End of Module A2 Quiz

What is the definition of ‘power’?

What information/parameters do you need before you can estimate the sample size needed for a particular study? Choose all that apply.

What information/parameters do you need before you can calculate the power of a given sample size and study? Choose all that apply.

Estimate the total sample size needed to compare the mean total cholesterol between two groups of men, with versus without coronary heart disease (CHD). Based on a previous study, you think the mean total cholesterol will be 5.5 mmol/l (SD: 1 mmol/l) amongst men without CHD, and 6.0 mmol/l (SD: 1 mmol/l) amongst men with CHD. Assume you want 90% power, and a 5% significance level. (Answers below are approximate as different software will give slightly different answers, but they should be close)

Which of the following increases the power of a study (when you are comparing a difference in means)? (choose all that apply)

What is a Type II error?

In order to estimate the sample size for a case-control study, you need to run a test for a difference in means.

In order to calculate the needed sample size for a study to have 90% power, you need estimate the expected difference in means/proportions you are likely to find in the study. With this information, you can run a sample size calculation. How do you estimate the expected difference in means/proportions you are likely to find in your study? (choose all that apply)

What is an acceptable level of power for a study? (choose all that apply)

You are running a placebo-controlled randomised trial of a new therapeutic drug that does actually reduce the risk of having a heart attack. However, over the course of the trial, 10% of the participants in your intervention group stop taking the new drug due to side effects. Will this reduce the power of your study to detect an effect?