Statistical Sampling

Posted on June 5, 2009 by


There is much to know about selecting a good sample and it should be done carefully by someone knowledgeable. A good sample can accurately describe the whole population with a fraction of the cost and time it would take to measure the entire population. Project managers may use statistical sampling as part of performing the fourth edition PMBOK®’s quality processes. The first step is to define the population of interest, such as every full-time employee in a certain organization. Then a sampling strategy, such as random sampling, stratified sampling, cluster sampling or others, is used to select the individuals who will be in the sample. Random sampling selects a sample of employees with each one having an equal chance of being selected. Stratified sampling first divides the population into strata—for example the organization’s population could be grouped by men and women, and then a separate sample drawn from each group to assure that the desired number of men and women are included in the sample. Cluster sampling samples a sub-group of the population. For example, everyone in the marketing department could be chosen instead of choosing people from throughout the organization. It is also important to determine the correct size of sample. Too small, and your data will be misleading, while an unnecessarily large sample will waste resources.