Screening, Data Extraction and Risk of Bias

Between a search and a synthesis lies the real work: screening hundreds of records down to the studies that qualify, pulling their data into a consistent form, and judging how far each one can be trusted. This level walks you through all three, the craft that makes a review defensible.

Who this level is for

Reviewers and their teams doing screening, extraction and appraisal, plus data staff supporting them. Recommended after Levels 1-2.

What you'll learn

• Two-stage screening (titles and abstracts, then full text) and how to apply criteria consistently
• Screening tools, from simple spreadsheets to Rayyan and Covidence, and how to handle disagreements
• What data to extract, and how to design and pilot a clean extraction form
• Why risk of bias matters and how it shapes confidence in the pooled result
• Risk-of-bias tools for randomized trials (Cochrane RoB 2) and observational studies (Newcastle-Ottawa, ROBINS-I), at a practical level
• How to summarise risk of bias in clear tables and traffic-light figures

You finish with: Screening decisions, a piloted extraction form, and a risk-of-bias table.