Methods and Applications of Statistics in Clinical Trials, Volume 1: Concepts, Principles, Trials, and Designs

Chapter 68

Randomization Procedures

Vance W. Berger, William C. Grant, Therese Dupin-Spriet, and Rick Chappell

68.1 Basics

Randomization is an important aspect of any experiment that compares the effects of some treatments on some outcome or outcomes. Note that any evaluation of a treatment is necessarily comparative, because a treatment cannot be “good” or “bad” in a vacuum. Rather, it is good or bad relative to other possible treatments (including the lack of any treatment), so randomization is a key component of the evaluation of any treatment. Through randomization, an experiment’s observer can better distinguish between the treatment effects and the effects of other factors that influence the outcome. In this section, we will explore how randomization accomplishes this.

68.1.1 Rationale for Randomization

Without randomization, treatments tend to be assigned based on prognostic factors, such as severity of disease. The experiment is confounded if sicker patients receive one treatment (perhaps an elective surgery) and healthier patients receive another treatment, because it can be difficult to separate the effects of the treatments from the effects of underlying differences across the treatment groups. Randomization is meant to prevent this confounding by removing the influence of patient characteristics (including preferences) on treatment assignments.

68.1.2 Defining Randomization

Randomization is the use of chance (probability) for assigning treatments. A more precise definition ...