NYFERA’s research standards support accurate, transparent explanations of New York education. They apply to policy guides, school and district profiles, research summaries, data analysis, and supporting resources.
Our purpose is to help readers understand what the evidence shows, how it was obtained, and where uncertainty remains. These standards establish expectations for research and publication; they do not imply that every article is an academic study or systematic review.
Research begins with a clearly defined question. The question should identify the relevant subject, population, location, and period. Researchers should distinguish questions about how a policy operates from questions about whether it produces a particular outcome.
The scope must be appropriate to the available evidence. A school profile, statewide statistical analysis, and evaluation of an educational intervention require different sources and methods. Conclusions must remain within those boundaries.
Source quality depends on the claim being investigated. Official records are generally the starting point for school identity, published requirements, accountability designations, and administrative data. Original research and well-conducted evidence reviews are important when examining effectiveness or educational outcomes.
An official publication establishes what an institution reports; it does not automatically establish that every interpretation or policy claim is correct. Likewise, a peer-reviewed study requires examination of its design, limitations, and relevance.
School websites, advocacy publications, press releases, interviews, and news reporting may provide useful information. Their purpose and interests must be considered, and promotional claims must remain distinguishable from independently supported findings.
Important claims must be verified against the underlying material. Researchers should read the relevant document, table, or study rather than rely solely on a headline, abstract, search snippet, or another publication’s summary.
Verification includes checking authorship, publication date, document version, definitions, quotations, calculations, and whether a source actually supports the proposed wording. Consequential or disputed claims require additional corroboration where available. Several websites repeating the same unsupported statement do not constitute independent confirmation.
Where a reliable primary source is sufficient for a straightforward fact, researchers should identify it directly rather than imply that multiple sources were necessary.
Searches must not be designed only to confirm a preferred conclusion. Research should seek relevant contradictory findings, alternative explanations, and evidence identifying limitations.
For substantial research reviews, the published methodology should explain the search period, principal sources searched, selection criteria, and important exclusions. A selective overview must not be described as comprehensive or systematic.
Formal systematic reviews require a defined protocol and documented search, screening, appraisal, and synthesis procedures. The What Works Clearinghouse describes these as distinct stages in its own systematic review process. Referencing that framework does not imply that NYFERA has WWC certification or endorsement.
Studies must be assessed beyond their conclusions. Review should consider:
- The research question and study design.
- The participants, setting, sample size, and selection process.
- The comparison group and relevant baseline differences.
- The outcome measures and length of follow-up.
- Missing data, participant attrition, and analytical assumptions.
- Effect sizes, uncertainty, and practical importance.
- Funding, conflicts of interest, and publication status.
A finding should not be described as causal merely because one group performs better than another. Research designs differ in their ability to support causal conclusions. Qualitative evidence can explain experiences and implementation without establishing a numerical treatment effect.
Statistical significance does not, by itself, establish educational importance. An inconclusive result must not automatically be described as proof that an intervention has no effect.
Evidence must be relevant to the population discussed. Findings from another state, age group, admissions setting, or historical period may offer context without directly answering a question about New York schools today.
Researchers should explain meaningful differences in implementation, student characteristics, resources, and institutional conditions. Results must not be generalized beyond what the study and surrounding evidence reasonably support.
Education data require their own documentation. Definitions, reporting periods, eligibility rules, and aggregation methods must be checked before figures are compared or interpreted.
NYSED provides glossaries, business rules, and database ReadMe documents explaining these elements. Research using its report-card data should consult the documentation for the relevant release.
School identifiers must be matched carefully. Missing and suppressed values must remain distinct from zero. Changes in assessments, cohorts, boundaries, or reporting practices must be considered before presenting a trend. NYFERA’s Data Methodology sets out additional requirements.
Research records should make material conclusions traceable. Supporting notes should identify the documents, versions, source locations, selection decisions, and calculations used. Published explanations should provide enough information for readers to understand and assess the approach.
Translated material must preserve the original meaning and qualifications. AI-generated summaries, citations, and calculations require independent checking under NYFERA’s AI Policy.
Confidential student information must not be solicited or exposed to substantiate a public article. Public availability does not remove the need to consider privacy and potential harm.
Relevant conflicts of interest must be disclosed and managed. Sources, schools, sponsors, and contributors must not receive authority to determine findings in exchange for access, funding, or cooperation.
Before publication, the responsible publisher or editor should assess whether the central claims are supported, limitations are visible, and the wording matches the strength of the evidence. Specialist review must only be claimed when it actually occurred.
New evidence, revised datasets, retractions, and substantiated reader concerns may require reassessment. Material errors must be addressed under our Corrections Policy. Readers can submit supporting documents or research concerns through our Contact Us page.