Read the research. Understand the evidence.
Understand the study. See how far its findings reach.
What we assess.
Record each judgment, and the reason behind it.
Risk of bias
Could study design, conduct or selective reporting affect credibility?
Consistency
Do studies reach similar results? Can differences be explained?
Directness
Do the population, intervention, comparison and outcome directly address the question?
Precision
Are the sample and uncertainty intervals sufficient to support this claim?
Publication bias
Could unpublished, negative or unfavourable results be missing?
These dimensions help readers examine the support for a conclusion. Assessment records the evidence, applicability and uncertainty, with reasons when a dimension does not apply.Methods reference: Cochrane Handbook, Chapter 14 ↗
Grading framework and application status
A–D summarizes an assessment of evidence for a specific question, published with the search, reasoning and verification record. Research notes present sources, findings and applicability directly.
The A–D proposal below is in methodological development and has not completed methods review; no formal ratings are currently published. It is TAP’s editorial communication framework and is distinguished from formal GRADE assessment.
AHigh certaintyHigh confidence in this specific claim.
The main limitations identified in the appraisal are unlikely to change the judgment. High certainty does not mean the claim applies to everyone or that the effect is large.
BModerate certaintySome confidence, with limitations that still matter.
Further research or more direct evidence may change the judgment. Read the scope and grading rationale alongside the grade.
CLow certaintyLimited confidence; the conclusion may change substantially.
Study limitations, inconsistency, indirectness or imprecision affect the judgment. Preliminary findings should not be presented as assured effects.
DVery low certaintySubstantial uncertainty; a reliable judgment remains difficult.
Grade D does not mean an intervention has been shown to be ineffective. Evidence of no important effect and a lack of reliable evidence are different situations.
UnassessedThe assessment is not complete. This does not mean low certainty.
Not applicableOutside the scope of this grading process. A reason must be provided.
Content type and grade, explained separately.
Scientific concept
Explains concepts, definitions and boundaries used in research. A concept does not receive a single overarching evidence grade.
Research explainer
Organises and appraises research around a specific question, explaining the evidence and certainty for each important claim.
Project model
An explanatory framework, metaphor or educational simulation developed by this project, with explicit assumptions and limits.
For example, the Allostatic Load feature explains a scientific concept and connects specific statements to research. The recovery curve is an educational model for exploring change over time, not measured data.
How a grade is reached.
Keep a record of the process before displaying a formal grade.
Explore a claim and its sources- Define the question
State the claim, population or scope, comparison and outcome. First determine whether grading is appropriate.
- Document the search
Record databases, search terms, dates and eligibility criteria, including how null findings, contrary evidence and unpublished studies are handled.
- Assess each domain
The assessor records the evidence for each judgment, key limitations and the final grading rationale. Use methods appropriate to the question.
- Have another person verify it
Record the actual assessor, verifier, date and method version. Incomplete records remain “Unassessed”.
- Updates and corrections
Reassess when new evidence emerges or an error is found. Keep a record of important changes and their reasons.
Read findings in context.
Grades help readers understand uncertainty. They are not diagnostic tools and cannot be converted directly into product efficacy ratings or purchasing recommendations.