Recently added item(s)
You have no items in your shopping cart.
Personalize your experience with Penn State Extension. Subscribe today!
Save For Later Print
Updated: December 6, 2017
Generally, random samples are the best way to ensure that the sample represents all your customers. This means that you can safely assume that the patterns you see in your fi ndings are the same as the patterns you would have seen had you contacted all your customers. Random sampling means that each customer has an equal probability of being selected into the sample. This works best if you have a complete list of your customers.
Approaches not based on random sampling techniques may also be used. These are appropriate when you don't have contact information for your customers, or when you're interested in information from a specific set of customers.
The limitation of these non-probabilistic approaches is the potential for bias in the results because not every individual has an equal chance to be chosen. For example, you might give a survey to people who purchase products during one day at the farmers' market. However, this approach excludes all those who did not attend the farmers market that day and all those who did not buy products. The survey results would only represent the small subset of people who purchased products that particular day, and not all customers of your booth.
You should decide if obtaining a random sample is important or not. This decision can be based on your reason for gathering feedback as well as the practical costs of doing so. Regardless of your ultimate choice. Some thought should be given to the appropriate sample size and how you will obtain your sample.