A pharmaceutical company has a drug it claims will reduce your cholesterol and needs to do a test to show the government, stockholders, and public how effective it is.
They need 100 people for the test, so they offer it to 200 people. Of those 200 test subjects, 80 of them had horrible side effects, and many became very sick.
The pharmaceutical company tells the 80 subjects that they are concerned about them and does not want to cause more harm, so they let them go. The company also asks 20 other subjects that are likely to have bad results to leave as well.
Now they have their prime 100 candidates to choose from and when the test results come out they will use these numbers - Unless - those numbers are bad in which case they will use "Relative Risk" factoring to make the numbers look better.
Do your research BEFORE you do anything!!!