Identify the population of interest. For most studies, you should be able to cut the population into groups, such as sex, class or size. The results of your study will be applicable only to the population you are sampling from. For example, if you intend to say something about bankruptcy in America, your population must include all known bankrupt American companies, past and present.
List every individual in the population.
Assign a number to each individual on the list.
Determine the sample size of the study. Most samples should contain at least 20 items. Call the sample size of your study "n."
Randomize the set of numbers on your list. Many statistical software programs can take a list of numbers and output it in a random ordering. Some websites also offer such applications.
Record the first n numbers on the randomized list.
Find the random sample, which consists of the items corresponding to the numbers that you recorded. All other items can be discarded.