Alternating patients is not the same as random

Question

A clinic assigns patients to treatment or control by strict alternation: the first arrival gets treatment, the second control, the third treatment, and so on. A statistician warns that this is not truly randomized and can be gamed.

Explain how a predictable alternating rule can be corrupted, and why concealed randomization avoids it.

Accepted Answer

Predictable assignment can be steered. With strict alternation, staff always know what the next patient will receive. A well-meaning doctor who believes the treatment helps might, consciously or not, delay a frailer patient so a healthier one lands in the treatment slot, or nudge borderline patients toward the arm they prefer. Once assignment can be anticipated, the person controlling the door can shape who ends up where. That reintroduces exactly the kind of imbalance randomization is meant to prevent: sicker or healthier patients concentrating in one arm, a confound with the treatment. Why …

Alternating patients is not the same as random

Your answer

More Statistics questions