ZACH / Free learning tool 04
Does this study actually
support the claim?
Seven questions to bring a research headline back to the people, methods and results behind it. Read a paper alongside the tool—or start with a teaching example.
Your entries stay in this page. No account is needed. The tool does not fetch or automatically assess a paper. Download your notes before leaving; entries are not saved.
Step 1 of 7 · The claim
Teaching example loaded. The office survey is fictional; it illustrates a reasoning problem, not a finding about movement breaks.
A narrower, more useful conclusion
Your reading notes.
These prompts reflect the answers you selected. They are not an independent appraisal, a validated score or a verdict on the paper.
Try a careful sentence.
In [population] , [intervention or exposure] compared with [comparator] was [associated with / estimated to change] [measured outcome] over [time]. The estimate was [effect and uncertainty]. Important limits are [limits].
Choose causal language only when the design, conduct and wider evidence justify it. One paper does not settle every practical decision.
Your additional notes
Why there is no score.
A sum of ticks can hide a decisive weakness. A study may be carefully conducted and still answer a different question. This worksheet makes the reasoning visible, with room for “I do not know”.
It is an original ZACH learning aid, not a substitute for formal risk-of-bias assessment or a systematic review. GRADE evaluates certainty across a body of evidence for a particular outcome; this worksheet does not assign a GRADE rating.
Go to the methods.
The questions draw on established principles of transparent reporting and evidence interpretation. These sources have different purposes; none endorses this tool.
- CONSORT 2025 — reporting randomised trials
- STROBE — reporting observational studies
- Cochrane Handbook — effects and certainty
- American Statistical Association — interpreting p-values
Reporting checklists improve transparency; they are not automatic study-quality scores. Reviewed 4 October 2026.