ABSTRACT
In any clinical trial, reasonable allocation of subjects to treatments under consideration is often the most challenging task. Most of the early clinical experiments adopted arbitrary schemes for treatment assignment. Randomization, the systematic procedure of assigning subjects to different treatments, was contributed by Sir R. A. Fisher (1935) in the context of assigning treatments to plots in agricultural experiments. However, in a clinical trial’s scenario the purpose of randomization is to produce groups similar with respect to all risk factors that might affect the outcome of interest, apart from the treatments. Randomization provides a degree of trade-off between balance, ensuring that the desired proportion of subjects per treatment arm is approximately achieved, and nonpredictability, ensuring that clinicians referring subjects for enrollment cannot dictate the next assignment.
