A researcher wants to perform a randomised controlled trial of a new lipid reducing agent in diabetics. On the basis of a pilot study in 25 diabetic patients she estimated the mean total cholesterol as 5.8 mmol/L with a standard deviation of 1.5 mmol/L. She is limited to 200 patients in total for her randomised controlled trial. With this sample size, what is the minimum mean difference in cholesterol she would be 80% sure of detecting between the treatment and placebo groups based on a 5% two-sided level of significance?
Could anyone explain this to me?
How are you celebrating?