Survey: Sampling Issues

Sampling Frame

Specifying the sampling frame is key to overcoming sampling issues in survey research. A sampling frame is all of the available elements or units of a population that have a chance of being selected for the survey. Researchers need to think about the access they have to these elements. First, researchers should ensure that all members of a population can be sampled and that there is a complete list or record of population units. Second, researchers need to ascertain what information they have about the possible participants in terms of contact information (names, addresses, phone numbers, and e-mail addresses). Finally, researchers need to establish who the actual participant is. For example, if they are surveying small businesses in a community (the unit or element), who is the actual person that they want to talk to from each business. Is it the owner, a manager, or an employee? Additionally, they need to decide what their course of action is if that person is unwilling or unavailable. Failure to think through these questions and not have a complete sampling frame may lead to poor representation.

Sampling Technique

Once the sampling frame has been established, researchers need to strategize how they will select those elements of the population that are available to them. The goal of survey research is to select a sample that allows researchers to generalize their findings to the larger population. Survey sampling can fall into one of two categories: probability or nonprobability. Probability sampling is when all elements of a sampling frame have equal chance of being selected for the sample. This type of sampling is random, reducing bias and increasing the likelihood that the sample is representative. Random sampling could involve using random numbers to select the sample (simple random sample), choosing every nth numbered person to select the sample (systematic random sample), or randomly selecting survey participants from groups within the population (e.g., stratified random sampling or cluster sampling). When probability sampling techniques are not feasible, researchers rely on nonprobability techniques such as convenience sampling, in which participants who are accessible are used; quota sampling, in which convenience samples within subsets of the population are surveyed; and network or snowball sampling, in which current participants pass the survey on to their contacts. Nonprobability techniques are popular, especially with the increases in online surveys, but pose several issues when it comes to representativeness of the findings.

...