Step 4: Sampling
Sampling, the process of selecting individuals, households or non-households to be interviewed, is critical to the success of any survey. Sampling is also a topic that is difficult for many people to understand. When good data on a population of interest are available, sampling can be carried out according to well-established statistical principles. Such samples--obtained using what is either termed “formal” or “probability” sampling--can be used to carry out analysis from which confident statements about the characteristics of the population can be made.
Below are some terms that are used in surveys and sampling:
- A survey is some means of obtaining information about the characteristics of a group by observing members of the group.
- Population is a term that denotes a group of interest or focus within a survey.
- A sample is a subset of a population.
- Sampling refers to the process of selecting cases from a population that will be interviewed within a survey.
- A sampling frame is a list of members of a population which can be sampled.
- Sample size refers to the number of cases interviewed within a survey.
- A population characteristic is a measurement of a characteristic of a population.
- A sample characteristic is a measurement of a characteristic of a sample.
These terms, their meanings, their uses, and their importance will become clear as we use them in the discussions and in the examples which follow. Please refer to these definitions as needed when reading the text below. We will define other terms as we proceed.
The fundamental goal of sampling is to select a subset of the population, a sample, that is suitable for analysis. A sample is suitable if it (1) is of sufficient size and (2) is selected from the population such that the risk of obtaining results from analysis of the sample that do not accurately reflect the characteristics of the population is within an acceptable range.