From Institutional Observation to Reproducible Citizen Auditing: Consolidated Operational Methodology and Evidence-Based Implementation Protocol

This publication presents a consolidated operational methodology for conducting reproducible citizen auditing through longitudinal observation of institutions, digital systems, and AI-related activities. Developed from documented independent research practice, the methodology transforms institutional communications, public records, digital interactions, technical observations, and archival materials into structured, traceable, and verifiable research evidence. The protocol establishes a nine-stage research workflow: OBSERVE → RECORD → PUBLISH → NOTIFY → ACKNOWLEDGE → RESPOND → VERIFY → COMPARE → LEARN It introduces an evidence classification framework (E1–E8), a structured institutional response coding system, documentation and provenance requirements, bounded inference principles, and procedures for preserving research records and assessing evidence integrity. The methodology is grounded in documented implementation cases involving institutional observation, NIST/NCCoE-related research, NAIAC correspondence, the Figshare account suspension and subsequent data export, institutional and court correspondence, OpenAIRE and DataCite support interactions, and a five-case Citizen Auditing pilot. The publication also defines the relationship between operational observation, evidence preservation, institutional response analysis, open research infrastructure, and AI accountability research. Scope and limitations: This is an operational methodology developed through independent research practice. Its documented cases demonstrate practical application, not universal effectiveness. Independent external validation, systematic inter-coder reliability testing, and comparative implementation studies remain areas for further research. The protocol is intended to support researchers, educators, citizen auditors, civil society organizations, and institutional stakeholders seeking transparent, reproducible, and evidence-based approaches to institutional accountability and responsible AI.