Sampling question #3: How do I draw a simple random sample?
There is a method to the madness of drawing a random sample. You must abide by the rules of the game in sampling in order for your sample to be representative of the population.
1. The first step is to identify all of the members in your population. You must be able to list them in what is called a sampling frame. The frame should have the names without order to them and non-overlapping (no duplicates). Alphabetizing the list by surname is a way to insure a random order in the sampling frame.
2. Second, you must give each name an identification number. Start with “1” and continue.
3. Third, you must decide what the size will be for your sample. You can use the table suggested in the Krejcie (1970) article, or whatever feels right for you to believe in the results you obtain. As a rule of thumb, use as large a sample size as possible.
4. Fourth, you need to get a Table of Random Numbers. Many are located at the end of statistical or mathematical textbooks. The Table of Random Numbers consists of rows and columns of numbers arranged at random so that they can be used as any point by reading in any direction left or right, up or down. We are now ready to draw our random sample.
A Really Simple Example
Fifth, we know that our largest ID number has two digits (20). So we are going to need a two-digit column in the Table of Random Numbers. We close our eyes and put our finger down anyplace in the Table of Random numbers. This is our starting point. We have decided in advance whether we would use two digits going up, down, left or right. So we begin.
Sixth, Number #6 is the first ID number that is within our band of between 1 and 20. This is the first member of our random sample and it is Deanna. We continue 82, 56, 96, 66, 46 until we and come up with the next is ID #13 Krista. Our three last members of the sample are ID #8 (Emily), ID #5 (Brett) and ID #4 (Amelia). We have five randomly selected members: Deanna, Krista, Emily, Brett and Amelia.
Return from random sample to samples and sample size.