Research methodology: how Bestead handles evidence and future original data

Current publication type

Bestead currently publishes narrative evidence syntheses: focused editorial explanations that connect a practical question with selected research and public-health guidance. A narrative synthesis can help orient readers across different kinds of evidence. It is more vulnerable to source-selection and interpretation bias than a well-conducted systematic review.

The SANRA authors describe narrative reviews as non-systematic and propose attention to the review's importance, aims, literature search, referencing, evidence level, and endpoint data. We use those ideas as editorial prompts. We do not claim a formal SANRA score, external peer review, or certification.

Question and scope

Before searching, the editor writes a narrow question, the intended reader, the practical decision the page might inform, and what the page must not claim. Scope includes relevant parent group, child age where applicable, outcome, context, and safety boundary.

For example, “Does this exact five-step routine improve father-child relationships?” requires direct comparative evidence. “What does work-recovery literature suggest about making a calmer transition?” can be answered as a cautious evidence note, provided the page labels inference and does not call the routine proven.

Search and selection

The current workflow uses targeted searches in PubMed and on relevant government, public-health, and professional-body sites. Queries combine the fatherhood or parenting concept with the outcome, population, and evidence type. Editors may follow reference lists to locate a foundational study or more direct source.

Priority generally goes to:

1. current public-health or professional guidance for safety and established practice boundaries;
2. systematic reviews and meta-analyses with visible methods;
3. father-specific primary studies when the design directly matches the claim;
4. foundational measurement or theory papers when needed for a definition.

Recency alone does not make a source better, and older foundational work is not automatically current guidance. Commercial blogs, unattributed summaries, and search snippets are not used as the evidentiary endpoint for a health-adjacent claim.

Extraction and claim writing

For each included source, editors record the design, population, setting, sample information when relevant, publication date, measured outcome, and limitations needed to interpret the nearby claim. Language follows design:

• Observational: “was associated with,” not “caused.”
• Qualitative: “participants described,” not a population prevalence.
• Experiment: describe the tested intervention, comparator, duration, and population before generalizing.
• Systematic review: report heterogeneity, population gaps, and the quality or limits the authors identified.
• Public-health advisory: identify it as guidance or synthesis, not a new controlled trial.

We distinguish a source-backed finding from an editorial inference. Practical scripts and routines are usually translations or combinations of evidence and professional guidance; unless directly tested, they are not described as validated interventions.

What this process does not currently include

The current pages do not have a prospectively registered protocol, multi-database reproducible query appendix, duplicate screening, duplicate extraction, formal risk-of-bias instrument, certainty-of-evidence rating, quantitative synthesis, or academic peer review. PRISMA 2020 is a reporting guideline for systematic reviews; because this library does not meet that method, it should not borrow the label.

Updates and corrections

Each page displays a reviewed date. That date changes only after a substantive source and claim review. A correction should identify the affected claim, inspect the primary or authoritative source, update the language and citation, and record the new review date. Readers can report issues through support.

Minimum bar for future original research

If Bestead later publishes an anonymous survey or privacy-safe aggregate product analysis, it must be visibly separate from these narrative pages and include, before any headline findings:

• a dated protocol and research question;
• recruitment channel, field dates, eligibility, consent basis, and sample characteristics;
• the full questionnaire or data dictionary where safe;
• prespecified exclusions, missing-data rules, analysis code or a detailed analysis plan;
• privacy protections, minimum cohort thresholds, retention rules, and suppression of risky small cells;
• funding, conflicts, incentives, deviations from protocol, and limitations;
• aggregate outputs sufficient for independent scrutiny without exposing individual data.

Product telemetry collected for operations does not automatically become research data. Consent, legal basis, reasonable user expectations, minimization, and re-identification risk require separate review. No individual mental-health entry should be published or made inferable.

Safety and clinical boundary

Research synthesis cannot diagnose a father, assess a child, or replace care. Pages should not convert group averages into personal predictions. When a topic can involve violence, self-harm, abuse, medical symptoms, or inability to supervise safely, the page states the urgent boundary and points to qualified local help.