Seven myths of randomisation in clinical trials

Abstract

I consider seven misunderstandings that may be encountered about the nature, purpose and properties of randomisation in clinical trials. Some concern the practical realities of clinical research on patients. Others are to do with the value and purpose of balance. Still others are to do with a confusion about the role of conditioning in valid statistical inference. I consider a simple game of chance involving two dice to illustrate some points about inference and then consider the seven misunderstandings in turn. I conclude that although one should not make a fetish of randomisation, when proposing alternatives to randomisation in clinical trials, one should be very careful to be precise about the exact nature of the alternative being considered if one is to avoid the danger of underestimating the advantages that randomisation can offer.