ABSTRACT
The whole enterprise of sampling is a major preoccupation in health research, and many difficult mathematical problems bedevil it. In doing clinical research, for example, you might want to ascertain the effect of a particular treatment on males over the age of 50 years who are suffering from arthritis of the shoulder. Possibly, you will have difficulty even finding many shoulder arthritis sufferers who are 50 or over and who are also male. Then, on the other hand, you might have access to more than you can possibly test. If the former situation prevails, your sample will need careful scrutiny to make sure that it is not biased.
