Research terms explained
A short reference for the terms you meet in the wiki. Read a definition, then return to the study.
Explore the studies- Randomized trial
A study that uses chance to allocate participants to intervention groups. Random allocation helps prevent systematic differences at the start, but other sources of bias can remain.
Read the reference- Sham comparison
A simulated intervention used as a comparison. Its usefulness depends partly on whether participants can distinguish it from the active treatment.
Read the reference- Outcome
What the study measures, such as pain, walking ability or an adverse event. Always read the measurement scale and assessment time.
Read the reference- Mean difference
The difference between two group averages, expressed in the original scale’s units. Its direction depends on how that scale is scored.
Read the reference- Confidence interval
A range expressing statistical uncertainty around an estimate. A wide interval usually means less precision; it does not remove bias in the study.
Read the reference- Statistical significance
A statistical convention tied to a P-value threshold. It does not, by itself, establish a large effect, a clinically important benefit or high-certainty evidence.
Read the reference- Risk of bias
The possibility that study design, conduct or reporting systematically distorts a result. In RoB 2, assessment concerns a particular result, not just a paper as a whole.
Read the reference- Certainty of evidence
Confidence in an estimate for a particular outcome across the evidence. GRADE uses high, moderate, low and very low certainty, considering limitations beyond study design alone.
Read the reference- Systematic review
A review using explicit methods to find, select and assess relevant studies. A meta-analysis may combine results statistically; combining studies does not automatically make the conclusion reliable.
Read the reference- Clinical review
The editor’s review of a study page, dated at the end of the page. It is separate from checking the PDF and the extracted numbers, and it is removed if the page changes until it is done again.
Sources for these definitions
- Cochrane · Risk of bias
- Cochrane · Effect measures
- Cochrane · Interpreting results
- Cochrane · Certainty of evidence
Sources consulted on 22 October 2026.