Data Collection Methods

Sampling

Sampling is the method of gathering information from a small subset of a group rather than polling every member. This technique allows the analyst to gather statically accurate data in a relatively non-invasive fashion. Sampling is a popular method used in both process analysis and quality control and also allows the analyst an opportunity to gather a lot of data with only a small commitment of time. However there are some drawbacks. For instance, sampling data can be skewed if you don't poll the correct number of people. You can also miss important contributors whose individual opinion could be valuable and it can take time if your sample size is large. However sampling is still a highly effective method to evaluate all the employees involved in a particular process with a limited amount of resources.

Ideally, if appropriate to the research question, a research effort would randomly sample. However, keep in mind that a so-called random sample rarely is truly random. First, research generally randomly samples from a defined population, such as particular area codes or particular military units, schools, etc. The defined nature of the population itself precludes total randomness. Second, the contact method also removes some randomness. For example, if respondents are to be contacted by telephone landlines, the portion of the population that does not have a landline is inherently excluded. A researcher may not be able to address these issues, but he or she should acknowledge such weaknesses in the study sample's "randomness."

Margaret C. Harrell, Melissa A. Bradley (2009) Data Collection Methods: Semi-Structured Interviews and Focus Groups, 39.

With sampling there is always the chance that your sample will contain views which might skew your data collection. It's important to review all the sample data and establish a set of norms or baselines. From this you can derive a standard deviation to know when the data is out of line. When data falls outside the norms, it should be examined closely to make sure it's not the result of bias. While fluctuations are expected, but shouldn't be ignored. Despite this possibility for error, sampling is still a proven tool for gathering accurate information with minimal resources and a valuable tool every analyst should understand and utilized especially in a volatile organization.