Our Methodology

AlbertaTruthRadar monitors four major platforms: X/Twitter, TikTok, YouTube, and Reddit: using a set of 12 keyword clusters related to the Alberta independence referendum. Our automated system scans for coordinated posting patterns, unusual amplification spikes, and content that matches known disinformation templates documented by CSIS, DisinfoWatch, and the Canadian Digital Media Research Network.

A narrative enters our radar when it meets at least two of the following criteria: (1) it is amplified by accounts with behavioral indicators of inauthenticity; (2) it contains factually verifiable false claims; (3) it originates from or is amplified by state-linked media outlets; or (4) it matches known foreign influence operation templates.

Origin attribution is one of the most methodologically sensitive aspects of disinformation research. We use a multi-signal approach and apply a confidence threshold before publishing any attribution.

Russia: Attribution is based on amplification by RT, Sputnik, or Pravda; use of Storm-1516 network templates; posting patterns consistent with the Internet Research Agency playbook; and coordination with known Russian-linked Telegram channels.

China: Attribution is based on amplification by state-linked accounts documented in the Stanford Internet Observatory's China dataset; use of WeChat/Weibo cross-posting patterns; and content consistent with United Front Work Department messaging priorities.

USA (foreign networks): Attribution is based on documented coordination patterns with American media ecosystems that have amplified Alberta separation narratives; cross-referencing with public records, financial disclosures, and Canadian parliamentary testimony.

Foreign / Automated: Attribution is based on automated detection of synthetic text and video patterns, including metadata examination, stylometric analysis, and cross-platform posting velocity.

We do not publish origin attributions below 70% confidence. All attributions are reviewed by at least one human analyst before publication.

Our behavioral scoring model assigns a suspicion score (0–100) to flagged accounts based on seven weighted indicators:

A score above 70 triggers automatic flagging for human review. Accounts are only marked "Confirmed" after manual verification by a trained analyst. All flagging decisions include a full evidence trail accessible to journalists.

All fact-checks published on AlbertaTruthRadar follow a four-step verification process:

1. Claim identification: Claims are submitted by the public, journalists, or detected automatically from monitored platforms. Each claim must be specific, verifiable, and relevant to the referendum context.

2. Source research: Our analysts consult primary sources: government documents, academic research, Statistics Canada data, Elections Alberta records: before consulting secondary sources. We do not rely on other fact-checking organizations as primary sources.

3. Expert consultation: For claims requiring specialized knowledge (economics, constitutional law, energy policy), we consult subject-matter experts from Alberta universities and research institutions.

4. Editorial review: All fact-checks are reviewed by a second analyst before publication. The verifying analyst's name and credentials are displayed on each fact-check.

Verdict definitions:
- True: The claim is accurate and supported by multiple primary sources.
- False: The claim is demonstrably incorrect based on primary source evidence.
- Misleading: The claim contains factual elements but omits critical context that changes its meaning.
- Unverified: We cannot confirm or deny the claim with available evidence. We explain what we know and what remains uncertain.

AlbertaTruthRadar draws on the following data sources:

Primary monitoring: X/Twitter API, TikTok Research API, YouTube Data API, Reddit API: all accessed through legitimate developer agreements.

Reference databases: DisinfoWatch (Canadian disinformation tracking), Stanford Internet Observatory (global influence operations), EU DisinfoLab (European state-sponsored disinformation), CSIS public threat assessments.

Statistical data: Statistics Canada, Alberta Energy Regulator, Elections Alberta, Parliament of Canada Hansard records.

Academic partners: We collaborate with researchers at the University of Alberta, University of Calgary, and the Canadian Digital Media Research Network. All partnerships are disclosed and researchers operate independently.

Funding: AlbertaTruthRadar is funded by independent research grants and individual donations. We do not accept funding from political parties, governments, or organizations with a declared position on Alberta independence. Our full financial disclosures are available upon request at [email protected].

In the interest of transparency, we want to be explicit about the limits of our methodology:

- We do not take a position on whether Alberta should or should not become independent. Our mission is to ensure the debate is conducted on accurate information, not to influence the outcome.
- We do not monitor private communications. All our monitoring is limited to public posts on public platforms.
- We do not identify individuals as disinformation agents without a documented evidence trail. We flag accounts, not people.
- We do not claim perfect accuracy. Our attribution confidence scores reflect genuine uncertainty. We correct errors publicly and promptly.
- We do not suppress legitimate political speech. Advocating for Alberta independence is legal and legitimate. We only flag content that is demonstrably false, coordinated, or inauthentic.

When we make errors, we correct them publicly and promptly. Corrections are noted directly on the affected content with a timestamp and description of what changed. We do not delete incorrect content: we update it and explain the correction.

To report an error or dispute a finding, contact us at [email protected]. We commit to responding within 48 hours and publishing a correction within 5 business days if warranted.

Journalists and researchers who identify methodological concerns are encouraged to contact us directly. We welcome scrutiny.