint; findings that belong in adjacent workstreams are routed there, not stretched here |
| Political and institutional sensitivity | Exposing respondents to institutional risk by publishing sensitive government or regulatory views | Data handling protocols (Section 20); explicit consent management; separation of public and confidential findings |
The final report will include a Limitations section that explicitly states what the research can and cannot support, where confidence is Low, and where further diligence is required before action.