A random sample is defined as a sample in which any person has an equal chance of being included. For example, if your station is surveying members, and each of their names was put into a barrel, mixed up throughly, and selected individually, this would be a random sample. Conversely, If only members with six- letter last names or with first names beginning with the letters B, G, R, and T were eligible for selection, this would NOT be a random sample.

Assuming that a database of names is being used, the best way to get a random sample is by having a computer program select names at random. Some software has a "random number generator" that will weed through your list and select names randomly. If your computer system has this capability, use it!

If your computer does not have this capability, the next best uncomplicated, manual method to achieve random results is:

• Use a random number table (this can be found in many statistics books).
• Let's assume this will be a telephone survey of members. Print out your membership list. Give each caller a stack of, let's say, 20 sheets with names and phone numbers on them.
• Have the callers shuffle and mix up the sheets, almost like they were shuffling a deck of cards.
• Using the random number table, have them close their eyes and point to a number on the random number table. Or just start at one corner of it.
• Assume the first number selected is #27. Starting from the first list of member names, go to the 27th name on the list and call that person. Mark off their name on the list.
• Following that call, look at the next number on the random number table. Maybe it's 3. Go three names down the list, and make the next call.
• Keep going through the random number table to select numbers and down the list of names using those numbers until it is exhausted. Then start the process again, if necessary, with a new list of names.

An alternate method that many stations use is to take a list and choose every seventh or every hundredth name for the sample. This is called a systematic sample, and while it does not result in a sample as random as the methods mentioned above, it is the next best technique available.

[ Chapter 3 Table of Contents | Next Section | Toolkit Home ]