How Secondata finds and evaluates research

Secondata searches publicly available sources in real time and scores each one against category-specific quality criteria. Here is how that works.

Source discovery

Each search covers a broad set of source types retrieved fresh in real time — results are not cached or reused between searches.

• Academic databases: CrossRef (100M+ scholarly works) and Semantic Scholar (200M+ papers)
• Industry research: consulting firm publications, market research reports, trade press
• Practitioner sources: professional associations, think tanks, policy institutes
• Market data: public statistical sources, government data

Because sources are retrieved in real time, each report reflects what is publicly available at the moment of the search, not a fixed snapshot from a prior index.

Source verification

Academic sources are verified against CrossRef and Semantic Scholar APIs to confirm:

• The source exists and is indexed
• Author, title, year, and DOI match
• The source is genuinely peer-reviewed or formally published

Industry and practitioner sources are not verifiable via academic APIs. They are evaluated by the AI based on publisher credibility, methodology transparency, and recency. Verified and unverified status is shown explicitly for each source in the report.

Quality scoring

Each source receives a quality score from 0–100 based on category-specific criteria:

• Academic sources: Methodology rigor, sample size, peer review status, replication potential, and independence from commercial interests. Meta-analyses and RCTs score highest.
• Industry sources: Publisher credibility, sample size transparency, methodology disclosure, and recency. Practitioner sources are capped at 75 to reflect lower methodological rigor compared to peer-reviewed research.
• Recency weighting: Recent sources receive a small boost to reflect current market conditions — particularly relevant for fast-moving topics such as technology, consumer behaviour, or regulatory environments.

Evidence certainty and decision confidence

Evidence certainty (Strong / Moderate / Limited) reflects the collective strength and consistency of the evidence base — not any single source. It accounts for methodological quality across all retrieved sources, degree of agreement or contradiction between findings, and how well the evidence covers the specific question asked.

Decision confidence (High / Medium / Low) translates evidence certainty into decision relevance. It accounts for question specificity, sector evidence availability, and how identified gaps may materially affect the specific decision being evaluated.

Limitations

• Secondata searches publicly available sources. Proprietary datasets, paywalled research, and unpublished studies are not included.
• AI-assisted summaries should be verified against primary sources for high-stakes decisions.
• Source quality reflects publicly available information at the time of the search.

Questions about methodology? Contact us at hello@secondata.ai