Research

How We Review Evidence

We use research to ask better questions. These are evidence reviews, not formal systematic reviews.

  1. QuestionNeeds, alternatives and proposed responses
  2. SourcesStudies, datasets and access limits
  3. LimitsUncertainty, interests and corrections

Start with a specific question.

Each report separates the underlying need, existing alternatives and the proposed EAG response. A problem can be real without proving that a particular new service is the best solution.

Keep the setting attached to the finding.

Reports identify who was studied, where, when and under what conditions. Some include Arkansas evidence as a particular school-system context. Those findings are not automatically estimates for Boston, Massachusetts or every school. Adult, college and international studies keep those boundaries too.

Read what the source actually supports.

We prioritize original studies, official datasets, evidence syntheses and relevant institutional sources. A provider’s own description can explain its stated service; it does not independently prove effectiveness. A research recommendation is different from a tested result. Abstract-only access and unresolved retrieval limits stay visible.

Show supportive and challenging findings.

A report should explain favorable, contradictory and null results. An association does not establish cause. A result for a whole teaching program does not establish the effect of a short website feature. Figures retain their units, measured periods, populations and text alternatives.

Separate research from our own proposals.

A section about how EAG could respond is a proposal, not a finding from an outside study. Earlier product definitions remain attached to the edition that examined them. A new product direction needs a current scope note or a separately reviewed edition, not a new label pasted over old evidence.

Be clear about who is publishing.

Education Angel Group develops the products discussed here and has an interest in their success. These are EAG reviews. The reports do not claim independent peer review or formal methodological certification.

Correct the record.

An edition may receive a correction, be marked as needing an update, or be withdrawn. Send the title, section and source behind a correction request. A substantive change needs a visible version record; publication and review dates should describe work that actually happened.