ABSTRACT
The main aim of most well-conducted randomised trials is to obtain unbiased estimates of treatment effects that are sufficiently precise to answer the research questions of interest. Eliminating or minimising bias rests on careful trial design and implementation, and relies on standard methods, such as random allocation of treatments, blinded evaluations and maximising follow-up rates. There is little point in minimising bias, however, if random errors are large due to an inadequate sample size.
