Early in the development of clinical research it became clear that it can be very difficult to disentangle the real effect of an intervention or treatment from the natural background variation in the way patients feel. In other words patients may feel better or worse and it is nothing to do with the study treatment. To address this problem, study designers often observe patients or volunteers who have not been given the active treatment and compare their progress with those that have. The former group is termed a control group and such a study a controlled study. To minimise bias such studies are often randomised (there is an equal chance of any individual being allocated to either the active or the control group) and (double or single blind – see ‘blinded trial’) placebo controlled.