. . . . "The researcher needs to adequately address problems such as sample size, blinding, randomisation and appropriate use of control and treatment." .