Randomization

Randomization protects a comparative study from selection bias, because neither investigators nor participants decide who gets which treatment. Reviewers look at how the allocation was generated and concealed, not only at whether the protocol calls the study randomized.

process of assigning subjects (3.50) to the investigational medical device (3.29) or control groups (3.15) using an established recognized statistical method using an element of chance to determine the unforeseeable assignment in order to reduce bias

ISO 14155:2020, clause 3.42

The process of assigning participants to groups such that each participant has a known, and usually an equal, chance of being assigned to a given group.

FDA guidance: Design Considerations for Pivotal Clinical Investigations for Medical Devices (November 2013), Section 11, Glossary