Research operating system
Study registry, versioned instruments, sample plans, quality review, release gates, and audit trail.
Demosight Trust Center
Research earns authority through repeatable controls, visible limits, and the willingness to publish uncomfortable corrections.
No push polls, guaranteed conclusions, disguised advocacy, or client control over findings.
Open policy →02Every public result must carry sponsor, population, fieldwork, instrument, sample, quality, weighting, uncertainty, and limitations.
Open policy →03Versioned consent, separate recontact choice, data minimization, withdrawal, access, correction, and deletion review.
Open policy →04Survey and poll labels are reserved for human-origin responses. Synthetic cases remain clearly labeled simulations.
Open policy →05Every predictive claim needs a target, version, data cutoff, uncertainty range, backtest, and observed outcome.
Open policy →06AI-assisted design, coding, explanation, and modeling stay disclosed, evaluated, and under accountable human review.
Open policy →07Material errors and changes are recorded visibly instead of silently replacing a result.
Open policy →Methodology Passport
Standards basis
The operating standard is aligned toward the AAPOR Transparency Initiative, the ESOMAR/WAPOR Guide to Opinion Polls, and ISO 20252 research-quality practice. Application security is measured against OWASP ASVS 5.0 and NIST CSF 2.0; accessibility targets WCAG 2.2 AA. Privacy operations are being prepared against U.S. requirements, Philippine law, and the National Privacy Commission’s current guidance.
This is a standards-alignment roadmap—not a claim of certification, accreditation, legal review, association membership, regulatory approval, completed penetration testing, or completed accessibility conformance testing.
Founding launch register
Study registry, versioned instruments, sample plans, quality review, release gates, and audit trail.
Direct identity is stored outside the research profile and excluded from answers, weighting data, dashboards, and research exports.
U.S.-adult, Philippine-resident, current-OFW, returnee, and wider-diaspora cohorts remain separate and explicitly nonrepresentative.
Production dependencies, route protections, session handling, identity separation, release controls, and migration integrity are tested. Independent penetration testing remains required.
Register the entity and publish accountable leadership, official address, and contact channels.
Appoint the privacy lead or DPO, complete the data-processing inventory, and confirm applicable NPC registration.
Establish external methodology, ethics, and privacy advisers with a disclosed conflicts policy.
No payments, invoices, or automatic project acceptance before registration and contract review.