For example, if you are using a survey to gather patient satisfaction feedback by email, you would not send a survey t o ever y patient. You would start by sending surveys t o roughly 50 percent of the patients an see how many are returned. This limited survey allows you to determine the response rate. Assume that 25 percent of these patients return the surveys. The next task is to determine how representative of the total population these respondents are. To test this question, you need to develop a profile of the total population. Typically, this profile is based on standard demographics such as gender, age, type of visit, payer class, and whether the respondent is a new or returning patient. If the distribution of these characteristics in the sample is similar (within 5 percent) to that found in the total population, you can be comfort able t hat your sample is reasonably representative of the population. If the characteristics of the sample and the population show considerable variation, however, you should adjust your sampling plan. This example clarifies that:
- A well-drawn sample, therefore, should be representative of the larger population
- The basic purpose of sampling is to be able to draw a limited number of observations
- Sampling is probably the most important thing you can do to reduce the amount of time and resources spent on data collection
- Sampling consists of series of compromises and tradeoffs